26th Edition - BaseballHQ.com
Transcription
26th Edition - BaseballHQ.com
26th Edition Copyright © 2011, Fantasy Sports Ventures, Inc. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form by any means, electronic, mechanical, photocopying, or otherwise, without the prior written permission of the publisher, Triumph Books, 542 South Dearborn Street, Suite 750, Chicago, Illinois 60605. Triumph Books and colophon are registered trademarks of Random House, Inc. This book is available in quantity at special discounts for your group or organization. For further information, contact: Triumph Books 542 South Dearborn Street Suite 750 Chicago, Illinois 60605 (312) 939-3330 Fax (312) 663-3557 Printed in U.S.A. ISBN: 978-1-60078-587-0 Rotisserie League Baseball is a registered trademark of the Rotisserie League Baseball Association, Inc. Statistics provided by Baseball Info Solutions Cover design by Veronica Corzo-Duchardt <[email protected]> Front cover photograph by Steven King/Icon Sports Media, Inc. Author photograph by Kevin Hurley Ron Shandler’s BASEBALL FORECASTER Editor Ray Murphy Acknowledgments If I were to list everyone who has contributed to 26 years of this publication, it would be 10 pages long. So I am going to keep it short and sweet, comparatively. But you should read this. It’s like the cast listing in the credits at the end of a movie. And you always read the cast listing. To your left—the stars of 2012 Baseball Forecaster. They are the names that appear at the beginning of the movie, the ones that matter the most. Their hard work is what you hold in your hands. If you ever cross paths with them, online or in the street, thank them. tttttt Thank you to the rest of the BaseballHQ cast: Andy Andres, Matt Baic, Matt Beagle, Alex Beckey, Bob Berger, Rob Carroll, Pat DiCaprio, Matt Dodge, Jim Driscoll, Scott Gellman, Dylan Hedges, Phil Hertz, Joe Hoffer, Brandon Kruse, Troy Martell, Scott Monroe, Mike Murray, Craig Neuman, Harold Nichols, Frank Noto, Paul Petera, Greg Pyron, Nick Richards, Brian Rudd, Mike Shears, Peter Sheridan, Skip Snow, Tom Todaro, Jeffrey Tomich and Michael Weddell. Technical Wizard Thank you to BaseballHQ.com’s tech team: Mike Krebs, Rob Rosenfeld and the entire site redesign group. Associate Editors Rod Truesdell Brent Hershey Rob Rosenfeld Design Brent Hershey Data and Charts Matt Cederholm Brent Hershey Joshua Randall Player Commentaries Dave Adler Matt Cederholm Neil FitzGerald Brent Hershey Tom Kephart Ray Murphy Stephen Nickrand Kristopher Olson Josh Paley Joshua Randall Jock Thompson Rod Truesdell Research and Articles Patrick Davitt Doug Dennis Ed DeCaria Bill Macey Ray Murphy Stephen Nickrand Brandon Wilson Prospects Rob Gordon Jeremy Deloney Tom Mulhall Injuries Rick Wilton Thank you to Lynda Knezovich, the Queen of Qustomer Service. Thank you to all my industry colleagues and competitors, past and present. Rather than listing all five thousand of you—like I do each year—and invariably leaving someone out—like I do each year—I’ll just thank you all en masse. We belong to a great industry and you have been constant inspirations. Thank you to everyone at Big Lead Sports/Fantasy Sports Ventures and Triumph Books. Thank you to all my contacts at USA Today, the Huffington Post and anywhere else my byline may have appeared this past year. Thank you to my ladies, who we’ve all been following on this page now for more than two decades. Darielle just finished a fall semester spent in Tel Aviv, London, Rome, Naples, Zurich, Vienna and Frankfort (okay, those last three were just layovers, but still). Returning to New Jersey for the rest of her junior year might seem tame in comparison but she’ll be directing her first play in February, so there’s that. Justina has been avoiding all the scandals at University of Miami, but apparently the music students are considered the real celebrities there anyway. She’s been gigging all over south Florida, taking master classes with the likes of Jackson Browne, Livingston Taylor and Bruce Hornsby, and already had two music videos produced. So there’s that, too. Thank you to Skype. Thank you to Sue, who has been there with me during these 26 fall fire drills. It never really got easier, but as new empty-nesters, at least we now have another new beginning to look forward to. Finally, thank you… for reading this far, and providing kind words and support for all these years. You will never know how grateful I am that you have allowed me to lead this charmed life. Music rises to a crescendo. Fat lady sings. Curtain falls. Fade to black. Ron Shandler November 2011 TABLE OF CONTENTS Simulators. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Encyclopedia of Fanalytics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Fundamentals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Batters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Pitchers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Prospects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Gaming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Statistical Research Abstracts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Gaming Research Abstracts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 Major Leagues Batters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Pitchers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Injuries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 Prospects Top Impact Prospects for 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 Top Japanese Prospects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 Major League Equivalents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234 Leaderboards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 Draft Guides. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 Blatant Advertisements for Other Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275 PQS Across America . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280 A Simulators batting average on balls-in-play (hit rate) and the ability to s fantasy owners, we take pride in our ability to assess baseprevent baserunners from scoring (strand rate). More recently, ball players and to project their performance. Our success in we’ve started analyzing batters’ ability to hit the ball hard and piloting our teams to titles and money finishes demonstrates our pitchers’ ability to prevent the ball from being hit hard. prowess at these skills. These are all component skills of baseball performance. But it really doesn’t. In the same way, we can break down playing fantasy baseball Player evaluation and forecasting are vital parts of building into a series of component skills. the foundation of a winning team. But we don’t give ourselves the There are the skills of draft management. At auctions or snake opportunity to measure the results of that evaluation process. Yes, drafts, owners have to manage budgets, opponents’ moves, and we get a “report card” every October, but it’s deceptive. Our final game strategies to build competitive rosters. point totals are amassed not only by the players we draft—the Then there’s in-season management—a package of skills that results of our painstaking off-season research—but also by many includes evaluating free agents, analyzing the standings to idenother players who churn through our roster during the season. tify opportunities, and, of course, trading. We only truly exercise our player evaluation muscles during It’s easy for most of us to identify fellow owners the off-season. The stats have stopped accumuwho have—or don’t have—these individual pockets lating, allowing us to dig into them, evaluating each Every player we of expertise, right? player’s individual performance in deeper context, You know guy in your league who consistently and thus to project probable future performance. add to our roster drafts solid, well-balanced teams—but then trades Then we draft players based on that comprehensive himself out of contention. You know the guy who analysis, and our expectation of what they are going during the season comes to the draft table and knows exactly what to to do over six months. pay for any player, but is clueless about the marketBut we typically base our in-season roster moves means we’re place when bidding his FAAB dollars. And you on far shorter periods, and far smaller sample know the guy who drafts terrible teams, but wears sizes, making these decisions less accurate and the everyone down with his incessant trade offers, en results more volatile. Every player we add to our eliminating route to money finishes every year. (Yes, I hate him roster during the season means we’re eliminating too.) one player we likely rostered through more careful one player we You know who these guys are because, anecdotanalysis. And given how much turnover a typical ally, you’ve seen these traits in action, year after year. fantasy roster undergoes during the course of a year, likely rostered And having this knowledge helps you navigate your our in-season roster management can negate much dealings with the others in your league. of our off-season player evaluation efforts. through more But it’s difficult to observe, identify and assess As a result, we have only limited insight into how any owner’s competency at player evaluation. It’s not good anyone is at player evaluation. Even ourselves. careful analysis. always apparent at the draft, and then it is completely We need a way to more accurately evaluate this lost in the blizzard of in-season player moves. core competency, and at the same time, hone our And yet, this component skill is a core requirement for fantasy skills and gauge our efforts at assembling an optimal roster. We league success. could just take all our pre-season projections and compare them to the end-of-season actuals. That’s easy. However, that wouldn’t The impact of roster churn assess our skill within the context of building a team. Consider the mad rush to roster Sam Fuld after two great The answer is not immediately apparent but is intuitively weeks in April 2011. This career minor-leaguer did not even logical. merit a projection in last year’s Forecaster and showed up in our Major League Equivalents section with batting averages over the Core fantasy skills previous two years of .239 and .208. Longtime Forecaster readers and BaseballHQ.com subscribers By contrast, we projected Casper Wells to get 240 at-bats last are well versed in our mantra for fantasy success: “Draft skills, not year, with some promising power-speed skills. When I had to fill stats.” That means evaluating players by looking at the underlying my last outfield spot in a 12-team AL-only league, I thought Wells’ component skills behind the statistical output. playing-time projection and skills justified a $1 bid. For pitchers, we know to ignore ERA, to focus instead on But our 240-AB projection didn’t necessarily mean we thought strikeouts (“Dominance,” or k/9), walks (“Control,” or bb/9) and he’d get 40 AB per month; it could mean 20 AB per month for the the ratio of strikeouts to walks (Command,” or k/bb). For batters, first half of the season and 60 AB per month during the second we ignore batting average, and focus on how often they put the half. ball into play (“Contact rate”), draw walks (“Walk Rate”) and, That’s exactly what happened. During the first half with again, the ratio of one to the other (“Batting Eye,” or bb/k). Detroit, Wells was only getting 7 AB per week. In August, he was We’ve also learned to evaluate pitchers and hitters by their traded to Seattle and became a full-timer. He finished just 25 ABs ground ball rates, flyball rates and HR-to-flyball rates, and by 1 Simulators / 2012 Baseball Forecaster day-to-day) micro-managing that fantasy baseball seems to have become.” Now, I won’t say that this level of roster management is wrong within the context of winning our fantasy leagues; in fact, it’s usually necessary. Many owners find it enjoyable, and many more like the idea that they can be competitive in their leagues by being willing to outwork their opponents. But the foundation for any successful fantasy season is the assessing of player talent and projecting player performance. All this micro-managing activity is seriously undermining any effort to evaluate how good we are at that key process. shy of projection (due to a late-season injury) and still provided a nice profit on my $1 investment. Rather, he would have provided me profit. In my league, like most, I couldn’t afford to wait and see if he would end up with the playing time I projected back in March. In order to pump up my numbers, I had to cut him and grab a free agent who was getting it done now. As it turned out, that free agent was, um… Sam Fuld. It was just in time for his subsequent 9-for-77 run, which helped bring him down to a .240 full-season average. Similarly, many teams had rostered Javier Vazquez, based on his excellent skills when not playing for the Yankees. Last year’s Forecaster projected Vazquez to finish 12-9, 3.92. But many teams dropped Vazquez in May or June when his ERA was hovering around 7.00. Those owners missed Vazquez’ sterling performance from June 21 on: 10-4, 1.86, which brought him to 13-11, 3.69 for the season, slightly better than his projection. Again, we play this game for six months. If cutting Wells or Vazquez, or rostering Fuld, had been based solely on our off-season assessments of each player, we would not have made any of those moves. Instead, Fuld would have remained safely in the free agent pool. Wells would have stayed on the roster, ready to contribute in the second half, and Vazquez would have remained rostered during his steaming summer streak. And as a result, our prognosticating report card would have looked so much better. But most fantasy leagues require some level of micro-management, forcing us into decisions that greatly inhibit our ability to evaluate our own forecasting skills. Even if our off-season analyses paint glorious pictures about the players we like, our limited roster space and quest for the best statistics force us to toss some of them away to take chances on hot hands. And even though we do look at past track record, the “hot hand” analysis—based on small samples of season-to-date performance data—tends to be the overwhelming decision-maker. In other words, we put in obsessive draft prep to analytically determine which players are going to give us the best six months of production. Then we’ll undo all that work by gambling on a fringe player having a hot April. (You know the thought process: “What if this is real?!”) I wrote about the idea of baseball smarts being undone by the vagaries of competition at BaseballHQ.com, and got plenty of feedback. Two examples from our subscribers: “It became clear after winning or finishing second every year for six straight years that it was hard to separate out my baseball acumen from my tenacity on the waiver wire.” “I haven’t enjoyed playing this season. I can’t stand scanning the free agent wires for the next ‘flavor of the week.’ I don’t like the thought of finding this week’s best two-start pitchers. I shudder at nabbing a guy in hopes of squeezing an extra week out of a hot streak that screams anomaly. I like poring over stats in the fall and winter. I like reading the hopeful quotes of players, managers and columnists during spring training. I like preparing for drafts. And I like watching baseball, which is a game made for sitting back and taking it all in. None of this is remotely close to the week-to-week (or The solution: More drafts Lots more drafts. Not more active leagues, more drafts. The process of measuring our player evaluation skills begins with the draft—and it needs to end there. In-season player moves add a huge variable in overall team performance, as we’ve seen. So to evaluate our projection skills, we need to draft more teams and then see how they fare, without in-season management. Drafting just for the sake of drafting? How unfulfilling is that? Not as much as you might think. Many of us already participate in mock drafts, often by the dozen; they let us practice our drafting without burdening ourselves with playing out the leagues. They can have far greater value, though, if only we’d let them. We just need to convert them into real “draft-and-hold” (D&H) leagues—leagues with rosters but limited or no free-agent or waiver-wire moves. D&H leagues are growing very quickly in popularity. Many owners like testing their mettle purely as projectors and evaluators, and many owners like being able to play the game without committing to hours of weekly or daily effort during the season. One BaseballHQ.com subscriber wrote me to say: “My first reaction was pleasure at the idea of more drafts; I think this is why people prefer the idea of drafting in many leagues; it creates more opportunities to be right.” I like the line, “It creates more opportunities to be right.” That’s why we play: to find out if we’re right. Each year, I value several dozen players more highly than others do. I’d love to figure out how to roster all of them and to test my expectations against others. But I’m limited by the number of leagues I’m in and I can’t commit to more leagues than my time will allow. Playing multiple D&H leagues would let me accomplish those goals. D&H leagues are truer representations of our prognosticating prowess. They let us build initial rosters based solely on our offseason analytical efforts. These leagues let us watch and analyze, but remove the temptation to continually tinker and tweak. Of course, one league would clearly not be enough. We’d want to be in many leagues. Because really, what is the most fun part of this hobby, anyway? The drafting! So let’s draft lots and lots of teams. More D&H leagues give us more chances to roster more of the players we like, not just the 23 we end up with. More D&H leagues give us more chances to experiment with various strategies. More chances to be right. 2 2012 Baseball Forecaster / Simulators the biggest growth in baseball next year as well. They are so popular that we have created an overall championship for the format. There’s no doubt in my mind that this format is here to stay in the NFBC and in time it could attract more teams than our Main Event.” OnRoto.com is a commissioner service that can accommodate private leagues of this type. Limited FA D&H: It seems unfair to penalize teams that get hit with injuries. Limited FA D&H leagues allow weekly or monthly free-agent access for injury replacements only. Of course, the more access to free agents, the less valuable these leagues are as evaluation tools. Injuries occur all the time, so roster decisions will still be made using small data sets. My recommendation here is to combine this with the Deep D&H format, draft 40-man rosters to cover most injuries, then allow access to free agents on a monthly basis only. Monthly D&H: After the initial pre-season draft, this format conducts subsequent re-drafts at the beginning of each month. Then you would total the results of all six drafts. The goal here is to predict how players are going to do over the balance of the season. You have to assess the values of the hot and cold starters versus their potential. Each draft forces you to make a commitment to a group of players for five months, four months, and successively shorter time periods. This is a great way to develop your skills for regular transaction leagues. Solo D&H: You can even play D&H as a type of Rotisserie Solitaire. Using the Rotisserie 500 price list in the back of this book, assemble a legal 23-man roster for $260. Now do it again with 23 completely different players. Now do it again while shifting your budget $20 more to pitchers. Now shift another $20 to pitching. Now increase your roster to 40 players and your budget to $350. Designate 23 as active. Freeze all of these rosters and check back in October. Or make this a monthly exercise and use season to-date dollar values for every player. That’s an even greater challenge. While you can use this to test your own abilities, you can also grab a few friends and, working independently, compete against each other. I don’t suggest that D&H leagues replace our existing fantasy baseball experiences. However, I wouldn’t be surprised if some variant of this format becomes much more popular because of the reduced time commitment. At minimum, I do believe that playing D&H leagues will enhance our league experiences. They will allow us to sharpen the assessment and projection skills that lead to championships in our other leagues. And they are a challenging cerebral exercise in their own right. It’s time to put your tray tables and seat backs in their upright and locked position. Strap yourself into a fantasy flight simulator and draft a solo roster right now. Flip back to the Rotisserie 500 price list on page 269 and give it a go. Online services can run both snake drafts and auctions; they do the heavy lifting. Then we just need to approach our mocks with a bit more purpose. You might consider D&H leagues as the “flight simulators” for fantasy baseball success. You’d never get on a plane with a pilot who hadn’t first spent many hours in simulators. Similarly, you shouldn’t come to the draft table without having also practiced these important skills over many, many trials. D&H variations Of course, there is one major downside to drafting in the D&H format: Spectating for six months is boring. But there are variations springing up that make this experience more interesting, informative, challenging and fun. Pure D&H: Draft a 23-man roster and watch it for six months. Given that attrition typically takes over as injuries mount, Pure D&H is often referred to as “Last Man Standing.” This format is probably what you immediately think of when you hear “draft and hold.” And while it is the purest form of the concept, post-draft activity is about as exciting as watching grass grow. However, if you view it strictly as a means to practice your drafting and player evaluation skills, Pure D&H is a useful exercise. To that end, it’s probably best to draft several teams, trying out different roster construction strategies for each one. Experiment with going “stars & scrubs” or “spread the risk” or variants on budgeting; go heavy on pitchers; or play batting average-andspeed over power. While the player pool changes every year, the results of your efforts in one season can provide some insight into possible strategies for future years—not just in other D&H leagues, but in your money leagues. (A side benefit of this format is that you can easily export draft results from a website like MockDraftCentral.com and set up the teams in an Excel spreadsheet. Then you could just enter or import players’ current statistics as often as you want. There is little data manipulation so you don’t need a commissioner service to run these leagues.) Deep D&H: For owners who want a more interactive experience, Deep D&H drafts 40- or even 50-man rosters, with 23 active. Weekly moves of players from your reserve roster to your active roster and vice-versa are allowed, but there’s no access to free agents. The 40-50 you draft are the only players you get to play with all year. This format gives you the opportunity to gauge your skills in projecting more players. While in-season roster management will still affect the final standings, you won’t be tempted by the hot new names that hit the free agent wire each week. Deep D&H still requires you to do your homework up front, while giving you added flexibility. The National Fantasy Baseball Championship (NFBC) offers a Deep D&H game that is surging in popularity. It is a 15-team, 50-round serpentine “slow” draft, conducted online during the winter, where owners get eight hours to submit each pick. According to NFBC founder Greg Ambrosius, “These leagues showed the biggest growth this year and will undoubtedly show 3 2012 Baseball Forecaster / Simulators Welcome to Edition #26 These are spreadsheet data files that will be posted on or about March 1, 2012. Remember to keep the book handy when you visit as the access codes are hidden within these pages. Electronic book: The complete PDF version of the Forecaster— plus MS Excel versions of most key charts—is available free to those who bought the book directly through the BaseballHQ.com website. These files will be available in January 2012; contact us if you do not receive information via e-mail about accessing them. If you purchased the book through an online vendor or bookstore, or would like these files earlier, you can purchase them from us for $9.95. Call 1-800-422-7820 for more information. If you are new to the Forecaster, this book may seem overwhelming. There is a ton of information here. But you’re not going to be tested on this tomorrow, so take your time. It is advisable to attack things methodically. The best place to start is with the Encyclopedia of Fanalytics, which provides the foundation concepts for everything else that appears in these pages. Take a cursory read-through, stopping at any section that looks interesting. You’ll keep coming back here every so often. Then it’s okay to jump in. Close your eyes, flip to a random page, and put your finger down anywhere. Power develops late in catchers so Yadier Molina’s outburst could be for real? See, you’ve learned something already! Beyond the Forecaster The Baseball Forecaster is just the beginning. The following companion products and services are described in more detail in the back of the book. BaseballHQ.com is our *newly designed!* home website. It provides regular updates to everything in this book, including daily updated statistics and projections. A subscription to HQ gets you more than 1,000 articles over the course of the year, customized tools, access to data going back over a decade, plus a ton more. First Pitch Forums are a series of conferences we run all over the country, where you can meet some of the top industry analysts and network with fellow fantasy leaguers in your area. We’ll be in cities from coast to coast in February and March. Our big annual symposium at the Arizona Fall League is the first weekend in November. RotoHQ.com is a very, very large online library of fantasy strategy essays and tools. The 7th edition of the Minor League Baseball Analyst is by Rob Gordon and Jeremy Deloney. It is a minor league version of the Forecaster, with stat boxes for 1,000-plus prospects, and more. Available in January. We still have copies available of How to Value Players for Rotisserie Baseball, Art McGee’s ground-breaking book on valuation theory. They are still on closeout at 50% off. RotoLab is the best draft software on the market, period. We’re on Facebook! http://www.facebook.com/baseballhq. “Like” the Baseball HQ page for updates, photos from First Pitch events and links to other important stuff. I have my own personal page as well, but I reserve that for personal stuff so I do not accept friend requests from folks I don’t know personally. I will accept requests from those on LinkedIn, though. We’re also on Twitter. In fact, some of the BaseballHQ. com analysts are tweeting all over the freakin’ place. Address is @baseballhq. I think that’s it. Now go out there and find out something meaningful. And win your league. Because all this hoopla is about winning. Only winning. Always winning. What’s New in 2012? Expanded player boxes: After 18 years of squeezing nine players onto a page, then overflowing our borders by adding supplementary data sections, it was time to reassess. Thanks to the technical wizardry of programmer Rob Rosenfeld, we’ve centralized each player’s profile in newly expanded player boxes. These boxes now include things like complete power and speed profiles, DOM/ DIS consistency ratios, and much more. And as an added benefit of moving to seven players per page, we’ve opened up more real estate for player commentary as well. Expanded prospect section: Based on last spring’s reader survey results, we’ve doubled the number of scouting reports and ratings for upwardly mobile minor leaguers. If you need even more prospect information (about 900 players more), there’s still the Minor League Baseball Analyst. Universal draft grid: Here was our challenge… How do you provide draft prep information that has value given this book’s early publication date, and at the same time is flexible enough to be used for a variety of fantasy formats? The answer is obvious for subscribers to BaseballHQ.com. Online, our Rotisserie grids provide precise valuations that you can customize to your own league parameters. Here, we have modified the format so that it captures all the performance and scarcity needs of most competitions. What’s Old from 2011? Remember PQS Across America? Well, you wouldn’t believe what happened. See the back of the book. Updates Content Updates: If there are corrections or clarifications on the information in this book, you will find them here: http://www. baseballhq.com/books/bfupdates.shtml Free Projections Update: As a buyer of this book, you get one free 2012 projections update, available online at http://www.baseballhq.com/books/freeupdate/index.shtml 5 CONSUMER ADVISORY AN IMPORTANT MESSAGE FOR FANTASY LEAGUERS REGARDING PROPER USAGE OF THE BASEBALL FORECASTER This document is provided in compliance with authorities to outline the prospective risks and hazards possible in the event that the Baseball Forecaster is used incorrectly. Please be aware of these potentially dangerous situations and avoid them. The publisher assumes no risk UHODWHGWRDQ\ÀQDQFLDOORVVRUVWUHVVLQGXFHGLOOQHVVHV caused by ignoring the items as described below. 1. The statistical projections in this book are intended as general guidelines, not as gospel. It is highly dangerous to use the projected statistics alone, and then live and die by them. That’s like going to a ballgame, being given a choice of any seat in the park, and delibHUDWHO\FKRRVLQJWKHODVWURZLQWKHULJKWÀHOGFRUQHUZLWK an obstructed view. The projections are there, you can look at them, but there are so many better places to sit. We have to publish those numbers, but they are stagnant, inert pieces of data. This book focuses on a live forecasting process that provides the tools so that you can understand the leading indicators and draw your own conclusions. If you at least attempt your own analyses of the data, and enhance them with the player commentaries, you can paint more robust, colorful pictures of the future. In other words... If you bought this book purely for the projected statistics and do not intend to spend at least some time learning about the process, then you might as well just buy an $8 magazine. 2. The player commentaries in this book are written by humans, just like you. These commentaries provide an overall evaluation of performance and likely future direction, but 60-word capsules cannot capture everything. Your greatest value will be to use these as a springboard to your own analysis of the data. Odds are, LI\RXWDNHWKHWLPH\RXҋOOÀQGKLGGHQLQGLFDWRUVWKDWZH might have missed. Forecaster veterans say that this self-guided excursion is the best part of owning the book. 3. This book does not attempt to tackle playing time. Rather than making arbitrary decisions about how roles will shake out, the focus is on performance. The playing time projections presented here are merely to help you better evaluate each player’s talent. Our online preseason projections update provides more current AB and IP expectations based on how roles are being assigned. 4. The dollar values in this book are intended solely for player-to-player comparisons. They are not driven by DÀQLWHSRRORISOD\LQJWLPH³ZKLFKLVUHTXLUHGIRUYDOXDWLRQV\VWHPVWRZRUNSURSHUO\³VRWKH\FDQQRWEHXVHG for bid values to be used in your own draft. There are two reasons for this: D 7KH ÀQLWH SRRO RI SOD\HUV WKDW ZLOO JHQHUDWH WKH ÀQLWH SRRO RI SOD\LQJ WLPH ZLOO QRW EH GHWHUPLQHG XQWLO much closer to Opening Day. And, if we are to be brutally KRQHVW WKHUH LV UHDOO\ QR VXFK WKLQJ DV D ÀQLWH SRRO RI players. b. Your particular league’s construction will drive the values; a $10 player in a 10-team mixed league will not be the same as a $10 player in a 13-team NL-only league. Note that book dollar values also cannot be compared to those published at BaseballHQ.com DVWKHRQOLQHYDOXHVDUHJHQHUDWHGE\DPRUHÀQLWH player pool. 5. Do not pass judgment on the effectiveness of this book based on the performance of a few individual players. The test, rather, is on the collective predictive value of the book’s methods. Are players with better base skills more likely to produce good results than bad ones? Years of research suggest that the answer is “yes.” Does that mean that every high skilled player will do well? No. But many more of them will perform well than will the average low-skilled player. You should always side with the better percentage plays, but recognize that there are factors we cannot predict. Good decisions that beget bad outcomes do not invalidate the methods. 6. If your copy of this book is not marked up and dogeared by Draft Day, you probably did not get as much value out of it as you might have. 7. This edition of the Forecaster is not intended to provide absorbency for spills of more than 7.5 ounces. 8. This edition is not intended to provide stabilizing weight for more than 18 sheets of 20 lb. paper in winds of more than 45 mph. 9. The pages of this book are not recommended for avian waste collection. In independent laboratory studies, 87% of migratory water fowl refused to excrete on interior pages, even when coaxed. 10. This book, when rolled into a cylindrical shape, is not intended to be used as a weapon for any purpose, including but not limited to insect extermination, canine WUDLQLQJ RU WR LQÁXHQFH ELGGLQJ EHKDYLRU DW D IDQWDV\ draft. ENC YCLOP EDIA For new readers... Everything begins here. The information in the following pages represents the foundation that powers everything we do. You’ll learn about the underlying concepts for our unique mode of analysis. You’ll find answers to long-asked questions, interesting insights into what makes players tick, and innovative applications for all this newfound knowledge. This Encyclopedia is organized into several logical sections: 1. Fundamentals 2. Batters 3. Pitchers 4. Prospects 5. Gaming Enough talking. Jump in. Remember to breathe. For veteran readers… As we do in each edition, this year’s everexpanding Encyclopedia includes relevant research results we’ve published over the past year. We’ve added some of the essays from the Research Abstracts and Gaming Abstracts sections in the 2011 Forecaster as well as some other essays from BaseballHQ.com. And we continue to mold the content to best fit how fantasy leaguers use their information. Many readers consider this their fantasy information bible. Okay, time to jump-start the analytical process for 2012. Remember to breathe— it’s always good advice. OF F A N A LY T I C S Abbreviations APC BA BABIP bb% bb/9 BF/G BIP BPI BPV BPX Cmd Ct% Ctl DIS% Dom DOM% ERA Eye FAAB FB% G/L/F GB% H% h% HH% hr/9 hr/f IP/G k/9 LI LD% MLE OB OBA OPS PA PQR PQS Pw PX QC qERA R$ RAR RC RC/G REff% S% SB% SBO Slg Spd Sv% vL, vR WHIP xBA xERA xPX 7 Average pitch count per game Batting average . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Batting average on balls-in-play (also h%, H%) . . . . . . . . . . . . . . . 19, 26 Walk rate (hitters) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Opposition walks per 9 IP (also Ctl) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Batters faced per game . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Balls-in-play . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19, 26 Base performance indicator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Base performance value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22, 28 Base performance index (BPV indexed to league average) Command ratio (K/BB) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Contact rate (AB-K)/AB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Control ratio (also bb/9) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 PQS disaster rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Dominance ratio (also k/9) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 PQS domination rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Earned run average . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Batting eye (bb/k) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Free agent acquisition budget . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Fly ball percentage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Ground balls, line drives, and fly balls as a percentages of total balls in play (hits and outs) . . . . . . . . . . . . . . . . . . . . . . . 18, 26 Ground ball percentage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Hits allowed per balls in play (pitchers) . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Hits per balls in play (batters) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Hard hit balls as a pct. of total balls in play . . . . . . . . . . . . . . . . . . . . . . . 61 Opposition home runs per 9 IP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Home runs hit (batters), or allowed (pitchers), per fly ball . . . . 20, 27 Innings pitched per game appearance Dominance ratio (also Dom) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Leverage index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Line drive percentage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Major league equivalency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 On base average (batters) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Opposition batting average (pitchers) . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 On base plus slugging average. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Plate appearances Pure Quality Relief . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Pure Quality Starts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Linear weighted power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Linear weighted power index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Quality/Consistency Score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23, 32 PQS earned run average . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Rotisserie value (also 15$, 12$, etc. for specific league sizes) . . . . . . 45 Runs above replacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22, 28 Runs created . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Runs created per game . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Reliever efficiency percentage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Strand rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Stolen base success rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Stolen base opportunity percent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Slugging average . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Statistically Scouted Speed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Saves conversion rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 BA vs left-handers, vs right-handers . . . . . . . . . . . . . . . . . . . . . . . . . 22, 29 Walks plus hits divided by innings pitched . . . . . . . . . . . . . . . . . . . . . . 26 Expected batting average. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Expected earned run average . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Expected skills-based power index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Fundamentals Surface Stats: Traditional gauges that the mainstream media uses to measure performance. Stats like batting average, wins, and ERA only touch the surface of a player’s skill and often distort the truth. To uncover a player’s true skill, you have to look at component skills statistics. What is Fanalytics? Fanalytics is the scientific approach to fantasy baseball analysis. A contraction of “fantasy” and “analytics,” fanalytic gaming might be considered a mode of play that requires a more strategic and quantitative approach to player analysis and game decisions. The three key elements of fanalytics are: 1. Performance analysis 2. Performance forecasting 3. Gaming analysis For performance analysis, we tap into the vast knowledge of the sabermetric community. Founded by Bill James, this area of study provides objective and progressive new ways to assess skill. What we do in this book is called “component skills analysis.” We break down performance into its component parts, then reverseengineer it back into the traditional measures with which we are more familiar. Our forecasting methodology is one part science and one part art. We start with a computer-generated baseline for each player. We then make subjective adjustments based on a variety of factors, such as discrepancies in skills indicators and historical guidelines gleaned from more than 20 years of research. We don’t rely on a rigid model; our method forces us to get our hands dirty. You might say that our brand of forecasting is more about finding logical journeys than blind destinations. Gaming analysis is an integrated approach designed to help us win our fantasy leagues. It takes the knowledge gleaned from the first two elements and adds the strategic and tactical aspect of each specific fantasy game format. Component Skills Analysis Familiar gauges like HR and ERA have long been used to measure skill. In fact, these gauges only measure the outcome of an individual event, or series of events. They represent statistical output. They are “surface stats.” Raw skill is the talent beneath the stats, the individual elements of a player’s makeup. Players use these skills to create the individual events, or components, that we record using measures like HR and ERA. Our approach: 1. It’s not about batting average, it’s about seeing the ball and making contact. We target hitters based on elements such as their batting eye (walks to strikeouts ratio), how often they make contact and the type of contact they make. We then combine these components into an “expected batting average.” By comparing each hitter’s actual BA to how he should be performing, we can draw conclusions about the future. 2. It’s not about home runs, it’s about power. From the perspective of a round bat meeting a round ball, it may be only a fraction of an inch at the point of contact that makes the difference between a HR or a long foul ball. When a ball is hit safely, often it is only a few inches that separate a HR from a double. We tend to neglect these facts in our analyses, although the outcomes—the doubles, triples, long fly balls—may be no less a measure of that batter’s raw power skill. We must incorporate all these components to paint a complete picture. 3. It’s not about ERA, it’s about getting the ball over the plate and keeping it in the park. Forget ERA. You want to draft pitchers who walk few batters (control), strike out many (dominance) and succeed at both in tandem (command). You also want pitchers who keep the ball on the ground (because home runs are bad). All of this translates into an “expected ERA” that you can use to compare to a pitcher’s actual performance. 4. It’s never about wins. For pitchers, winning ballgames is less about skill than it is about offensive support. As such, projecting wins is a very high-risk exercise and valuing hurlers based on their win history is dangerous. Target skill; wins will come. 5. It’s not about saves, it’s about opportunity first and skills second. While the highest skilled pitchers have the best potential to succeed as closers, they still have to be given the ball with the game on the line in the 9th inning, and that is a decision left to others. Over the past 10 years, about 40% of relievers drafted for saves failed to hold the role for the entire season. The lesson: Don’t take chances on draft day. There will always be saves in the free agent pool. Definitions Base Performance Indicator (BPI): A statistical formula that measures an isolated aspect of a player’s situation-independent raw skill or a gauge that helps capture the effects that random chance has on skill. Leading Indicator: A statistical formula that can be used to project potential future performance. Noise: Irrelevant or meaningless pieces of information that can distort the results of an analysis. In news, this is opinion or rumor that can invalidate valuable information. In forecasting, these are unimportant elements of statistical data that can artificially inflate or depress a set of numbers. Situation Independent: Describing performance that is separate from the context of team, ballpark, or other outside variables. Strikeouts and walks, as they are unaffected by the performance of a batter’s team, are often considered situation independent stats. Conversely, RBIs are situation dependent because individual performance varies greatly by the performance of other batters on the team (you can’t drive in runs if there is nobody on base). Situation independent gauges are important for us to be able to isolate and judge performance on its own merits. Soft Skills: BPIs with levels below established minimums for acceptable performance. Accounting for “luck” Luck has been used as a catch-all term to describe random chance. When we use the term here, we’re talking about unexplained variances that shape the statistics. While these variances may be random, they are also often measurable and projectable. 8 2012 Baseball Forecaster / Encyclopedia of Fanalytics To get a better read on “luck,” we use formulas that capture the external variability of the data. Through our research and the work of others, we have learned that when raw skill is separated from statistical output, what’s remaining is often unexplained variance. The aggregate totals of many of these variances, for all players, is often a constant. For instance, while a pitcher’s ERA might fluctuate, the rate at which his opposition’s batted balls fall for hits will tend towards 30%. Large variances can be expected to regress towards 30%. Why is all this important? Analysts complain about the lack of predictability of many traditional statistical gauges. The reason they find it difficult is that they are trying to project performance using gauges that are loaded with external noise. Raw skills gauges are more pure and follow better defined trends during a player’s career. Then, as we get a better handle on the variances— explained and unexplained—we can construct a complete picture of what a player’s statistics really mean. means all the advanced systems are fighting for occupation of the remaining 5%. But there is a bigger question… what exactly are we measuring? When we search for accuracy, what does that mean? In fact, any quest for accuracy is going to run into a brick wall of paradoxes… t *G B TMVHHJOH BWFSBHF QSPKFDUJPO JT EFBE PO CVU UIF player hits 10 fewer HRs than expected (and likely, 20 more doubles), is that a success or a failure? t *GBQSPKFDUJPOPGIJUTBOEXBMLTBMMPXFECZBQJUDIFSJT on the mark, but the bullpen and defense implodes, and inflates his ERA by a run, is that a success or a failure? t *G UIF QSPKFDUJPO PG B TQFFETUFST SBUF PG TUPMFO CBTF success is perfect, but his team replaces the manager with one that doesn’t run, and the player ends up with half as many SBs as expected, is that a success or a failure? t *GBCBUUFSJTUSBEFEUPBIJUUFSTCBMMQBSLBOEBMMUIFUPVUT project an increase in production, but he posts a statistical line exactly what would have been projected had he not been traded to that park, is that a success or a failure? t *GUIFQSPKFDUJPOGPSBCVMMQFODMPTFST&3"8)*1BOE peripheral numbers is perfect, but he saves 20 games instead of 40 because the GM decided to bring in a high-priced free agent at the trading deadline, is that a success or a failure? t *GBQMBZFSJTQSPKFDUFEUPIJUJO"#BOEPOMZIJUT .249, is that a success or failure? Most will say “failure.” But wait a minute! The real difference is only two hits per month. That shortfall of 23 points in batting average is because a fielder might have made a spectacular play, or a screaming liner might have been hit right at someone, or a long shot to the outfield might have been held up by the wind... once every 14 games. Does that constitute “failure”? Even if we were to isolate a single statistic that measures “overall performance” and run our accuracy tests on it, the results will still be inconclusive. According to OPS, these players are virtually identical: Baseball Forecasting Forecasting in perspective Forecasts. Projections. Predictions. Prognostications. The crystal ball aura of this process conceals the fact it is a process. We might define it as “the systematic process of determining likely end results.” At its core, it’s scientific. However, the outcomes of forecasted events are what is most closely scrutinized, and are used to judge the success or failure of the forecast. That said, as long as the process is sound, the forecast has done the best job it can do. In the end, forecasting is about analysis, not prophecy. Baseball performance forecasting is inherently a high-risk exercise with a very modest accuracy rate. This is because the process involves not only statistics, but also unscientific elements, from random chance to human volatility. And even from within the statistical aspect there are multiple elements that need to be evaluated, from skill to playing time to a host of external variables. Every system is comprised of the same core elements: t 1MBZFST XJMM UFOE UP QFSGPSN XJUIJO UIF GSBNFXPSL PG past history and/or trends. t 4LJMMTXJMMEFWFMPQBOEEFDMJOFBDDPSEJOHUPBHF t 4UBUJTUJDT XJMM CF TIBQFE CZ B QMBZFST IFBMUI FYQFDUFE role and venue. While all systems are built from these same elements, they also are constrained by the same limitations. We are all still trying to project a bunch of human beings, each one... t XJUIIJTPXOJOEJWJEVBMTLJMMTFU t XJUIIJTPXOSBUFPGHSPXUIBOEEFDMJOF t XJUIIJTPXOBCJMJUZUPSFTJTUBOESFDPWFSGSPNJOKVSZ t MJNJUFEUPPQQPSUVOJUJFTEFUFSNJOFECZPUIFSQFPQMF t HFOFSBUJOH B HSPVQ PG TUBUJTUJDT MBSHFMZ BČFDUFE CZ external noise. Based on the research of multiple sources, the best accuracy rate that can be attained by any system is about 70%. In fact, a simple system that uses three-year averages adjusted for age (“Marcel the Monkey”) can attain a success rate of 65%. This %$77(5 .LQVOHU, -R\FH0 %XWOHU% =REULVW% +5 5%, 6% %$ 2%$ 6/* 236 If I were to project Ian Kinsler-caliber stats and ended up with Billy Butler numbers, I’d hardly call that an accurate projection, especially if my fantasy team was in dire need of home runs and stolen bases. According to Roto dollars, these players are also dead-on: %$77(5 8SWRQ%- %RQLIDFLR( 7HL[HLUD0 %HOWUDQ& +5 5%, 5XQV 6% %$ 5 It’s not so simple for someone to claim they have accurate projections. And so, it is best to focus on the bigger picture, especially when it comes to winning at fantasy baseball. More on this: “The Great Myths of Projective Accuracy” KWWSZZZEDVHEDOOKTFRPERRNVP\WKVVKWPO 9 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Baseball Forecaster’s forecasting process is healthy enough to play. These words from experts have some element of truth, but cannot be wholly relied upon to provide an accurate expectation of future events. As such, it is often helpful to develop an appropriate cynicism for what you read. For instance, if a player is struggling for no apparent reason and there are denials about health issues, don’t dismiss the possibility that an injury does exist. There are often motives for such news to be withheld from the public. And so, as long as we do not know all the facts, we cannot dismiss the possibility that any one fact is true, no matter how often the media assures it, deplores it, or ignores it. Don’t believe everything you read; use your own judgment. If your observations conflict with what is being reported, that’s powerful insight that should not be ignored. Also remember that nothing lasts forever in major league baseball. Reality is fluid. One decision begets a series of events that lead to other decisions. Any reported action can easily be reversed based on subsequent events. My favorite examples are announcements of a team’s new bullpen closer. Those are about the shortest realities known to man. We need the media to provide us with context for our analyses, and the real news they provide is valuable intelligence. But separating the news from the noise is difficult. In most cases, the only thing you can trust is how that player actually performs. We are all about component skills. Our approach is to assemble these evaluators in such a way that they can be used to validate our observations, analyze their relevance and project a likely future direction. In a perfect world, if a player’s raw skills improve, then so should his surface stats. If his skills decline, then his stats should follow as well. But, sometimes a player’s skill indicators increase while his surface stats decline. These variances may be due to a variety of factors. Our forecasting process is based on the expectation that events tend to move towards universal order. Surface stats will eventually approach their skill levels. Unexplained variances will regress to a mean. And from this, we can identify players whose performance may potentially change. For most of us, this process begins with the previous year’s numbers. Last season provides us with a point of reference, so it’s a natural way to begin the process of looking at the future. Component skills analysis allows us to validate those numbers. A batter with few HRs but a high linear weighted power level has a good probability of improving his future HR output. A pitcher whose ERA was poor while his command ratio was solid is a good bet for ERA improvement. Of course, these leading indicators do not always follow the rules. There are more shades of grey than blacks and whites. When indicators are in conflict—for instance, a pitcher who is displaying both a rising strikeout rate and a rising walk rate— then we have to find ways to sort out what these indicators might be saying. It is often helpful to look at leading indicators in a hierarchy, of sorts. In fact, a hierarchy of the most important pitching base performance indicators might look like this: command (k/bb), dominance (k/9), control (bb/9) and GB/FB rate. For batters, contact rate might top the list, followed by power, walk rate and speed. Embracing imprecision Precision in baseball prognosticating is a fool’s quest. There are far too many unexpected variables and noise that can render our projections useless. The truth is, the best we can ever hope for is to accurately forecast general tendencies and percentage plays. However, even when you follow an 80% percentage play, for instance, you will still lose 20% of the time. That 20% is what skeptics use as justification to dismiss prognosticators; they conveniently ignore the more prevalent 80%. The paradox, of course, is that fantasy league titles are often won or lost by those exceptions. Still, long-term success dictates that you always chase the 80% and accept the fact that you will be wrong 20% of the time. Or, whatever that percentage play happens to be. For fantasy purposes, playing the percentages can take on an even less precise spin. The best projections are often the ones that are just far enough away from the field of expectation to alter decision-making. In other words, it doesn’t matter if I project Player X to bat .320 and he only bats .295; it matters that I project .320 and everyone else projects .280. Those who follow my less-accurate projection will go the extra dollar to acquire him in their draft. Or, perhaps we should evaluate projections based upon their intrinsic value. For instance, coming into 2011, would it have been more important for me to tell you that Albert Pujols was going to hit 40 HRs or that Jacoby Ellsbury would hit 20 HRs? By season’s end, the Pujols projection would have been more accurate, but the Ellsbury projection—even though it was off by 12 HR—would have been far more valuable. And that has to be enough. Any tout who projects a player’s statistics dead-on will have just been lucky with his dart throws that day. Assimilating additional research Once we’ve painted the statistical picture of a player’s potential, we then use additional criteria and research results to help us add some color to the analysis. These other criteria include the player’s health, age, changes in role, ballpark and a variety of other factors. We also use the research results described in the following pages. This research looks at things like traditional periods of peak performance and breakout profiles. The final element of the process is assimilating the news into the forecast. This is the element that many fantasy leaguers tend to rely on most since it is the most accessible. However, it is also the element that provides the most noise. Players, management and the media have absolute control over what we are allowed to know. Factors such as hidden injuries, messy divorces and clubhouse unrest are routinely kept from us, while we are fed red herrings and media spam. We will never know the entire truth. Quite often, all you are reading is just other people’s opinions... a manager who believes that a player has what it takes to be a regular or a team physician whose diagnosis is that a player 10 2012 Baseball Forecaster / Encyclopedia of Fanalytics Perpetuity Melky Cabrera, Mark Trumbo, Daniel Murphy, Gerardo Parra, Matt Joyce, Ryan Roberts, Cory Luebke, Emilio Bonifacio. Unexpected return to form: These players had displayed solid skill at some point in the past. But that skill had been M.I.A. long enough that we began to doubt that I would ever return; our projections model got sick of waiting. Or previous skill displays were so inconsistent that projecting an “up year” would have been a shot in the dark; our projections model got sick of guessing. Yes, “once you display a skill, you own it” but still... David Ortiz, Todd Helton, Jeff Francoeur, Lance Berkman, Jhonny Peralta, Francisco Cordero, James Shields, Ervin Santana. Unexpected role: Every year, there are relief pitchers who are on nobody’s radar to become a frontline closer but are suddenly thrust into the role with great success... Fernando Salas, Craig Kimbrel, Jordan Walden, Sergio Santos, Drew Storen. The unexplained: How these players put up the numbers they did is a mystery, but the odds of a comparable follow-up—particularly those with soft peripherals—will be miniscule: Asdrubal Cabrera’s power, Mike Napoli’s batting average, Doug Fister’s Motor City surge and Ryan Vogelsong’s roster spot. Forecasting is not an exercise that produces a single set of numbers. It is dynamic, cyclical and ongoing. Conditions are constantly changing and we must react to those changes by adjusting our expectations. A pre-season projection is just a snapshot in time. Once the first batter steps to the plate on Opening Day, that projection has become obsolete. Its value is merely to provide a starting point, a baseline for what is about to occur. During the season, if a projection appears to have been invalidated by current performance, the process continues. It is then that we need to ask... What went wrong? What conditions have changed? In fact, has anything changed? We need to analyze the situation and revise our expectation, if necessary. This process must be ongoing. When good projections go bad We cannot predict the future; all we can do is provide a sound process for constructing a “most likely expectation for future performance.” All we can control is the process. As such, there is a limit to how much blame we can shoulder for each year’s misses. If we’ve captured as much information as is available, used the best methodology and analyzed the results correctly, that’s about the best we can do. We simply can’t control outcomes. What we can do is analyze all the misses to see why they occurred. This is always a valuable exercise each year. It puts a proper focus on the variables that were out of our control as well as providing perspective on those players that we might have done a better job with. In general, we can organize our forecasting misses into several categories. To demonstrate, here are the players whose 2011 Rotisserie earnings varied from projection by at least $10: The performances that fell short of expectation The DL denizens: These are players who got hurt, may not have returned fully healthy, or may have never been fully healthy (whether they’d admit it or not)... Joe Mauer, Shin Soo-Choo, Justin Morneau, Alex Rodriguez, Buster Posey, Brian Roberts, Brett Anderson, Stephen Drew, Andrew Bailey, Francisco Liriano, Clay Buchholz, Vernon Wells, Denard Span, Casey McGehee (sprained thumb at end of April might have wiped out 2-plus months before he started coming around), Tsuyoshi Nishioka, David Wright, Matt Holliday, Rafael Furcal, Jason Heyward, Kevin Youkilis, Ryan Zimmerman, Rajai Davis, Hanley Ramirez, Carl Crawford, Hong-Chih Kuo, Tommy Hanson, Josh Johnson, Roy Oswalt. Note that some of these players seemed to be putting up sub-par numbers before they actually hit the DL. Many were likely playing through the hurt before breaking down. Accelerated skills erosion: These are players who we knew were on the downside of their careers but who we did not think would plummet so quickly. In some cases, there were injuries involved, but all in all, 2011 was probably the beginning of the end... Miguel Tejada, Magglio Ordonez. Marginal players who just had bad years: We knew these guys were not that good; we just didn’t anticipate they’d be this bad... Trevor Cahill, Bronson Arroyo. Inflated expectations: Here are players who we really should not have expected much more than what they produced. Some had short or spotty track records, others had soft peripherals coming into 2011, and still others were inflated by media hype... Omar Infante, Travis Snider, Aubrey Huff, Delmon Young, Jayson Werth, Brett Myers, Matt Latos, Pedro Alvarez. Unexpected loss of role: These are the relief pitchers who were drafted for saves but were traded, got hurt or were moved out of the role after a few bad outings... Francisco Rodriguez, Matt Thornton, Ryan Franklin, Brandon Lyon, David Aardsma, The performances that exceeded expectation Development beyond the growth trend: These are young players for whom we knew there was skill. Some were prized prospects in the past who took their time ascending the growth curve. Others were a surprise only because their performance spike arrived sooner than anticipated... Alex Gordon, Justin Upton, Alex Avila, Cameron Maybin, Starlin Castro, Ricky Romero, Ian Kennedy, Mike Morse, Brennan Boesch, Elvis Andrus, Justin Masterson. Skilled players who just had big years: We knew these guys were good too; we just didn’t anticipate they’d be this good... Justin Verlander, Matt Kemp, Jose Bautista, Michael Bourn, Curtis Granderson, Jacoby Ellsbury, Clayton Kershaw, Erick Aybar, Tyler Clippard, C.J. Wilson. Unexpected health: We knew these players had the goods; we just didn’t think they’d stay healthy... Carlos Beltran, Coco Crisp, Josh Beckett, Kyle Lohse. Unexpected playing time: These players had the skills—and may have even displayed them at some time in the past—but had questionable playing time potential coming into this season. Some benefited from another player’s injury, a rookie who didn’t pan out or leveraged a short streak into a regular gig... Eric Hosmer, 11 Encyclopedia of Fanalytics / 2012 Baseball Forecaster 4. Historical trends: Have his skill levels over time been on an upswing or downswing? 5. Component skills indicators: Looking beyond batting averages and ERAs, what do his support ratios look like? 6. Ballpark, team, league: Pitchers going to Texas will see their ERA spike. Pitchers going to PETCO Park will see their ERA improve. 7. Team performance: Has a player’s performance been affected by overall team chemistry or the environment fostered by a winning or losing club? 8. Batting stance, pitching style: Has a change in performance been due to a mechanical adjustment? 9. Usage pattern, lineup position, role: Has a change in RBI opportunities been a result of moving further up or down in the batting order? Has pitching effectiveness been impacted by moving from the bullpen to the rotation? 10. Coaching effects: Has the coaching staff changed the way a player approaches his conditioning, or how he approaches the game itself? 11. Off-season activity: Has the player spent the winter frequenting workout rooms or banquet tables? 12. Personal factors: Has the player undergone a family crisis? Experienced spiritual rebirth? Given up red meat? Taken up testosterone? Jonathan Broxton, Brad Lidge, Kevin Gregg, Frank Francisco, Fernando Rodney and Matt Capps. The unexplained: These are the players for whom we have no rational explanation for what happened. We can speculate that they hid an injury, went off of PEDs, or just didn’t have their head on right in 2011: Chone Figgins, Martin Prado, Evan Longoria, Alex Rios, John Danks, Chris Carpenter, Ubaldo Jimenez, Chad Billingsley and Adam Dunn. For some, there are insights to be found in the BPIs and you can read about them later on. For others, the only explanation we can find starts with alien abduction. About fantasy baseball touts As a group, there is a strong tendency for all pundits to provide numbers that are publicly palatable, often at the expense of realism. That’s because committing to either end of the range of expectation poses a high risk. Few touts will put their credibility on the line like that, even though we all know that those outliers are inevitable. Among our projections, you will find few .350 hitters and 70-steal speedsters. Someone is going to post a 2.25 ERA next year, but damned if any of us will commit to that. So we take an easier road. We’ll hedge our numbers or split the difference between two equally possible outcomes. In the world of prognosticating, this is called the comfort zone. This represents the outer tolerances for the public acceptability of a set of numbers. In most circumstances, even if the evidence is outstanding, prognosticators will not stray from within the comfort zone. As for this book, occasionally we do commit to outlying numbers when we feel the data support it. But on the whole, most of the numbers here can be nearly as cowardly as everyone else’s. We get around this by providing “color” to the projections in the capsule commentaries. That is where you will find the players whose projection has the best potential to stray beyond the limits of the comfort zone. As analyst John Burnson once wrote: “The issue is not the success rate for one player, but the success rate for all players. No system is 100% reliable, and in trying to capture the outliers, you weaken the middle and thereby lose more predictive pull than you gain. At some level, everyone is an exception!” Skills ownership Once a player displays a skill, he owns it. That display could occur at any time—earlier in his career, back in the minors, or even in winter ball play. And while that skill may lie dormant after its initial display, the potential is always there for him to tap back into that skill at some point, barring injury or age. That dormant skill can reappear at any time given the right set of circumstances. Caveats: 1. The initial display of skill must have occurred over an extended period of time. An isolated 1-hit shut-out in Single-A ball amidst a 5.00 ERA season is not enough. The shorter the display of skill in the past, the more likely it can be attributed to random chance. The longer the display, the more likely that any re-emergence is for real. 2. If a player has been suspected of using performance enhancing drugs at any time, all bets are off. Corollaries: 1. Once a player displays a vulnerability or skills deficiency, he owns that as well. That vulnerability could be an old injury problem, an inability to hit breaking pitches, or just a tendency to go into prolonged slumps. 2. The probability of a player addressing and correcting a skills deficiency declines with each year he allows that deficiency to continue to exist. Validating Performance Performance validation criteria The following is a set of support variables that helps determine whether a player’s statistical output is an accurate reflection of his skills. From this we can validate or refute stats that vary from expectation, essentially asking, is this performance “fact or fluke?” 1. Age: Is the player at the stage of development when we might expect a change in performance? 2. Health: Is he coming off an injury, reconditioned and healthy for the first time in years, or a habitual resident of the disabled list? 3. Minor league performance: Has he shown the potential for greater things at some level of the minors? Or does his minor league history show a poor skill set that might indicate a lower ceiling? Contract year performance (Tom Mullooly) There is a contention that players step up their game when they are playing for a contract. Research looked at contract year players and their performance during that year as compared to career levels. Of the batters and pitchers studied, 53% of the batters 12 2012 Baseball Forecaster / Encyclopedia of Fanalytics performed as if they were on a salary drive, while only 15% of the pitchers exhibited some level of contract year behavior. How do players fare after signing a large contract (minimum $4 million per year)? Research from 2005-2008 revealed that only 30% of pitchers and 22% of hitters exhibited an increase of more than 15% in base performance values after signing a large deal either with their new team, or re-signing with the previous team. But nearly half of the pitchers (49%) and nearly half of the hitters (47%) saw a drop in BPV of more than 15% in the year after signing. downward trend cannot be considered truly consistent as we do not know whether those trends will continue. Typically, they don’t. “A” level players are those whose runs created per game level (xERA for pitchers) has fluctuated by less than half a run during each of the past three years. “F” grades go to those whose RC/G or xERA has fluctuated by two runs or more. Remember that these grades have nothing to do with quality of performance; they strictly refer to confidence in our expectations. So a grade of AAA for Joe Saunders, for instance, only means that there is a high probability he will perform as poorly as we’ve projected. Risk management and reliability grades Forecasts are constructed with the best data available, but there are factors that can impact the variability around that projection. One way we manage this risk is to assign each player Reliability Grades. The more certainty we see in a data set, the higher the reliability grades assigned to that player. The following variables are evaluated: Health: Players with a history of staying healthy and off the disabled list are valuable to own. Unfortunately, while the ability to stay healthy can be considered skill, it is not very projectable. We can track the number of days spent on the disabled list and draw only rough conclusions. The grades in the player boxes also include an adjustment for older players, who have a higher likelihood of getting hurt. That is the only forward-looking element of the grade. “A” level players would have accumulated fewer than 30 days on the major league DL over the past five years. “F” grades go to those who’ve spent more than 120 days on the DL. Recent DL stays are given a heavier weight in the calculation. Playing Time and Experience (PT/Exp): The greater the pool of major league history to draw from, the greater our ability to construct a viable forecast. Length of service is important, as is length of consistent service. So players who bounce up and down from the majors to the minors are higher risk players. And rookies are all high risk. For batters, we simply track plate appearances. Major league PAs have greater weight than minor league PAs. “A” level players would have averaged at least 550 major league PAs per year over the past three years. “F” graded players averaged fewer than 250 major league PA per year. For pitchers, workload can be a double-edged sword. On one hand, small IP samples are deceptive in providing a read on a pitcher’s true potential. Even a consistent 65-inning reliever can be considered higher risk since it would take just one bad outing to skew an entire season’s work. On the flipside, high workload levels also need to be monitored, especially in the formative years of a pitcher’s career. Exceeding those levels elevates the risk of injury, burnout, or breakdown. So, tracking workload must be done within a range of innings. The grades capture this. Consistency: Consistent performers are easier to project and garner higher reliability grades. Players that mix mediocrity with occasional flashes of brilliance or badness generate higher risk projections. Even those who exhibit a consistent upward or Reliability and age Peak batting reliability occurs at ages 29 and 30, followed by a minor decline for four years. So, to draft the most reliable batters, and maximize the odds of returning at least par value on your investments, you should target the age range of 28-34. The most reliable age range for pitchers is 29-34. While we are forever looking for “sleepers” and hot prospects, it is very risky to draft any pitcher under 27 or over 35. Health Analysis Disabled list statistics <HDU 3OD\HUV \U$YJ '/'D\V \U$YJ D.L. days as a leading indicator (Bill Macey) Players who are injured in one year are likely to be injured in a subsequent year: '/EDWWHUVLQ<HDUZKRDUHDOVR'/LQ\HDU 8QGHUDJH $JHDQGROGHU '/EDWWHUVLQ<HDUDQGZKRDUHDOVR'/LQ\HDU '/SLWFKHUVLQ<HDUZKRDUHDOVR'/LQ\HDU 8QGHUDJH $JHDQGROGHU '/SLWFKHUVLQ<UDQGZKRDUHDOVR'/LQ\HDU Previously injured players also tend to spend a longer time on the DL. The average number of days on the DL was 51 days for batters and 73 days for pitchers. For the subset of these players who get hurt again the following year, the average number of days on the DL was 58 days for batters and 88 days for pitchers. 13 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Spring training spin (Dave Adler) During this period, it is difficult to get a true read on how a player is going to ultimately perform. He is adjusting to the new situation. Things could be volatile during this time. For instance, a role change that doesn’t work could spur other moves. A rookie hurler might buy himself a few extra starts with a solid debut, even if he has questionable skills. It is best not to make a decision on a player who is going through a courtship period. Wait until his stats stabilize. Don’t cut a struggling pitcher in his first few starts after a managerial change. Don’t pick up a hitter who smacks a pair of HRs in his first game after having been traded. Unless, of course, talent and track record say otherwise. Spring training sound bites raise expectations among fantasy leaguers, but how much of that “news” is really “noise”? Thanks to a summary listed at RotoAuthority.com, we were able to compile the stats for 2009. Verdict: Noise. %$77(56 :HLJKWFKDQJH )LWQHVVSURJUDP (\HVXUJHU\ 3ODQVPRUH6% 1R ,03529(' '(&/,1(' 3,7&+(56 :HLJKWFKDQJH )LWQHVVSURJUDP (\HVXUJHU\ 1HZSLWFK 1R ,03529(' '(&/,1(' Half-season fallacies A popular exercise at the midpoint of each season is to analyze those players who are consistent first half to second half surgers or faders. There are several fallacies with this analytical approach. 1. Half-season consistency is rare. There are very few players who show consistent changes in performance from one half of the season to the other. Research results from a three-year study conducted in the late1990s: The test groups... batters with min. 300 AB full season, 150 AB first half, and pitchers with min. 100 IP full season, 50 IP first half. Of those groups (size noted): In-Season Analysis April performance as a leading indicator We isolated all players who earned at least $10 more or $10 less than we had projected in March. Then we looked at the April stats of these players to see if we could have picked out the $10 outliers after just one month. ,GHQWLILDEOHLQ$SULO (DUQHGPRUHWKDQSURMHFWHG %$77(56 3,7&+(56 (DUQHGOHVVWKDQSURMHFWHG %$77(56 3,7&+(56 \HDUFRQVLVWHQF\LQ VWDWFDWHJRU\ VWDWFDWHJRULHV VWDWFDWHJRULHV 3,7&+(56 When the analysis was stretched to a fourth year, only 1% of all players showed consistency in even one category. 2. Analysts often use false indicators. Situational statistics provide us with tools that can be misused. Several sources offer up 3- and 5-year stats intended to paint a picture of a long-term performance. Some analysts look at a player’s half-season swing over that multi-year period and conclude that he is demonstrating consistent performance. The fallacy is that those multi-year scans may not show any consistency at all. They are not individual season performances but aggregate performances. A player whose 5-year batting average shows a 15-point rise in the 2nd half, for instance, may actually have experienced a BA decline in several of those years, a fact that might have been offset by a huge BA rise in one of the years. 3. It’s arbitrary. The season’s midpoint is an arbitrary delineator of performance swings. Some players are slow starters and might be more appropriately evaluated as pre-May 1 and postMay 1. Others bring their game up a notch with a pennant chase and might see a performance swing with August 15 as the cut-off. Each player has his own individual tendency, if, in fact, one exists at all. There’s nothing magical about mid-season as the break point, and certainly not over a multi-year period. Nearly three out of every four pitchers who earned at least $10 less than projected also struggled in April. For all the other surprises—batters or pitchers—April was not a strong leading indicator. Another look: %DWWHUVZKRILQLVKHG 3LWFKHUVZKRILQLVKHG %DWWHUVZKRILQLVKHGXQGHU 3LWFKHUVZKRILQLVKHGXQGHU %$77(56 3FW April surgers are less than a 50/50 proposition to maintain that level all season. Those who finished April at the bottom of the roto rankings were more likely to continue struggling, especially pitchers. In fact, of those pitchers who finished April with a value under -$10, 91% finished the season in the red. Holes are tough to dig out of. Courtship period Any time a player is put into a new situation, he enters into what we might call a courtship period. This period might occur when a player switches leagues, or switches teams. It could be the first few games when a minor leaguer is called up. It could occur when a reliever moves into the rotation, or when a lead-off hitter is moved to another spot in the lineup. There is a team-wide courtship period when a manager is replaced. Any external situation that could affect a player’s performance sets off a new decision point in evaluating that performance. Half-season tendencies Despite the above, it stands to reason logically that there might be some underlying tendencies on a more global scale, first half to second half. In fact, one would think that the player population 14 2012 Baseball Forecaster / Encyclopedia of Fanalytics Plexiglass Principle (Bill James): If a player or team improves markedly in one season, it will likely decline in the next. The opposite is true but not as often (because a poor performer gets fewer opportunities to rebound). Whirlpool Principle (Bill James): All team and player performances are forcefully drawn to the center. For teams, that center is a .500 record. For players, it represents their career average level of performance. as a whole might decline in performance as the season drones on. There are many variables that might contribute to a player wearing down—workload, weather, boredom—and the longer a player is on the field, the higher the likelihood that he is going to get hurt. A recent 5-year study uncovered the following tendencies: Batting Overall, batting skills held up pretty well, half to half. There was a 5% erosion of playing time, likely due, in part, to September roster expansion.. Power: First half power studs (20 HRs in 1H) saw a 10% dropoff in the second half. 34% of first half 20+ HR hitters hit 15 or fewer in the second half and only 27% were able to improve on their first half output. Speed: Second half speed waned as well. About 26% of the 20+ SB speedsters stole at least 10 fewer bases in the second half. Only 26% increased their second half SB output at all. Batting average: 60% of first half .300 hitters failed to hit .300 in the second half. Only 20% showed any second half improvement at all. As for 1H strugglers, managers tended to stick with their full-timers despite poor starts. Nearly one in five of the sub-.250 1H hitters managed to hit more than .300 in the second half. Pitching Overall, there was some slight erosion in innings and ERA despite marginal improvement in some peripherals. ERA: For those who pitched at least 100 innings in the first half, ERAs rose an average of 0.40 runs in the 2H. Of those with first half ERAs less than 4.00, only 49% were able to maintain a sub-4.00 ERA in the second half. Wins: Pitchers who won 18 or more games in a season tended to pitch more innings in the 2H and had slightly better peripherals. Saves: Of those closers who saved 20 or more games in the first half, only 39% were able to post 20 or more saves in the 2H, and 26% posted fewer than 15 saves. Aggregate ERAs of these pitchers rose from 2.45 to 3.17, half to half. Trending into trouble Other Diamonds The Fanalytic Fundamentals 1. This is not a game of accuracy or precision. It is a game of human beings and tendencies. 2. This is not a game of projections. It is a game of market value versus real value. 3. Draft skills, not stats. Draft skills, not roles. 4. A player’s ability to post acceptable stats despite lousy BPIs will eventually run out. 5. Once you display a skill, you own it. 6. Virtually every player is vulnerable to a month of aberrant performance. Or a year. 7. Exercise excruciating patience. Aging Axioms 1. Age is the only variable for which we can project a rising trend with 100% accuracy. (Or, age never regresses.) 2. The aging process slows down for those who maintain a firm grasp on the strike zone. Plate patience and pitching command can preserve any waning skill they have left. 3. Negatives tend to snowball as you age. Steve Avery List Players who hang onto MLB rosters for six years searching for a skill level they only had for three. Bylaws of Badness 1. Some players are better than an open roster spot, but not by much. 2. Some players have bad years because they are unlucky. Others have many bad years because they are bad... and lucky. (Ed DeCaria) It is almost irresistibly tempting to look at three numbers moving in one direction and expect that the fourth will continue that progression. While such pattern recognition serves us well in other endeavors, it can lead to some dreadful decision-making in the world of fantasy baseball. In this 2003-2009 study, we see a forceful trend disruption in all metrics. Separating players by age, those 32 and younger hold their gains slightly better than those 33 and over, but the numbers do not fundamentally differ from the above. George Brett Path to Retirement Get out while you’re still putting up good numbers and the public perception of you is favorable. Like Mike Mussina. And Billy Wagner. (See Steve Carlton Path to Retirement.) Steve Carlton Path to Retirement Hang around the major leagues long enough for your numbers to become so wretched that people begin to forget your past successes. (See George Brett Path to Retirement.) Among the many players who have taken this path include Jose Mesa, Doc Gooden, Matt Morris, Nomar Garciaparra and of course, Steve Carlton. Current players who look to be on the same course include Jorge Posada, Ivan Rodriguez, Mike Cameron and Miguel Tejada. Teams Johnson Effect (Bryan Johnson): Teams whose actual won/loss record exceeds or falls short of their statistically projected record in one season will tend to revert to the level of their projection in the following season. Law of Competitive Balance (Bill James): The level at which a team (or player) will address its problems is inversely related to its current level of success. Low performers will tend to make changes to improve; high performers will not. This law explains the existence of the Plexiglass and Whirlpool Principles. 15 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Christie Brinkley Law of Statistical Analysis Never get married to the model. 5. Players with back problems are always worth $10 less. 6. “Opting out of surgery” usually means it’s coming anyway, just later. Employment Standards 1. If you are right-brain dominant, own a catcher’s mitt and are under 40, you will always be gainfully employed. 2. Some teams believe that it is better to employ a player with any experience because it has to be better than the devil they don’t know. 3. It’s not so good to go pffft in a contract year. The Health Hush Players get hurt and potentially have a lot to lose, so there is an incentive for them to hide injuries. HIPAA laws restrict the disclosure of health information. Team doctors and trainers have been instructed not to talk with the media. So, when it comes to information on a player’s health status, we’re all pretty much in the dark. Laws of Prognosticating Perspective t Berkeley’s 17th Law: A great many problems do not have accurate answers, but do have approximate answers, from which sensible decisions can be made. t Ashley-Perry Statistical Axiom #4: A complex system that works is invariably found to have evolved from a simple system that works. t Baseball Variation of Harvard Law: Under the most rigorously observed conditions of skill, age, environment, statistical rules and other variables, a ballplayer will perform as he damn well pleases. Hidden Injury Progression 1. Player’s skills implode. 2. Team and player deny injury. 3. More unexplained struggles. 4. Injury revealed; surgery follows. Law of Injury Estimation (Westheimer’s Rule) To calculate an accurate projection of the amount of time a player will be out of action due to injury, first take the published time estimate, double it and change the unit of measure to the next highest unit. Thus, a player estimated to be out two weeks will actually be out four months. Brad Fullmer List Players whose leading indicators indicate upside potential, year after year, but consistently fail to reach that full potential. Ricky Nolasco and Brandon Morrow are at the top of the list right now. The Livan Level The point when a player’s career Runs Above Replacement level has dropped so far below zero that he has effectively cancelled out any possible remaining future value. (Similarly, the Dontrelle Demarcation.) Ceiling The highest professional level at which a player maintains acceptable BPIs. Also, the peak performance level that a player will likely reach, given his BPIs. Monocarp A player whose career consists of only one productive season. Good Luck Truism Good luck is rare and everyone has more of it than you do. That’s the law. Paradoxes and Conundrums 1. Is a player’s improvement in performance from one year to the next a point in a growth trend, an isolated outlier or a complete anomaly? 2. A player can play through an injury, post rotten numbers and put his job at risk… or… he can admit that he can’t play through an injury, allow himself to be taken out of the lineup/rotation, and put his job at risk. 3. Did irregular playing time take its toll on the player’s performance or did poor performance force a reduction in his playing time? 4. Is a player only in the game versus right-handers because he has a true skills deficiency versus left-handers? Or is his poor performance versus left-handers because he’s never given a chance to face them? 5. The problem with stockpiling bench players in the hope that one pans out is that you end up evaluating performance using data sets that are too small to be reliable. 6. There are players who could give you 20 stolen bases if they got 400 AB. But if they got 400 AB, they would likely be on a bad team that wouldn’t let them steal. The Gravity Principles 1. It is easier to be crappy than it is to be good. 2. All performance starts at zero, ends at zero and can drop to zero at any time. 3. The odds of a good performer slumping are far greater than the odds of a poor performer surging. 4. Once a player is in a slump, it takes several 3-for-5 days to get out of it. Once he is on a streak, it takes a single 0-for-4 day to begin the downward spiral. Corollary: Once a player is in a slump, not only does it take several 3-for-5 days to get out of it, but he also has to get his name back on the lineup card. 5. Eventually all performance comes down to earth. It may take a week, or a month, or may not happen until he’s 45, but eventually it’s going to happen. Health Homilies 1. Staying healthy is a skill (and “DL Days” should be a Rotisserie category). 2. A $40 player can get hurt just as easily as a $5 player but is eight times tougher to replace. 3. Chronically injured players never suddenly get healthy. 4. There are two kinds of pitchers: those that are hurt and those that are not hurt... yet. Quack! An exclamation in response to the educated speculation that a player has used performance enhancing drugs. While it is rare to have absolute proof, there is often enough information to suggest 16 2012 Baseball Forecaster / Encyclopedia of Fanalytics UGLY (Unreasonable Good Luck Year) The driving force behind every winning team. It’s what they really mean when they say “winning ugly.” that, “if it looks like a duck and quacks like a duck, then odds are it’s a duck.” Tenets of Optimal Timing 1. If a second half fader had put up his second half stats in the first half and his first half stats in the second half, then he probably wouldn’t even have had a second half. 2. Fast starters can often buy six months of playing time out of one month of productivity. 3. Poor 2nd halves don’t get recognized until it’s too late. 4. “Baseball is like this. Have one good year and you can fool them for five more, because for five more years they expect you to have another good one.” — Frankie Frisch Walbeckian Possessing below replacement level stats, as in “Adam Dunn’s season was downright Walbeckian.” Alternate usage: “Adam Dunn’s stats were so bad that I might as well have had Walbeck in there.” (In fact, Dunn’s season was worthy of renaming this to Dunnian.) Wasted talent A player with a high level skill that is negated by a deficiency in another skill. For instance, base path speed can be negated by poor on base ability. Pitchers with strong arms can be wasted because home plate is an elusive concept to them. The Three True Outcomes 1. Strikeouts 2. Walks 3. Home runs Zombie A player who is indestructible, continuing to get work, year-afteryear, no matter how dead his BPIs are. Like Seth McClung. The Three True Handicaps 1. Has power but can’t make contact. 2. Has speed but can’t hit safely. 3. Has potential but is too old. 17 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Batters Ground ball, line drive, fly ball percentages (G/L/F) The percentage of all balls in play that are hit on the ground, as line drives and in the air. For batters, increased fly ball tendency may foretell a rise in power skills; increased line drive tendency may foretell an improvement in batting average. For a pitcher, the ability to keep the ball on the ground can contribute to his statistical output exceeding his demonstrated skill level. Batting Eye, Contact and Batting Average Batting average (BA, or Avg) This is where it starts. BA is a grand old nugget that has long outgrown its usefulness. We revere .300 hitting superstars and scoff at .250 hitters, yet the difference between the two is one hit every 20 ABs. This one hit every five games is not nearly the wide variance that exists in our perceptions of what it means to be a .300 or .250 hitter. BA is a poor evaluator of performance in that it neglects the offensive value of the base on balls and assumes that all hits are created equal. %,37\SH 7RWDO 2XW *URXQGEDOO /LQHGULYH )O\EDOO 727$/ 'DWDRQO\LQFOXGHVILHOGDEOHEDOOVDQGLVQHWRI+5V Batting eye (Eye) Batting average perception :DONV6WULNHRXWV A measure of a player’s strike zone judgment. BENCHMARKS: The best hitters have Eye ratios more than 1.00 (indicating more walks than strikeouts) and are the most likely to be among a league’s .300 hitters. Ratios less than 0.50 represent batters who likely also have lower BAs. Early season batting average strugglers who surge later in the year get no respect because they have to live with the weight of their early numbers all season long. Conversely, quick starters who fade late get far more accolades than they deserve. For instance, take Darwin Barney’s 2011 month-by-month batting averages. Perception, which is typically based solely on a player’s cumulative season stat line, was that he had a solid rookie campaign. Reality is different. He had only two truly good months, and they happened to occur at the beginning of the season. How many people knew he batted only .256 from June 1 on? 0RQWK $SULO 0D\ -XQH -XO\ $XJXVW 6HSW2FW %$ Batting eye as a leading indicator There is a strong correlation between strike zone judgment and batting average. However, research shows that this is more descriptive than predictive: &XP%$ %DWWLQJ(\H DQGRYHU %DWWLQJ$YHUDJH We can create percentage plays for the different levels: Walk rate (bb%) )RU(\H /HYHOVRI %%$%%% A measure of a batter’s plate patience. BENCHMARKS: The best batters will have levels more than 10%. Those with poor plate patience will have levels of 5% or less. On base average (OB) +%%$%%% Addressing a key deficiency with BA, OB gives value to events that get batters on base, but are not hits. An OB of .350 can be read as “this batter gets on base 35% of the time.” When a run is scored, there is no distinction made as to how that runner reached base. So, two-thirds of the time—about how often a batter comes to the plate with the bases empty—a walk really is as good as a hit. The official version of this formula includes hit batsmen. We do not include it because our focus is on skills-based gauges; research has shown that HBP is not a measure of batting skill but of pitching deficiency. BENCHMARKS: We know what a .300 hitter is, but what represents “good” for OB? That comparable level would likely be .400, with .275 representing the comparable level of futility. 3FWZKREDW Any batter with an eye ratio more than 1.50 has about a 4% chance of hitting less than .250 over 500 at bats. Of all .300 hitters, those with ratios of at least 1.00 have a 65% chance of repeating as .300 hitters. Those with ratios less than 1.00 have less than a 50% chance of repeating. Only 4% of sub-.250 hitters with ratios less than 0.50 will mature into .300 hitters the following year. In a 1995-2000 study, only 37 batters hit .300-plus with a sub-0.50 eye ratio over at least 300 AB in a season. Of this group, 30% were able to accomplish this feat on a consistent basis. For the other 70%, this was a short-term aberration. Contact rate (ct%) $%.$% Measures a batter’s ability to get wood on the ball and hit it into the field of play. BENCHMARKS: Those batters with the best contact 18 2012 Baseball Forecaster / Encyclopedia of Fanalytics P/PA as a leading indicator for BA skill will have levels of 90% or better. The hackers of society will have levels of 75% or less. Contact rate as a leading indicator The more often a batter makes contact with the ball, the higher the likelihood that he will hit safely. &RQWDFW5DWH 2YHU %DWWLQJ$YHUDJH 33$ 8QGHU /RZ33$DQG/RZ%$ +LJK33$DQG+LJK%$ A matrix of contact rates and walk rates can provide expectation benchmarks for a player’s batting average: &RQWDFWUDWHFW Expected batting average <($57:2 %$,PSURYHG %$'HFOLQHG (John Burnson) [&7>[+[+@ ZKHUH [+ *%[>3;OQ6;@ /'[>OQ6;@ )%[ DQG [+ )%[>3;6;@ *%[>3;@ A hitter’s batting average as calculated by multiplying the percentage of balls put in play (contact rate) by the chance that a ball in play falls for a hit. The likelihood that a ball in play falls for a hit is a product of the speed of the ball and distance it is hit (PX), the speed of the batter (SX), and distribution of ground balls, fly balls, and line drives. We further split it out by non-homerun hit rate (xH1%) and homerun hit rate (xH2%). BENCHMARKS: In general, xBA should approximate batting average fairly closely. Those hitters who have large variances between the two gauges are candidates for further analysis. LIMITATION: xBA tends to understate a batter’s true value if he is an extreme ground ball hitter (G/F ratio over 3.0) with a low PX. These players are not inherently weak, but choose to take safe singles rather than swing for the fences. A contact rate of 65% or lower offers virtually no chance for a player to hit even .250, no matter how high a walk rate he has. The .300 hitters most often come from the group with a minimum 86% contact and 11% walk rate. Balls in play (BIP) $%z. The total number of batted balls that are hit fair, both hits and outs. An analysis of how these balls are hit—on the ground, in the air, hits, outs, etc.—can provide analytical insight, from player skill levels to the impact of luck on statistical output. Batting average on balls in play %$ In these scenarios, there was a strong tendency for performance to normalize in year #2. :DONUDWHEE 2%$ Generally speaking, the more pitches seen,the lower the BA, but the higher the OBA. But what about the outliers, those players that bucked the trend in year #1? Contact rate and walk rate as leading indicators (Paul Petera) The art of working the count has long been considered one of the more crucial aspects of good hitting. It is common knowledge that the more pitches a hitter sees, the greater opportunity he has to reach base safely. (Voros McCracken) +z+5$%z+5z. Also called hit rate (h%). The percent of balls hit into the field of play that fall for hits. BENCHMARK: Every hitter establishes his own individual hit rate that stabilizes over time. A batter whose seasonal hit rate varies significantly from the h% he has established over the preceding three seasons (variance of at least +/- 3%) is likely to improve or regress to his individual h% mean (with over-performer declines more likely and sharper than under-performer recoveries). Three-year h% levels strongly predict a player’s h% the following year. Expected batting average variance [%$%$ The variance between a batter’s BA and his xBA is a measure of over- or under-achievement. A positive variance indicates the potential for a batter’s BA to rise. A negative variance indicates the potential for BA to decline. BENCHMARK: Discount variances that are less than 20 points. Any variance more than 30 points is regarded as a strong indicator of future change. 19 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Power Linear weighted power index (PX) %DWWHUV/:3ZU/HDJXH/:3ZU[ LWPwr is presented in this book in its normalized form to get a better read on a batter’s accomplishment in each year. For instance, a 30-HR season today is much more of an accomplishment than 30 HRs hit in a higher offense year like 2003. BENCHMARKS: A level of 100 equals league average power skills. Any player with a value more than 100 has above average power skills, and those more than 150 are the Slugging Elite. Slugging average (Slg) 6LQJOHV['RXEOHV[7ULSOHV[+5$% A measure of the total number of bases accumulated (or the minimum number of runners’ bases advanced) per at bat. It is a misnomer; it is not a true measure of a batter’s slugging ability because it includes singles. Slg also assumes that each type of hit has proportionately increasing value (i.e. a double is twice as valuable as a single, etc.) which is not true. For instance, with the bases loaded, a HR always scores four runs, a triple always scores three, but a double could score two or three and a single could score one, or two, or even three. BENCHMARKS: Top batters will have levels over .500. The bottom batters will have levels less than .300. Fly ball tendency and power Expected LW power index (xPX) (Bill Macey) ++/'++)% Previous research has shown that hard-hit balls are more likely to result in hits and hard-hit fly balls are more likely to end up as HRs. As such, we can use hard-hit ball data to calculate an expected skills-based power index. This metric starts with hardhit ball data, which measures a player’s fundamental skill of making solid contact, and then places it on the same scale as PX (xPX). In the above formula, HHLD% is calculated as the number of hard hit-line drives divided by the total number of balls put in play. HHFB% is similarly calculated for fly balls. (Mat Olkin) There is a proven connection between a hitter’s ground ball/fly ball tendencies and his power production. 1. Extreme ground ball hitters generally do not hit for much power. It’s almost impossible for a hitter with a ground/ fly ratio over 1.80 to hit enough fly balls to produce even 25 HRs in a season. However, this does not mean that a low G/F ratio necessarily guarantees power production. Some players have no problem getting the ball into the air, but lack the strength to reach the fences consistently. 2. Most batters’ ground/fly ratios stay pretty steady over time. Most year-to-year changes are small and random, as they are in any other statistical category. A large, sudden change in G/F, on the other hand, can signal a conscious change in plate approach. And so... 3. If a player posts high G/F ratios in his first few years, he probably isn’t ever going to hit for all that much power. 4. When a batter’s power suddenly jumps, his G/F ratio often drops at the same time. 5. Every so often, a hitter’s ratio will drop significantly even as his power production remains level. In these rare cases, impending power development is likely, since the two factors almost always follow each other. P/PA as a leading indicator for PX 33$ 3; 8QGHU As for the year #1 outliers: /RZ33$DQG+LJK3; +LJK33$DQG/RZ3; <($57:2 3;,PSURYHG 3;'HFOLQHG In these scenarios, there was a strong tendency for performance to normalize in year #2 Opposite field home runs (Ed DeCaria) From 2001-2008, nearly 75% of all HRs were hit to the batter’s pull field, with the remaining 25% distributed roughly evenly between straight away and opposite field. Left-handers accomplished the feat slightly more often than right-handers (including switchhitters hitting each way), and younger hitters did it significantly more often than older hitters. The trend toward pulled home runs was especially strong after age 36. Home runs to fly ball rate (hr/f) The percent of fly balls that are hit for HRs. HR/F rate as a leading indicator (Paul Petera) Working the count has a positive effect on power. (Joshua Randall) Each batter establishes an individual home run to fly ball rate that stabilizes over rolling three-year periods; those levels strongly predict the hr/f in the subsequent year. A batter who varies significantly from his hr/f is likely to regress toward his individual hr/f mean, with over-performance decline more likely and more severe than under-performance recovery. 3RZHU 4XDUWLOH 7RS QG UG %RW Linear weighted power (LWPwr) 'RXEOHV[7ULSOHV[+5[$WEDWV.[ A variation of the linear weights formula that considers only events that are measures of a batter’s pure power. BENCHMARKS: Top sluggers typically top the 17 mark. Weak hitters will have a LWPwr level of less than 10. $%+5 2SS 6WUDLJKW )LHOG $ZD\ 3XOO )LHOG Opposite field HRs serve as a strong indicator of overall home run power (AB/HR). Power hitters (smaller AB/HR rates) hit a far higher percentage of their HR to the opposite field or straight away (over 30%). Conversely, non-power hitters hit almost 90% of their home runs to their pull field. 20 2012 Baseball Forecaster / Encyclopedia of Fanalytics version) may be used as a leading indicator for stolen base output, SB attempts are controlled by managerial strategy which makes speed score somewhat less valuable. Speed score is calculated as the mean value of the following four elements: 1. Stolen base efficiency = 6%6%&6[ 2. Stolen base freq. = 6TXDUHURRWRI6%&66LQJOHV%% 3. Triples rating = %$%+5. and the result assigned a value based on the following chart: 3HUIRUPDQFHLQ<<RI*URXS <7ULJJHU $%+5 5&* 5 2SS+5 2SS+5 Players who hit just two or more OppHR in one season were 2-3 times as likely as those who hit zero or one OppHR to sustain strong AB/HR rates, RC/G levels, or R$ values over the following three seasons. <<%UHDNRXW3HUIRUPDQFH %UHDNRXWE\*URXS$JH 2QO\ $%+5 5&* 5 <7ULJJHU !WR WR WR 2SS+5 2SS+5 4. Runs scored as a percentage of times on base = 5+5+%%+5 Roughly one of every 3-4 batters age 26 or younger experiences a sustained three-year breakout in AB/HR, RC/G or R$ after a season in which they hit 2+ OppHR, far better odds than the one in 8-10 batters who experience a breakout without the 2+ OppHR trigger. Speed score index (SX) Power breakout profile %DWWHUVVSHHGVFRUH/HDJXHVSHHGVFRUH[ Normalized speed scores get a better read on a runner’s accomplishment in context. A level of 100 equals league average speed skill. Values more than 100 indicate above average skill, more than 200 represent the Fleet of Feet Elite. It is not easy to predict which batters will experience a power spike. We can categorize power breakouts to determine the likelihood of a player taking a step up or of a surprise performer repeating his feat. Possibilities: t *ODSFBTFJOQMBZJOHUJNF t )JTUPSZPGQPXFSTLJMMTBUTPNFUJNFJOUIFQBTU t 3FEJTUSJCVUJPO PG BMSFBEZ EFNPOTUSBUFE FYUSB CBTF IJU power t /PSNBMTLJMMTHSPXUI t 4JUVBUJPOBM CSFBLPVUT QBSUJDVMBSMZ JO IJUUFSGSJFOEMZ venues t *ODSFBTFEĘZCBMMUFOEFODZ t 6TFPGJMMFHBMQFSGPSNBODFFOIBODJOHTVCTUBODFT t .JTDFMMBOFPVTVOFYQMBJOFEWBSJBCMFT Statistically scouted speed (Spd) ^>5XQV+5DJHBZW5%,+5@OJBDY` ^>%DJHBZW%%@OJBDY` ^>6RIW0HG*%KLWVDJHBZW6RIW0HG*%@OJBDY` ^>:HLJKW/EV+HLJKW,QA@OJBDY` A skills-based gauge that measures speed without relying on stolen bases. Its components are: t 5XQV+55%,+5: This metric aims to minimize the influence of extra base hit power and team run-scoring rates on perceived speed. t % % %: No one can deny that triples are a fast runner’s stat; dividing them by 2B+3B instead of all balls in play dampens the power aspect of extra base hits. t 6RIW0HGLXP*URXQG%DOO+LWV6RIW0HGLXP*URXQG %DOOV: Faster runners are more likely than slower runners to beat out routine grounders. Hard hit balls are excluded from numerator and denominator. t %RG\0DVV,QGH[%0,: Calculated as :HLJKWOEV+HLJKW LQ . All other factors considered, leaner players run faster than heavier ones. In this book, the formula is scaled as an index with a midpoint of 100. Speed Wasted talent on the base paths We refer to some players as having “wasted talent,” a high level skill that is negated by a deficiency in another skill. Among these types are players who have blazing speed that is negated by a sub-.300 on base average. These players can have short-term value. However, their stolen base totals are tied so tightly to their “green light” that any change in managerial strategy could completely erase that value. A higher OB mitigates that downside; the good news is that plate patience can be taught. Players in 2011 who had at least 20 SBs with an OB less than .300, and whose SB output could be at risk, are Rajai Davis (34 SB, .272 OBP), Alcides Escobar (26, .286), Aaron Hill (21, .294), Ian Desmond (25, .296) and Tony Campana (24, .298). Speed score Stolen base opportunity percent (SBO) 6%&6%%6LQJOHV A rough approximation of how often a baserunner attempts a stolen base. Provides a comparative measure for players on a given team and, as a team measure, the propensity of a manager to give a “green light” to his runners. (Bill James) A measure of the various elements that comprise a runner’s speed skills. Although this formula (a variation of James’ original 21 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Overall Performance Analysis Replacement levels used in this book, for 2011: 326 & % % % 66 /) &) 5) '+ On base plus slugging average (OPS) A simple sum of the two gauges, it is considered one of the better evaluators of overall performance. OPS combines the two basic elements of offensive production—the ability to get on base (OB) and the ability to advance baserunners (Slg). BENCHMARKS: The game’s top batters will have OPS levels more than .900. The worst batters will have levels less than .600. Base Performance Value (BPV) 1/ RAR can also be used to calculate rough projected team wonloss records. (Roger Miller) Total the RAR levels for all the players on a team, divide by 10 and add to 53 wins. :DONUDWH[ &RQWDFWUDWH[ 3RZHU,QGH[[ 6SG[ A single value that describes a player’s overall raw skill level. This is more useful than traditional statistical gauges to track player performance trends and project future statistical output. The BPV formula combines and weights several BPIs. This formula combines the individual raw skills of batting eye, contact rate, power and speed. BENCHMARKS: The best hitters will have a BPV of 50 or greater. Linear weights $/ Runs created (Bill James) +%%&6[7RWDOEDVHV[6%$%%% A formula that converts all offensive events into a total of runs scored. As calculated for individual teams, the result approximates a club’s actual run total with great accuracy. Runs created per game Bill James version: 5XQV&UHDWHG$%+&6 Neil Bonner version (used in this book): 66[FW[EE[ where SS, or “swing speed” is defined as %[%[%[+5[$%. RC expressed on a per-game basis might be considered the hypothetical ERA compiled against a particular batter. Another way to look at it: A batter with a RC/G of 7.00 would be expected to score 7 runs per game if he were cloned nine times and faced an average pitcher in every at bat. Cloning batters is not a practice we recommend. BENCHMARKS: Few players surpass the level of a 10.00 RC/G, but any level more than 7.50 can still be considered very good. At the bottom are levels less than 3.00. (Pete Palmer) 6LQJOHV['RXEOHV[7ULSOHV[ +RPHUXQV[:DONV[6WROHQ%DVHV[ &DXJKW6WHDOLQJ[$WEDWV+LWV[1RUPDOL]LQJ)DFWRU (Also referred to as Batting Runs.) Formula whose premise is that all events in baseball are linear; that is, the output (runs) is directly proportional to the input (offensive events). Each of these events is then weighted according to its relative value in producing runs. Positive events—hits, walks, stolen bases—have positive values. Negative events—outs, caught stealing—have negative values. The normalizing factor, representing the value of an out, is an offset to the level of offense in a given year. It changes every season, growing larger in high offense years and smaller in low offense years. The value is about .26 and varies by league. LW is not included in the player forecast boxes, but the LW concept is used with the linear weighted power gauge. Handedness 1. While pure southpaws account for about 27% of total ABs (RHers about 55% and switch-hitters about 18%), they hit 31% of the triples and take 30% of the walks. 2. The average lefty posts a batting average about 10 points higher than the average RHer. The on base averages of pure LHers are nearly 20 points higher than RHers, but only 10 points higher than switch-hitters. 3. LHers tend to have a better batting eye ratio than RHers, but about the same as switch-hitters. 4. Pure righties and lefties have virtually identical power skills. Switch-hitters tend to have less power, on average. 5. Switch-hitters tend to have the best speed, followed by LHers, and then RHers. 6. On an overall production basis, LHers have an 8% advantage over RHers and a 14% edge over switch-hitters. Runs above replacement (RAR) An estimate of the number of runs a player contributes above a “replacement level” player. “Replacement” is defined as the level of performance at which another player can easily be found at little or no cost to a team. What constitutes replacement level is a topic that is hotly debated. There are a variety of formulas and rules of thumb used to determine this level for each position (replacement level for a shortstop will be very different from replacement level for an outfielder). Our estimates appear below. One of the major values of RAR for fantasy applications is that it can be used to assemble an integrated ranking of batters and pitchers for drafting purposes. To calculate RAR for batters: t 4UBSUXJUIBCBUUFSTSVOTDSFBUFEQFSHBNF3$( t 4VCUSBDUIJTQPTJUJPOTSFQMBDFNFOUMFWFM3$( t .VMUJQMZCZOVNCFSPGHBNFTQMBZFE"#)$4 22 2012 Baseball Forecaster / Encyclopedia of Fanalytics In-Season Analysis Optimal ages Every player develops at a different pace, but age can still be helpful to determine where they should be along the developmental curve. Bill James’ research said batters peak at about 27. Recent research suggests that average is now closer to 30. More tendencies: Power: Batting power tends to grow consistently between 24 and 29. Many batters experience a power peak around 30-31. Catchers often see a power spike in the mid-30s. Speed: Base-running and speed are skills of the young. When given the choice of two speedsters of equivalent abilities and opportunity, always go after the younger one. A sharp drop-off in speed skills typically occurs around age 34. Batting eye: For batters who continue to play into their 30s, this is a skill that can develop and grow throughout their career. A decline in this level, which can occur at any age, often indicates a decline in overall skills. Thirtysomethings (Ed Spaulding): Batters tend to lose points on their BA but draw more walks. While players on the outside of the defensive spectrum (1B, 3B, LF, RF, DH) often have their best seasons in their 30s, players in the middle (2B, SS, CF) tend to fade. Catchers (Ed Spaulding): Many catchers—particularly second line catchers—have their best seasons late in their careers. Some possible reasons why: 1. Catchers, like shortstops, often get to the big leagues for defensive reasons and not their offensive skills. These skills take longer to develop. 2. The heavy emphasis on learning the catching/ defense/ pitching side of the game detracts from their time to learn about, and practice, hitting. 3. Injuries often curtail their ability to show offensive skills, though these injuries (typically jammed fingers, bruises on the arms, rib injuries from collisions) often don’t lead to time on the disabled list. 4. The time spent behind the plate has to impact the ability to recognize, and eventually hit, all kinds of pitches. Batting order facts (Ed DeCaria) Eighty-eight percent of today’s leadoff hitters bat leadoff again in their next game, 78% still bat leadoff 10 games later, and 68% still bat leadoff 50 games later. Despite this level of turnover after 50 games, leadoff hitters have the best chance of retaining their role over time. After leadoff, #3 and #4 hitters are the next most likely to retain their lineup slots. On a season-to-season basis, leadoff hitters are again the most stable, with 69% of last year’s primary leadoff hitters retaining the #1 slot next year. Plate appearances decline linearly by lineup slot. Leadoff batters receive 10-12% more PAs than when batting lower in the lineup. AL #9 batters and NL #8 batters get 9-10% fewer PAs. These results mirror play-by-play data showing a 15-20 PA drop by lineup slot over a full season. Walk rate is largely unaffected by lineup slot in the AL. Beware strong walk rates by NL #8 hitters, as much of this “skill” will disappear if ever moved from the #8 slot. Batting order has no discernable effect on contact rate. Hit rate slopes gently upward as hitters are slotted deeper in the lineup. As expected, the #3-4-5 slots are ideal for non-HR RBIs, at the expense of #6 hitters. RBIs are worst for players in the #1-2 slots. Batting atop the order sharply increases the probability of scoring runs, especially in the NL. The leadoff slot easily has the highest stolen base attempt rate. #4-5-6 hitters attempt steals more often when batting out of those slots than they do batting elsewhere. The NL #8 hitter is a SB attempt sink hole. A change in batting order from #8 to #1 in the NL could nearly double a player’s SB output due to lineup slot alone. DOMination and DISaster rates A hitter’s spring training Slg .200 or more above his lifetime Slg is a leading indicator for a better than normal season. Week-to-week consistency is measured using a batter’s BPV compiled in each week. A player earns a DOMinant week if his BPV was greater or equal to 50 for that week. A player registers a DISaster if his BPV was less than 0 for that week. The percentage of Dominant weeks, DOM%, is simply calculated as the number of DOM weeks divided by the total number of weeks played. Overall batting breakout profile Is week-to-week consistency a repeatable skill? (Bill Macey) Spring training Slg as leading indicator (John Dewan) (Brandon Kruse) We define a breakout performance as one where a player posts a Roto value of $20+ after having never posted a value of $10. These criteria are used to validate an apparent breakout in the current season but may also be used carefully to project a potential upcoming breakout: t "HFPSZPVOHFS t "OJODSFBTFJOBUMFBTUUXPPGI19PS4QE t .JOJNVNMFBHVFBWFSBHF19PS4QE t .JOJNVNDPOUBDUSBUFPG t .JOJNVNY#"PG To test whether consistent performance is a repeatable skill for batters, we examined how closely related a player’s DOM% was from year to year. <5'20 z z $9*<5'20 Quality/consistency score (QC) '20[',6[ Using the DOM/DIS percentages, this score measures both the quality of performance as well as week–to-week consistency. 23 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Sample size reliability Other Diamonds (Russell Carleton) At what point during the season do statistics become reliable indicators of skill? Measured in plate appearances: 100: Contact rate 150: Strikeout rate, line drive rate, pitches/PA 200: Walk rate, ground ball rate, GB/FB 250: Fly ball rate 300: HR rate, hr/f 500: OBP, Slg, OPS 550: Isolated power Unlisted stats did not stabilize over a full season of play. Projecting RBIs It’s a Busy World Shortcut For marginal utility-type players, scan their PX and Spd history to see if there’s anything to mine for. If you see triple digits anywhere, stop and look further. If not, move on. Chronology of the Classic Free-Swinger with Pop 1. Gets off to a good start. 2. Thinks he’s in a groove. 3. Gets lax, careless. 4. Pitchers begin to catch on. 5. Fades down the stretch. Errant Gust of Wind A unit of measure used to describe the difference between your home run projection and mine. (Patrick Davitt) Evaluating players in-season for RBI potential is a function of the interplay among four factors: t 5FBNNBUFT BCJMJUZ UP SFBDI CBTF BIFBE PG IJN BOE UP run the bases efficiently t )JTPXOBCJMJUZUPESJWFUIFNJOCZIJUUJOHFTQFDJBMMZ9#) t /VNCFSPG(BNFT1MBZFE t 1MBDFJOUIFCBUUJOHPSEFS 3-4-5 Hitters: [*3[72%[,7%[+5z[*3 6-7-8 Hitters: [*3[72%[,7%[+5z[*3 9-1-2 Hitters: [*3[72%[,7%[+5z[*3 ...where *3 JDPHVSOD\HG72% WHDPRQEDVHSFW and ,7% LQGLYLGXDOWRWDOEDVHV,7%. Apply this pRBI formula after 70 games played or so (to reduce the variation from small sample size) to find players more than 9 RBIs over or under their projected RBI. There could be a correction coming. You should also consider other factors, like injury or trade (involving the player or a top-of-the-order speedster) or team SB philosophy and success rate. Remember: the player himself has an impact on his TOB. When we first did this study, we excluded the player from his TOB and got better results. The formula overestimates projected RBI for players with high OBP who skew his teams’ OBP but can’t benefit in RBI from that effect. Hannahan Concession Players with a .218 BA rarely get 500 plate appearances, but when they do, it’s usually once. Mendoza Line Named for Mario Mendoza, it represents the benchmark for batting futility. Usually refers to a .200 batting average, but can also be used for low levels of other statistical categories. Note that Mendoza’s lifetime batting average was actually a much more robust .215. Old Player Skills Power, low batting average, no speed and usually good plate patience. Young players, often those with a larger frame, who possess these “old player skills” tend to decline faster than normal, often in their early 30s. Small Sample Certitude If players’ careers were judged based what they did in a single game performance, then Tuffy Rhodes and Mark Whiten would be in the Hall of Fame. Esix Snead List Players with excellent speed and sub-.300 on base averages who get a lot of practice running down the line to first base, and then back to the dugout. Also used as an adjective, as in “Esix-Sneadian.” 24 2012 Baseball Forecaster / Encyclopedia of Fanalytics Pitchers We can create percentage plays for the different levels: )RU&PG ZLWK(5$RI /HYHOVRI Strikeouts and Walks Fundamental skills Unreliable pitching performance is a fallacy driven by the practice of attempting to project pitching stats using gauges that are poor evaluators of skill. How can we better evaluate pitching skill? We can start with the three statistical categories that are generally unaffected by external factors. These three stats capture the outcome of an individual pitcher versus batter match-up without regard to supporting offense, defense or bullpen: Walks Allowed, Strikeouts and Ground Balls Even with only these stats to observe, there is a wealth of insight that these measures can provide. Pitchers who maintain a Cmd over 2.5 have a high probability of long-term success. For fantasy drafting purposes, it is best to avoid pitchers with sub-2.0 ratios. Avoid bullpen closers if they have a ratio less than 2.5. 2. A pitcher’s command in tandem with dominance (strikeout rate) provides even greater predictive abilities. (DUQHG5XQ$YHUDJH &RPPDQG 'RP 'RP Control rate (Ctl, bb/9), or opposition walks per game %%DOORZHG[,3 Measures how many walks a pitcher allows per game equivalent. BENCHMARK: The best pitchers will have bb/9 levels of 3.0 or less. This helps to highlight the limited upside potential of softtossers with pinpoint control. The extra dominance makes a huge difference. 3. Research also suggests that there is a strong correlation between a pitcher’s command ratio and his propensity to win ballgames. Over three quarters of those with ratios over 3.0 post winning records, and the collective W/L record of those command artists is nearly .600. The command/winning correlation holds up in both leagues, although the effect was more pronounced in the NL. Over four times more NL hurlers than AL hurlers had Cmd over 3.0, and higher ratios were required in the NL to maintain good winning percentages. A ratio between 2.0 and 2.9 was good enough for a winning record for over 70% of AL pitchers, but that level in the NL generated an above-.500 mark slightly more than half the time. In short, in order to have at least a 70% chance of drafting a pitcher with a winning record, you must target NL pitchers with at least a 3.0 command ratio. To achieve the same odds in the AL, a 2.0 command ratio will suffice. Dominance rate (Dom, k/9), or opposition strikeouts/game 6WULNHRXWVUHFRUGHG[,3 Measures how many strikeouts a pitcher allows per game equivalent. BENCHMARK: The best pitchers will have k/9 levels of 5.6 or higher. Command ratio (Cmd) 6WULNHRXWV:DONV A measure of a pitcher’s ability to get the ball over the plate. There is no more fundamental a skill than this, and so it is used as a leading indicator to project future rises and falls in other gauges, such as ERA. BENCHMARKS: Baseball’s best pitchers will have ratios in excess of 3.0. Pitchers with ratios less than 1.0—indicating that they walk more batters than they strike out—have virtually no potential for long-term success. If you make no other changes in your approach to drafting pitchers, limiting your focus to only pitchers with a command ratio of 2.0 or better will substantially improve your odds of success. Command ratio as a leading indicator The ability to get the ball over the plate—command of the strike zone—is one of the best leading indicators for future performance. Command ratio (K/BB) can be used to project potential in ERA as well as other skills gauges. 1. Research indicates that there is a high correlation between a pitcher’s Cmd ratio and his ERA. &RPPDQG (DUQHG5XQ$YHUDJH Power/contact rating %%.,3 Measures the level by which a pitcher allows balls to be put into play. In general, extreme power pitchers can be successful even with poor defensive teams. Power pitchers tend to have greater longevity in the game. Contact pitchers with poor defenses behind them are high risks to have poor W-L records and ERA. BENCHMARKS: A level of 1.13+ describes pure throwers. A level of .93 or less describes high contact pitchers. 25 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Balls in Play pitchers will have levels less than .300; the worst pitchers levels more than .375. Balls in play (BIP) Walks plus hits divided by innings pitched (WHIP) ,3[+z. The total number of batted balls that are hit fair, both hits and outs. An analysis of how these balls are hit—on the ground, in the air, hits, outs, etc.—can provide analytical insight, from player skill levels to the impact of luck on statistical output. Batting average on balls in play Essentially the same measure as opposition on base average, but used for Rotisserie purposes. BENCHMARKS: A WHIP of less than 1.20 is considered top level; more than 1.50 indicative of poor performance. Levels less than 1.00—allowing fewer runners than IP—represent extraordinary performance and are rarely maintained over time. (Voros McCracken) +z+5,3[+z.z+5 Abbreviated as BABIP; also called hit rate (H%). The percent of balls hit into the field of play that fall for hits. BENCHMARK: The league average is 30%, which is also the level that individual performances will regress to on a year to year basis. Any +/- variance of 3% or more can affect a pitcher’s ERA. BABIP as a leading indicator Ground ball, line drive, fly ball percentage (G/L/F) The percentage of all balls-in-play that are hit on the ground, in the air and as line drives. For a pitcher, the ability to keep the ball on the ground can contribute to his statistical output exceeding his demonstrated skill level. Ground ball tendency as a leading indicator (John Burnson) (Voros McCracken) Ground ball pitchers tend to give up fewer HRs than do fly ball pitchers. There is also evidence that GB pitchers have higher hit rates. In other words, a ground ball has a higher chance of being a hit than does a fly ball that is not out of the park. GB pitchers have lower strikeout rates. We should be more forgiving of a low strikeout rate (under 5.5 K/9) if it belongs to an extreme ground ball pitcher. GB pitchers have a lower ERA but a higher WHIP than do fly ball pitchers. On balance, GB pitchers come out ahead, even when considering strikeouts, because a lower ERA also leads to more wins. In 2000, Voros McCracken published a study that concluded that “there is little if any difference among major league pitchers in their ability to prevent hits on balls hit in the field of play.” His assertion was that, while a Johan Santana would have a better ability to prevent a batter from getting wood on a ball, or perhaps keeping the ball in the park, once that ball was hit in the field of play, the probability of it falling for a hit was virtually no different than for any other pitcher. Among the findings in his study were: t ćFSF JT MJUUMF DPSSFMBUJPO CFUXFFO XIBU B QJUDIFS EPFT one year in the stat and what he will do the next. This is not true with other significant stats (BB, K, HR). t :PVDBOCFUUFSQSFEJDUBQJUDIFSTIJUTQFSCBMMTJOQMBZ from the rate of the rest of the pitcher’s team than from the pitcher’s own rate. This last point brings a team’s defense into the picture. It begs the question, when a batter gets a hit, is it because the pitcher made a bad pitch, the batter took a good swing, or the defense was not positioned correctly? Pitchers will often post hit rates per balls-in-play that are far off from the league average, but then revert to the mean the following year. As such, we can use that mean to project the direction of a pitcher’s ERA. Subsequent research has shown that ground ball or fly ball propensity has some impact on this rate. Groundball and strikeout tendencies as indicators (Mike Dranchak) Pitchers were assembled into 9 groups based on the following profiles (minimum 23 starts in 2005): 3URILOH *URXQG%DOO 1HXWUDO )O\%DOO *URXQG%DOO5DWH KLJKHUWKDQ WR OHVVWKDQ 3URILOH 6WULNHRXW $YHUDJH 6RIW7RVVHU 6WULNHRXW5DWHN KLJKHUWKDQN WRN OHVVWKDQN Findings: Pitchers with higher strikeout rates had better ERAs and WHIPs than pitchers with lower strikeout rates, regardless of ground ball profile. However, for pitchers with similar strikeout rates, those with higher ground ball rates had better ERAs and WHIPs than those with lower ground ball rates. Pitchers with higher strikeout rates tended to strand more baserunners than those with lower K rates. Fly ball pitchers tended to strand fewer runners than their GB or neutral counterparts within their strikeout profile. Ground ball pitchers (especially those who lacked highdominance) yielded more home runs per fly ball than did fly ball pitchers. However, the ERA risk was mitigated by the fact that ground ball pitchers (by definition) gave up fewer fly balls to begin with. Hit rate (See Batting average on balls in play) Opposition batting average (OBA) +LWV$OORZHG,3[+LWV$OORZHG A close approximation of the batting average achieved by opposing batters against a pitcher. BENCHMARKS: The best pitchers will have levels less than .250; the worst pitchers levels more than .300. Opposition on base average (OOB) +LWV$OORZHG%%,3[+%% A close approximation of the on base average achieved by opposing batters against a pitcher. BENCHMARK: The best 26 2012 Baseball Forecaster / Encyclopedia of Fanalytics Line Drive percentage as a leading indicator Expected ERA variance (Seth Samuels) Also beyond a pitcher’s control is the percentage of balls-in-play that are line drives. Line drives do the most damage; from 19942003, here were the expected hit rates and number of total bases per type of BIP. + 7RWDOEDVHV [(5$(5$ The variance between a pitcher’s ERA and his xERA is a measure of over or underachievement. A positive variance indicates the potential for a pitcher’s ERA to rise. A negative variance indicates the potential for ERA improvement. BENCHMARK: Discount variances that are less than 0.50. Any variance more than 1.00 (one run per game) is regarded as a indicator of future change. _{{{7\SHRI%,3{{{_ *% )% /' Despite the damage done by LDs, pitchers do not have any innate skill to avoid them. There is little relationship between a pitcher’s LD% one year and his rate the next year. All rates tend to regress towards a mean of 22.6%. However, GB pitchers do have a slight ability to prevent line drives (21.7%) and extreme GB hurlers even moreso (18.5%). Extreme FB pitchers have a slight ability to prevent LDs (21.1%) as well. Projected xERA or projected ERA? Home run to fly ball rate +%%(5+%%+5 Measures the percentage of allowed runners a pitcher strands (earned runs only), which incorporates both individual pitcher skill and bullpen effectiveness. BENCHMARKS: The most adept at stranding runners will have S% levels over 75%. Those with rates over 80% will have artificially low ERAs which will be prone to relapse. Levels below 65% will inflate ERA but have a high probability of regression. Which should we be using to forecast a pitcher’s ERA? Projected xERA is more accurate for looking ahead on a purely skills basis. Projected ERA includes situation-dependent events—bullpen support, park factors, etc.—which are reflected better by ERA. The optimal approach is to use both gauges as a range of expectation for forecasting purposes. Strand rate (S%) +5)% The percent of fly balls that are hit for home runs. HR/FB rate as a leading indicator (John Burnson) McCracken’s work focused on “balls in play,” omitting home runs from the study. However, pitchers also do not have much control over the percentage of fly balls that turn into HR. Research shows that there is an underlying rate of HR as a percentage of fly balls of about 10%. A pitcher’s HR/FB rate will vary each year but always tends to regress to that 10%. The element that pitchers do have control over is the number of fly balls they allow. That is the underlying skill or deficiency that controls their HR rate. Pitchers who keep the ball out of the air more often correlate well with Roto value. Expected strand rate (Michael Weddell) .%%%% *%A*% 0$;0,1,3* LIOHIWKDQGHG This formula is based on three core skills: strikeouts per nine innings, walks per nine innings, and groundballs per balls in play, with adjustments for whether the pitcher is a starter or reliever (measured by IP/G), and his handedness. Opposition home runs per game (hr/9) +5$OORZHG[,3 $OVRH[SHFWHGRSSRVLWLRQ+5UDWH )%[[,3 Measures how many HR a pitcher allows per game equivalent. Since FB tend to go yard at about a 10% rate, we can also estimate this rate off of fly balls. BENCHMARK: The best pitchers will have hr/9 levels of less than 1.0. Strand rate as a leading indicator (Ed DeCaria) Strand rate often regresses/rebounds toward past rates (usually 69-74%), resulting in Year 2 ERA changes: RI3LWFKHUVZLWK<HDU5HJUHVVLRQ5HERXQG <6 53 63 /5 Runs Expected earned run average Gill and Reeve version:[+>SHU,3@[+5>SHU,3@ [%%>SHU,3@[.>SHU,3@1RUPDOL]LQJ)DFWRU John Burnson version (used in this book): [(5[,3ZKHUH[(5LVGHILQHGDV [(5[)%[6[>[%,3)%%%@ ZKHUH[(5 [*%)% DQG [6 .[%%[%% [*%A[*% xERA represents the an equivalent of what a pitcher’s real ERA might be, calculated solely with skills-based measures. It is not influenced by situation-dependent factors. 7\SLFDO(5$5HJUHVVLRQ5HERXQGLQ<HDU <6 53 63 /5 QD Starting pitchers (SP) have a narrower range of strand rate outcomes than do relievers (RP) or swingmen/long relievers (LR). 27 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Relief pitchers with Y1 strand rates of <=67% or >=78% are likely to experience a +/- ERA regression in Y2. Starters and swingmen/ long relievers with Y1 strand rates of <=65% or >=75% are likely to experience a +/- ERA regression in Y2. Pitchers with strand rates that deviate more than a few points off of their individual expected strand rates are likely to experience some degree of ERA regression in Y2. Over-performing (or “lucky”) pitchers are more likely than underperforming (or “unlucky”) pitchers to see such a correction. Relievers’ overuse. Warning flags should be up for relievers who post in excess of 100 IP in a season, while averaging fewer than 2 IP per outing. When focusing solely on minor league pitchers, research results are striking: Stamina: Virtually every minor league pitcher who had a BF/G of 28.5 or more in one season experienced a drop-off in BF/G the following year. Many were unable to ever duplicate that previous level of durability. Performance: Most pitchers experienced an associated dropoff in their BPVs in the years following the 28.5 BF/G season. Some were able to salvage their effectiveness later on by moving to the bullpen. Wins Projecting/chasing wins There are five events that need to occur in order for a pitcher to post a single win... 1. He must pitch well, allowing few runs. 2. The offense must score enough runs. 3. The defense must successfully field all batted balls. 4. The bullpen must hold the lead. 5. The manager must leave the pitcher in for 5 innings, and not remove him if the team is still behind. Of these five events, only one is within the control of the pitcher. As such, projecting, or chasing wins can be an exercise in futility. Protecting young pitchers (Craig Wright) There is a link between some degree of eventual arm trouble and a history of heavy workloads in a pitcher’s formative years. Some recommendations from this research: Teenagers (A-ball): No 200 IP seasons and no BF/G over 28.5 in any 150 IP span. No starts on three days rest. Ages 20-22: Average no more than 105 pitches per start with a single game ceiling of 130 pitches. Ages 23-24: Average no more than 110 pitches per start with a single game ceiling of 140 pitches. When possible, a young starter should be introduced to the majors in long relief before he goes into the rotation. Home field advantage (John Burnson) A 2006 study found that home starting pitchers get credited with a win in 38% of their outings. Visiting team starters are credited with a win in 33% of their outings. Overall Performance Analysis Base Performance Value (BPV) Usage 'RPLQDQFH5DWH[ :DON5DWH[ *URXQGEDOOUDWHDVDZKROHQXPEHU A single value that describes a player’s overall raw skill level. This is more useful than traditional statistical gauges to track player performance trends and project future statistical output. The formula combines the individual raw skills of power, control and the ability to keep the ball down in the zone, all characteristics that are unaffected by most external factors. In tandem with a pitcher’s strand rate, it provides a more complete picture of the elements that contribute to ERA, and therefore serves as an accurate tool to project likely changes in ERA. BENCHMARKS: A BPV of 50 is the minimum level required for long-term success. The elite of the bullpen aces will have BPVs in excess of 100 and it is rare for these stoppers to enjoy long term success with consistent levels under 75. Batters faced per game (Craig Wright) ,3[+%%* A measure of pitcher usage and one of the leading indicators for potential pitcher burnout. Workload Research suggests that there is a finite number of innings in a pitcher’s arm. This number varies by pitcher, by development cycle, and by pitching style and repertoire. We can measure a pitcher’s potential for future arm problems and/or reduced effectiveness (burnout): Sharp increases in usage from one year to the next. Common wisdom has suggested that pitchers who significantly increase their workload from one year to the next are candidates for burnout symptoms. This has often been called the Verducci Effect, after writer Tom Verducci. BaseballHQ.com analyst Michael Weddell tested pitchers with sharp workload increases during the period 1988-2008 and found that no such effect exists. Starters’ overuse. Consistent “batters faced per game” (BF/G) levels of 28.0 or higher, combined with consistent seasonal IP totals of 200 or more may indicate burnout potential. Within a season, a BF/G of more than 30.0 with a projected IP total of 200 may indicate a late season fade. Runs above replacement (RAR) An estimate of the number of runs a player contributes above a “replacement level” player. Batters create runs; pitchers save runs. But are batters and pitchers who have comparable RAR levels truly equal in value? Pitchers might be considered to have higher value. Saving an additional run is more important than producing an additional run. A pitcher who throws a shutout is guaranteed to win that game, whereas no matter how many runs a batter produces, his team can still lose given poor pitching support. 28 2012 Baseball Forecaster / Encyclopedia of Fanalytics To calculate RAR for pitchers: 1. Start with the replacement level league ERA. 2. Subtract the pitcher’s ERA. (To calculate projected RAR, use the pitcher’s xERA.) 3. Multiply by number of games played, calculated as plate appearances (IP x 4.34) divided by 38. 4. Multiply the resulting RAR level by 1.08 to account for the variance between earned runs and total runs. t 1JUDIFST XIP BMSFBEZ IBWF HPPE TLJMMT TIPXO CZ BO xERA in the prior year of 4.25 or less. t 1JUDIFST DPNJOH PČ PG BU MFBTU B QBSUJBM TFBTPO JO UIF majors without a major health problem. t 5P UIF FYUFOU UIBU POF TQFDVMBUFT PO QJUDIFST XIP BSF one skill away, look for pitchers who only need to improve their control (bb/9). Overall pitching breakout profile (Brandon Kruse) A breakout performance is defined here as one where a player posts a Rotisserie value of $20 or higher after having never posted a value of $10 previously. These criteria are primarily used to validate an apparent breakout in the current season but may also be used carefully to project a potential breakout for an upcoming season. t "HFPSZPVOHFS t .JOJNVN%PN$NEISBOE#17 t .BYJNVNIJUSBUF t .JOJNVNTUSBOESBUF t 4UBSUFST TIPVME IBWF B IJU SBUF OP HSFBUFS UIBO UIF previous year’s hit rate. Relievers should show improved command t .BYJNVNY&3"PG Handedness 1 LHers tend to peak about a year after RHers. 2. LHers post only 15% of the total saves. Typically, LHers are reserved for specialist roles so few are frontline closers. 3. RHers have slightly better command and HR rate. 4. There is no significant variance in ERA. 5. On an overall skills basis, RHers have about a 6% advantage. Optimal ages As with batters, each pitcher develops at a different pace, but a look at their age can help determine where they should be along the developmental curve. Tendencies: While peaks vary, most all pitchers (who are still around) tend to experience a sharp drop-off in their skills at age 38. Thirtysomethings (Ed Spaulding): Older pitchers, as they lose velocity and movement on the ball, must rely on more variety and better location. Thus, if strikeouts are a priority, you don’t want many pitchers over 30. The over-30 set that tends to be surprising includes finesse types, career minor leaguers who break through for 2-3 seasons often in relief, and knuckleballers (a young knuckleballer is 31). The first 20-save season for a relief ace arrives at about age 26. About three of every four relievers who begin a run of 20-save seasons in their 20s will likely sustain that level for about four years, with their value beginning to decline at the beginning of the third year. Many aces achieve a certain level of maturity in their 30s and can experience a run of 20-save seasons between ages 33 and 36. For some, this may be their first time in the role of bullpen closer. However, those who achieve their first 20-save season after age 34 are unlikely to repeat. Career year drop-off (Rick Wilton) Research shows that a pitcher’s post-career year drop-off, on average, looks like this: t &3"JODSFBTFTCZ t 8)*1JODSFBTFTCZ t /FBSMZGFXFSXJOT Closers Saves There are six events that need to occur in order for a relief pitcher to post a single save: 1. The starting pitcher and middle relievers must pitch well. 2. The offense must score enough runs. 3. It must be a reasonably close game. 4. The manager must put the pitcher in for a save opportunity. 5. The pitcher must pitch well and hold the lead. 6. The manager must let him finish the game. Of these six events, only one is within the control of the relief pitcher. As such, projecting saves for a reliever has little to do with skill and a lot to do with opportunity. However, pitchers with excellent skills may create opportunity for themselves. Catchers’ effect on pitching (Thomas Hanrahan) A typical catcher handles a pitching staff better after having been with a club for a few years. Research has shown that there is an improvement in team ERA of approximately 0.37 runs from a catcher’s rookie season to his prime years with a club. Expect a pitcher’s ERA to be higher than expected if he is throwing to a rookie backstop. Saves conversion rate (Sv%) 6DYHV6DYH2SSRUWXQLWLHV The percentage of save opportunities that are successfully converted. BENCHMARK: We look for a minimum 80% for long-term success. First productive season (Michael Weddell) To find those starting pitchers who are about to post their first productive season in the majors (10 wins, 150 IP, ERA of 4.00 or less), look for: t 1JUDIFST FOUFSJOH UIFJS BHF TFBTPOT FTQFDJBMMZ those about to pitch their age 25 season. 29 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Leverage index (LI) (Tom Tango) Are the minor leagues a place to look at all? From 1990-2004, there were 280 twenty-save seasons in Double-A and Triple-A, accomplished by 254 pitchers. Of those 254, only 46 ever made it to the majors at all. Of those 46, only 13 ever saved 20 games in a season. Of those 13, only 5 ever posted more than one 20-save season in the majors: John Wetteland, Mark Wohlers, Ricky Bottalico, Braden Looper and Francisco Cordero. Five out of 254 pitchers, over 15 years—a rate of 2%. One of the reasons that minor league closers rarely become major league closers is because, in general, they do not get enough innings in the minors to sufficiently develop their arms into bigleague caliber. In fact, organizations do not look at minor league closing performance seriously, assigning that role to pitchers who they do not see as legitimate prospects. The average age of minor league closers over the past decade has been 27.5. Leverage index measures the amount of swing in the possible change in win probability indexed against an average value of 1.00. Thus, relievers who come into games in various situations create a composite score and if that average score is higher than 1.00, then their manager is showing enough confidence in them to try to win games with them. If the average score is below 1.00, then the manager is using them, but not showing nearly as much confidence that they can win games. Saves chances and wins (Craig Neuman) Should the quality of a pitcher’s MLB team be a consideration in drafting a closer? One school of thought says that more wins means more save opportunities. The flipside is that when poor teams win they do so by a small margin, which means more save opportunities. A six-season correlation yielded these results for saves, save opportunities, save percentage, wins, quality starts and run differential. (Any value above .50 suggests at least a moderate correlation.) 69 692 : 6 5' 46 6Y 6Y2 : 6Y 5' Elements of saves success The task of finding future closing potential comes down to looking at two elements: Talent: The raw skills to mow down hitters for short periods of time. Optimal BPVs over 100, but not under 75. Opportunity: The more important element, yet the one that pitchers have no control over. There are pitchers that have Talent, but not Opportunity. These pitchers are not given a chance to close for a variety of reasons (e.g. being blocked by a solid front-liner in the pen, being lefthanded, etc.), but are good to own because they will not likely hurt your pitching staff. You just can’t count on them for saves, at least not in the near term. There are pitchers that have Opportunity, but not Talent. MLB managers decide who to give the ball to in the 9th inning based on their own perceptions about what skills are required to succeed, even if those perceived “skills” don’t translate into acceptable BPI levels. Those pitchers without the BPIs may have some initial shortterm success, but their long-term prognosis is poor and they are high risks to your roster. Classic examples of the short life span of these types of pitchers include Matt Karchner, Heath Slocumb, Ryan Kohlmeier, Dan Miceli and Danny Kolb. More recent examples include Joe Borowski and Fernando Rodney. 46 Saves do correlate with wins. As for the theory that teams who play in close games would accumulate more saves, the low correlation between saves and run differential seems to dispel that a bit. On average, teams registered one save for every two wins. However, there is a relationship between wins and the number of saves per win a team achieves: :LQ7RWDO ! 6DYHV:LQ Teams with fewer wins end up with more saves per win. So, when poor teams do win games, they are more likely to get a save in their wins. Origin of closers Closers’ job retention (Michael Weddell) History has long maintained that ace closers are not easily recognizable early on in their careers, so that every season does see its share of the unexpected. Javy Guerra, Mark Melancon, Fernando Salas, Sergio Santos... who would have thought it a year ago? Accepted facts, all of which have some element of truth: t :PVDBOOPUĕOENBKPSMFBHVFDMPTFSTGSPNQJUDIFSTXIP were closers in the minors. t $MPTFSTCFHJOUIFJSDBSFFSTBTTUBSUFST t $MPTFSTBSFDPOWFSUFETFUVQNFO t $MPTFSTBSFQJUDIFSTXIPXFSFVOBCMFUPEFWFMPQBUIJSE effective pitch. More simply, closers are a product of circumstance. Of pitchers with 20 or more saves in one year, only 67.5% of these closers earned 20 or more saves the following year. The variables that best predicted whether a closer would avoid this attrition: t Saves history: Career saves was the most important factor. t Age: Closers are most likely to keep their jobs at age 27. For long-time closers, their growing career saves totals more than offset the negative impact of their advanced ages. Older closers without a long history of racking up saves tend to be bad candidates for retaining their roles. t Performance: Actual performance, measured by ERA+, was of only minor importance. 30 2012 Baseball Forecaster / Encyclopedia of Fanalytics t Being right-handed: Increased the odds of retaining the closer’s role by 9% over left-handers. How well can we predict which closers will keep their jobs? Of the 10 best closers during 1989-2007, 90% saved at least 20 games during the following season. Of the 10 worst bets, only 20% saved at least 20 games the next year. Though 20-saves success with a 75+ BPV is only a 66% percentage play in any given year, the below-75 group is composed of closers who are rarely able to repeat the feat in the following season: <HDU Closer volatility history <HDU 'UDIWHG 1XPEHURI&ORVHUV $YJ5 )DLOHG 1HZ6RXUFHV Projecting holds (Doug Dennis) Here are some general rules of thumb for identifying pitchers who might be in line to accumulate holds. The percentages represent the portion of 2003’s top holds leaders who fell into the category noted. 1. Left-handed set-up men with excellent BPIs. (43%) 2. A “go-to” right-handed set-up man with excellent BPIs. This is the one set-up RHer that a manager turns to with a small lead in the 7th or 8th innings. These pitchers also tend to vulture wins. (43%, but 6 of the top 9) 3. Excellent BPIs, but not a firm role as the main LHed or RHed set-up man. Roles change during the season; cream rises to the top. Relievers projected to post great BPIs often overtake lesser set-up men during the season. (14%) Reliever efficiency percent (REff%) :LQV6DYHV+ROGV:LQV/RVVHV6DYH2SSV+ROGV This is a measure of how often a reliever contributes positively to the outcome of a game. A record of consistent, positive impact on game outcomes breeds managerial confidence, and that confidence could pave the way to save opportunities. For those pitchers suddenly thrust into a closer’s role, this formula helps gauge their potential to succeed based on past successes in similar roles. BENCHMARK: Minimum of 80%. BPV as a leading indicator (Doug Dennis) Research has shown that base performance value (BPV) is an excellent indicator of long-term success as a closer. Here are 20-plus saves seasons, by year: <HDU 0($1 1R 1RZKRIROORZHGXS VDYHV%39 Other Relievers Drafted refers to the number of saves sources purchased in both LABR and Tout Wars experts leagues each year. These only include relievers drafted for at least $10, specifically for saves speculation. Avg R$ refers to the average purchase price of these pitchers in the AL-only and NL-only leagues. Failed is the number (and percentage) of saves sources drafted that did not return at least 50% of their value that year. The failures include those that lost their value due to ineffectiveness, injury or managerial decision. New Sources are arms that were drafted for less than $10 (if drafted at all) but finished with at least double-digit saves. The failed saves investments in 2011 were Jonathan Broxton, Matt Capps, Frank Francisco, Ryan Franklin, Kevin Gregg, Brad Lidge, Brandon Lyon, Jake McGee, Joe Nathan, Fernando Rodney, Joakim Soria and Matt Thornton. The new sources in 2011 were Kyle Farnsworth, Javy Guerra, Craig Kimbrel, Ryan Madson, Mark Melancon, Fernando Salas, Sergio Santos and Jordan Walden. 1RZLWK %39 Vulture A pitcher, typically a middle reliever, who accumulates an unusually high number of wins by preying on other pitchers’ misfortunes. More accurately, this is a pitcher typically brought into a game after a starting pitcher has put his team behind, and then pitches well enough and long enough to allow his offense to take the lead, thereby “vulturing” a win from the starter. _%39_ 31 Encyclopedia of Fanalytics / 2012 Baseball Forecaster In-Season Analysis Skill versus consistency Two pitchers have identical 4.50 ERAs and identical 3.0 PQS averages. Their PQS logs look like this: Pure Quality Starts We’ve always approached performance measures on an aggregate basis. Each individual event that our statistics chronicle gets dumped into a huge pool of data. We then use our formulas to try to sort and slice and manipulate the data into more usable information. Pure Quality Starts (PQS) take a different approach. It says that the smallest unit of measure should not be the “event” but instead be the “game.” Within that game, we can accumulate all the strikeouts, hits and walks, and evaluate that outing as a whole. After all, when a pitcher takes the mound, he is either “on” or “off ” his game; he is either dominant or struggling, or somewhere in between. In PQS, we give a starting pitcher credit for exhibiting certain skills in each of his starts. Then by tracking his “PQS Score” over time, we can follow his progress. A starter earns one point for each of the following criteria: 1. The pitcher must go a minimum of 6 innings. This measures stamina. If he goes less than 5 innings, he automatically gets a total PQS score of zero, no matter what other stats he produces. 2. He must allow no more than an equal number of hits to the number of innings pitched. This measures hit prevention. 3. His number of strikeouts must be no fewer than two less than his innings pitched. This measures dominance. 4. He must strike out at least twice as many batters as he walks. This measures command. 5. He must allow no more than one home run. This measures his ability to keep the ball in the park. A perfect PQS score is 5. Any pitcher who averages 3 or more over the course of the season is probably performing admirably. The nice thing about PQS is it allows you to approach each start as more than an all-or-nothing event. Note the absence of earned runs. No matter how many runs a pitcher allows, if he scores high on the PQS scale, he has hurled a good game in terms of his base skills. The number of runs allowed—a function of not only the pitcher’s ability but that of his bullpen and defense—will tend to even out over time. It doesn’t matter if a few extra balls got through the infield, or the pitcher was given the hook in the fourth or sixth inning, or the bullpen was able to strand their inherited baserunners. When we look at performance in the aggregate, those events do matter, and will affect a pitcher’s BPIs and ERA. But with PQS, the minutia is less relevant than the overall performance. In the end, a dominating performance is a dominating performance, whether Josh Beckett is hurling a 1-hit shutout or giving up 4 runs while striking out 8 in 7 IP. And a disaster is still a disaster, whether Brian Duensing gets a second inning hook after giving up 5 runs, or “takes one for the team” getting shelled for 7 runs in 4.2 IP. 3,7&+(5$ 3,7&+(5% Which pitcher would you rather have on your team? The risk-averse manager would choose Pitcher A as he represents the perfectly known commodity. Many fantasy leaguers might opt for Pitcher B because his occasional dominating starts show that there is an upside. His Achilles Heel is inconsistency—he is unable to sustain that high level. Is there any hope for Pitcher B? t *G B QJUDIFST JODPOTJTUFODZ JT DIBSBDUFSJ[FE CZ NPSF poor starts than good starts, his upside is limited. t 1JUDIFST XJUI FYUSFNF JODPOTJTUFODZ SBSFMZ HFU B GVMM season of starts. t )PXFWFSJODPOTJTUFODZJTOFJUIFSDISPOJDOPSGBUBM The outlook for Pitcher A is actually worse. Disaster avoidance might buy these pitchers more starts, but history shows that the lack of dominating outings is more telling of future potential. In short, consistent mediocrity is bad. PQS DOMination and DISaster rates (Gene McCaffrey) DOM% is the percentage of a starting pitcher’s outings that rate as a PQS-4 or PQS-5. DIS% is the percentage that rate as a PQS-0 or PQS-1. DOM/DIS percentages open up a new perspective, providing us with two separate scales of performance. In tandem, they measure consistency. PQS ERA (qERA) A pitcher’s DOM/DIS split can be converted back to an equivalent ERA. By creating a grid of individual DOM% and DIS% levels, we can determine the average ERA at each cross point. The result is an ERA based purely on PQS. Quality/consistency score (QC) '20[',6[ Using PQS and DOM/DIS percentages, this score measures both the quality of performance as well as start-to-start consistency. PQS correlation with Quality Starts (Paul Petera) 346 46 Forward-looking PQS (John Burnson) PQS says whether a pitcher performed ably in a past start—it doesn’t say anything about how he’ll do in the next start. We built a version of PQS that attempts to do that. For each series of five starts for a pitcher, we looked at his average IP, K/9, HR/9, H/9, and K/BB, and then whether the pitcher won his next start. We catalogued the results by indicator and calculated the observed future winning percentage for each data point. 32 2012 Baseball Forecaster / Encyclopedia of Fanalytics This research suggested that a forward-looking version of PQS should have these criteria: t ćFQJUDIFSNVTUIBWFMBTUFEBUMFBTUJOOJOHT t )FNVTUIBWFSFDPSEFEBUMFBTU*1TUSJLFPVUT t )FNVTUIBWFBMMPXFE[FSPIPNFSVOT t )FNVTUIBWFBMMPXFEOPNPSFIJUTUIBO*1 t )FNVTUIBWFIBEB$PNNBOE,## PGBUMFBTU of high quality. In fact, once a pitcher has posted six DOM starts in a row, there is greater than a 70% probability that the streak will continue. When it does end, there is less than a 10% probability that the streak-breaker is going to be a DISaster. Once a pitcher enters into a DIS streak of any length, the probability is that his next start is going to be below average, even if it breaks the streak. However, DIS streaks end quickly. Once a pitcher hits the skids, odds are low for him to start posting good numbers in the short term, though the duration of the plummet itself should be brief. Pure Quality Relief (Patrick Davitt) A system for evaluating reliever outings. The scoring : 1. Two points for the first out, and one point for each subsequent out, to a maximum of four points. 2. One point for having at least one strikeout for every four full outs (one K for 1-4 outs, two Ks for 5-8 outs, etc.). 3. One point for zero baserunners, minus one point for each baserunner, though allowing the pitcher one unpenalized runner for each three full outs (one baserunner for 3-5 outs, two for 6-8 outs, three for nine outs) 4. Minus one point for each earned run, though allowing one ER for 8– or 9-out appearances. 5. An automatic PQR-0 for allowing a home run. 5-game PQS predictability (Bill Macey) *DPHDYJ346 $YJ346 /HVVWKDQ %HWZHHQDQG %HWZHHQDQG %HWZHHQDQG RUJUHDWHU '20 ',6 Pitchers with higher PQS scores in their previous 5 starts tended to pitch better in their next start. But the relative parity of subsequent DOM and DIS starts for all but the hottest of streaks warn us not to put too much effort into predicting any given start. That more than a quarter of pitchers who had been awful over their previous 5 starts still put up a dominating start next shows that anything can happen in a single game. Avoiding relief disasters (Ed DeCaria) Relief disasters (defined as ER>=3 and IP<=3), occur in 5%+ of all appearances. The chance of a disaster exceeds 13% in any 7-day period. To minimize the odds of a disaster, we created a model that produced the following list of factors, in order of influence: 1. Strength of opposing offense 2. Park factor of home stadium 3. BB/9 over latest 31 days (more walks is bad) 4. Pitch count over previous 7 days (more pitches is bad) 5. Latest 31 Days ERA>xERA (recent bad luck continues) Daily league owners who can slot relievers by individual game should also pay attention to days of rest: pitching on less rest than one is accustomed to increases disaster risk. Pitch counts as a leading indicator Long-term analysis of workload is an ongoing science. However, there have also been questions whether we can draw any conclusions from short-term trends. For this analysis, all pitching starts from 2005-2006 were isolated—looking at pitch counts and PQS scores—and compared side-by-side with each pitcher’s subsequent outing. We examined two-start trends, the immediate impact that the length of one performance would have on the next start. 3LWFK&W Sample size reliability (Russell Carleton) At what point during the season do statistics become reliable indicators of skill? Measured in batters faced: 150: K/PA, ground ball rate, line drive rate 200: Fly ball rate, GB/FB 500: K/BB 550: BB/PA Unlisted stats did not stabilize over a full season of play. (Note that 150 BF is roughly equivalent to six outings for a starting pitcher; 550 BF would be 22 starts, etc.) 3FW 1H[W6WDUW 346 '20 ',6 T(5$ There does appear to be merit to the concern over limiting hurlers to 120 pitches per start. The research shows a slight dropoff in performance in those starts following a 120+ pitch outing. However, the impact does not appear to be all that great and the fallout might just affect those pitchers who have no business going that deep into games anyway. Additional detail to this research (not displayed) showed that higher-skilled pitchers were more successful throwing over 120 pitches but less-skilled pitchers were not. Pitching streaks It is possible to find predictive value in strings of DOMinating (PQS 4/5) or DISaster (PQS 0/1) starts: Once a pitcher enters into a DOM streak of any length, the probability is that his subsequent start is going to be a betterthan-average outing. The further a player is into a DOM streak, the higher the likelihood that the subsequent performance will be Days of rest as a leading indicator Workload is only part of the equation. The other part is how often a pitcher is sent out to the mound. For instance, it’s possible that a hurler might see no erosion in skill after a 120+ pitch outing if he had enough rest between starts: 33 Encyclopedia of Fanalytics / 2012 Baseball Forecaster 3,7&+&28176 7KUHHGD\VUHVW )RXU'D\VUHVW )LYH'D\VUHVW 6L['D\VUHVW 'D\VUHVW 1(;767$57 '20 ',6 3FW 346 Chaconian Having the ability to post many saves despite sub-Mendoza BPIs and an ERA in the stratosphere. T(5$ Vintage Eck Territory A BPV greater than 200, a level achieved by Dennis Eckersley for four consecutive years. Edwhitsonitis A dreaded malady marked by the sudden and unexplained loss of pitching ability upon a trade to the New York Yankees. ERA Benchmark A half run of ERA over 200 innings comes out to just one earned run every four starts. Gopheritis (also, Acute Gopheritis and Chronic Gopheritis) The dreaded malady in which a pitcher is unable to keep the ball in the park. Pitchers with gopheritis have a FB rate of at least 40%. More severe cases have a FB% over 45%. The Knuckleballers Rule: Knuckleballers don’t follow no stinkin’ rules. Managers are reluctant to put a starter on the mound with any fewer than four days rest, and the results for those who pitched deeper into games shows why. Four days rest is the most common usage pattern and even appears to mitigate the drop-off at 120+ pitches. Perhaps most surprising is that an extra day of rest improves performance across the board and squeezes even more productivity out of the 120+ pitch outings. Performance begins to erode at six days (and continues at 7-20 days, though those are not displayed). The 20+ Days chart represents pitchers who were primarily injury rehabs and failed call-ups, and the length of the “days rest” was occasionally well over 100 days. This chart shows the result of their performance in their first start back. The good news is that the workload was limited for 85% of these returnees. The bad news is that these are not pitchers you want active. So for those who obsess over getting your DL returnees activated in time to catch every start, the better percentage play is to avoid that first outing. Brad Lidge Lament When a closer posts a 62% strand rate, he has nobody to blame but himself. LOOGY (Lefty One Out GuY) A left-handed reliever whose job it is to get one out in important situations. Vin Mazzaro Vindication Occasional nightmares (2.1 innings, 14 ER) are just a part of the game. Meltdown Any game in which a starting pitcher allows more runs than innings pitched. Lance Painter Lesson Six months of solid performance can be screwed up by one bad outing. (In 2000, Painter finished with an ERA of 4.76. However, prior to his final appearance of the year—in which he pitched 1 inning and gave up 8 earned runs—his ERA was 3.70.) The Five Saves Certainties 1. On every team, there will be save opportunities and someone will get them. At a bare minimum, there will be at least 30 saves to go around, and not unlikely more than 45. 2. Any pitcher could end up being the chief beneficiary. Bullpen management is a fickle endeavor. 3. Relief pitchers are often the ones that require the most time at the start of the season to find a groove. The weather is cold, the schedule is sparse and their usage is erratic. 4. Despite the talk about “bullpens by committee,” managers prefer a go-to guy. It makes their job easier. 5. As many as 50% of the saves in any year will come from pitchers who are unselected at the end of Draft Day. Other Diamonds The Pitching Postulates 1. Never sign a soft-tosser to a long-term contract. 2. Right-brain dominance has a very long shelf life. 3. A fly ball pitcher who gives up many HRs is expected. A GB pitcher who gives up many HRs is making mistakes. 4. Never draft a contact fly ball pitcher who plays in a hitter’s park. 5. Only bad teams ever have a need for an inning-eater. 6. Never chase wins. Ricky Bones List Pitchers with BPIs so incredibly horrible that you have to wonder how they can possibly draw a major league paycheck year after year. (Given that Bones was named in the Mitchell Report for taking steroids/HGH and was a pitching coach in the Arizona Fall League—think about those two facts for a second—we may need to rename this the Dontrelle Willis List.) Soft-tosser A pitcher with a strikeout rate of 5.5 or less. Soft-tosser land The place where feebler arms leave their fortunes in the hands of the defense, variable hit and strand rates, and park dimensions. It’s a place where many live, but few survive. 34 2012 Baseball Forecaster / Encyclopedia of Fanalytics Prospects 11. Decide where your priorities really are. If your goal is to win, prospect analysis is just a part of the process, not the entire process. General Factors affecting minor league stats (Terry Linhart) 1. Often, there is an exaggerated emphasis on short-term performance in an environment that is supposed to focus on the long-term. Two poor outings don’t mean a 21-year-old pitcher is washed up. 2. Ballpark dimensions and altitude create hitters parks and pitchers parks, but a factor rarely mentioned is that many parks in the lower minors are inconsistent in their field quality. Minor league clubs have limited resources to maintain field conditions, and this can artificially depress defensive statistics while inflating stats like batting average. 3. Some players’ skills are so superior to the competition at their level that you can’t get a true picture of what they’re going to do from their stats alone. 4. Many pitchers are told to work on secondary pitches in unorthodox situations just to gain confidence in the pitch. The result is an artificially increased number of walks. 5. The #3, #4, and #5 pitchers in the lower minors are truly longshots to make the majors. They often possess only two pitches and are unable to disguise the off-speed offerings. Hitters can see inflated statistics in these leagues. Minor league prospecting in perspective In our perpetual quest to be the genius who uncovers the next Albert Pujols when he is in A-ball, there is an obsessive fascination with minor league prospects. That’s not to say that prospecting is not important. The issue is perspective: 1. During the 10 year period of 1996 to 2005, only 8% of players selected in the first round of the Major League Baseball First Year Player Draft went on to become stars. 2. Some prospects are going to hit the ground running (Ryan Braun) and some are going to immediately struggle (Alex Gordon), no matter what level of hype follows them. 3. Some prospects are going to start fast (since the league is unfamiliar with them) and then fade (as the league figures them out). Others will start slow (since they are unfamiliar with the opposition) and then improve (as they adjust to the competition). So if you make your free agent and roster decisions based on small early samples sizes, you are just as likely to be an idiot as a genius. 4. How any individual player will perform relative to his talent is largely unknown because there is a psychological element that is vastly unexplored. Some make the transition to the majors seamlessly, some not, completely regardless of how talented they are. 5. Still, talent is the best predictor of future success, so major league equivalent base performance indicators still have a valuable role in the process. As do scouting reports, carefully filtered. 6. Follow the player’s path to the majors. Did he have to repeat certain levels? Was he allowed to stay at a level long enough to learn how to adjust to the level of competition? A player with only two great months at Double-A is a good bet to struggle if promoted directly to the majors because he was never fully tested at Double-A, let alone Triple-A. 7. Younger players holding their own against older competition is a good thing. Older players reaching their physical peak, regardless of their current address, can be a good thing too. The R.A. Dickeys and Ryan Ludwicks can have some very profitable years. 8. Remember team context. A prospect with superior potential often will not unseat a steady but unspectacular incumbent, especially one with a large contract. 9. Don’t try to anticipate how a team is going to manage their talent, both at the major and minor league level. You might think it’s time to promote Bryce Harper and give him an everyday role. You are not running the Nationals. 10. Those who play in shallow, one-year leagues should have little cause to be looking at the minors at all. The risk versus reward is so skewed against you, and there is so much talent available with a track record, that taking a chance on an unproven commodity makes little sense. Minor league level versus age When evaluating minor leaguers, look at the age of the prospect in relation to the median age of the league he is in: /RZOHYHO$ %HWZHHQ 8SSHUOHYHO$ $URXQG 'RXEOH$ 7ULSOH$ These are the ideal ages for prospects at the particular level. If a prospect is younger than most and holds his own against older and more experienced players, elevate his status. If he is older than the median, reduce his status. Triple-A experience as a leading indicator The probability that a minor leaguer will immediately succeed in the majors can vary depending upon the level of Triple-A experience he has amassed at the time of call-up. %$77(56 <U )XOO 3HUIRUPHGZHOO 3HUIRUPHGSRRUO\ QGKDOIGURSRII 3,7&+(56 <U )XOO The odds of a batter achieving immediate MLB success was slightly more than 50-50. More than 80% of all pitchers promoted with less than a full year at Triple-A struggled in their first year in the majors. Those pitchers with a year in Triple-A succeeded at a level equal to that of batters. 35 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Major League Equivalency (MLE) in Triple-A is just as likely to peak at age 27 as a major leaguer of the same age. The question becomes one of opportunity—will the parent club see fit to reap the benefits of that peak performance? Prospecting these players for your fantasy team is, admittedly, a high risk endeavor, though there are some criteria you can use. Look for a player who is/has: t 0QUJNBMMZ BHF GPS PWFSBMM QFBL TLJMMT BHF for power skills, or age 28-31 for pitchers. t "U MFBTU UXP TFBTPOT PG FYQFSJFODF BU 5SJQMF" $BSFFS Double-A players are generally not good picks. t 4PMJECBTFTLJMMTMFWFMT t 4IBMMPXPSHBOJ[BUJPOBMEFQUIBUUIFJSQPTJUJPO t /PUBCMFXJOUFSMFBHVFPSTQSJOHUSBJOJOHQFSGPSNBODF Players who meet these conditions are not typically draftable players, but worthwhile reserve or FAAB picks. (Bill James) A formula that converts a player’s minor or foreign league statistics into a comparable performance in the major leagues. These are not projections, but conversions of current performance. MLEs contain adjustments for the level of play in individual leagues and teams. They work best with Triple-A stats, not quite as well with Double-A stats, and hardly at all with the lower levels. Foreign conversions are still a work in process. James’ original formula only addressed batting. Our research has devised conversion formulas for pitchers, however, their best use comes when looking at BPIs, not traditional stats. Adjusting to the competition All players must “adjust to the competition” at every level of professional play. Players often get off to fast or slow starts. During their second tour at that level is when we get to see whether the slow starters have caught up or whether the league has figured out the fast starters. That second half “adjustment” period is a good baseline for projecting the subsequent season, in the majors or minors. Premature major league call-ups often negate the ability for us to accurately evaluate a player due to the lack of this adjustment period. For instance, a hotshot Double-A player might open the season in Triple-A. After putting up solid numbers for a month, he gets a call to the bigs, and struggles. The fact is, we do not have enough evidence that the player has mastered the Triple-A level. We don’t know whether the rest of the league would have caught up to him during his second tour of the league. But now he’s labeled as an underperformer in the bigs when in fact he has never truly proven his skills at the lower levels. Batters MLE PX as a leading indicator (Bill Macey) Looking at minor league performance (as MLE) in one year and the corresponding MLB performance the subsequent year: 2EVHUYDWLRQV 0HGLDQ3; 3HUFHQW3;! <HDU0/( <HDU0/% In addition, 53% of the players had a MLB PX in year 2 that exceeded their MLE PX in year 1. A slight bias towards improved performance in year 2 is consistent with general career trajectories. Rookie playing time Weaker-performing teams have historically (1976-2009) been far more dependent on debut rookies than stronger-performing teams. This makes sense, as non-contenders have greater incentive and flexibility to allow young players to gain experience at the MLB level. Additionally, individual player characteristics can provide clues as to which debut rookies will earn significant PA or IP: t 3PPLJFTXIPDBOQMBZVQUIFNJEEMF$'#44 PSPO the left side (LF, 3B) are likely to see more debut playing time than those at 1B, RF, C, or DH. t -) CBUUFST BOE QJUDIFST BSF TMJHIUMZ NPSF MJLFMZ UIBO righties to earn significant PA or IP in their debuts. t 3PPLJFTVOEFSBHFFBSOOFBSMZUXJDFUIF1"TPGUIPTF aged 25-26 in their debut season. Pitchers under age 23 earn twice the innings of those aged 26-27. <HDU 0/(3; ! <HDU 0/%3; 3FW ,QFU 3FW0/% 3;! Slicing the numbers by performance level, there is a good amount of regression to the mean. Players rarely suddenly develop power at the MLB level if they didn’t previously display that skill at the minor league level. However, the relatively large gap between the median MLE PX and MLB PX for these players, 125 to 110, confirms the notion that the best players continue to improve once they reach the major leagues. MLE contact rate as a leading indicator (Bill Macey) There is a strong positive correlation (0.63) between a player’s MLE ct% in Year 1 and his actual ct% at the MLB level in Year 2. Bull Durham prospects 0/(FW 727$/ There is some potential talent in older players—age 26, 27 or higher—who, for many reasons (untimely injury, circumstance, bad luck, etc.), don’t reach the majors until they have already been downgraded from prospect to suspect. Equating potential with age is an economic reality for major league clubs, but not necessarily a skills reality. Skills growth and decline is universal, whether it occurs at the major league level or in the minors. So a high-skills journeyman 36 <HDU 0/(FW <HDU 0/%FW 2012 Baseball Forecaster / Encyclopedia of Fanalytics There is very little difference between the median MLE BA in Year 1 and the median MLB BA in Year 2: 0/(FW 727$/ <HDU 0/(%$ There is some erosion, but a 0.26 run rise in qERA is hardly cause for panic. If we re-focus our study class, the qERA variance increased to 4.44-5.08 for those who made at least 16 starts before July 31. The variance also was larger (3.97-4.56) for those who had a PQS-3 average prior to July 31. The pitchers who intersected these two sub-groups: <HDU 0/%%$ 346!*6! EHIRUH DIWHU )DUHGSRRUO\ The data used here were MLE levels from the previous two years, not the season in which they were called up. The significance? Solid current performance is what merits a call-up, but this is not a good indicator of short-term MLB success, because a) the performance data set is too small, typically just a few month’s worth of statistics, and b) for those putting up good numbers at a new minor league level, there has typically not been enough time for the scouting reports to make their rounds. (Al Melchior) There is a link between minor league skill and how a pitching prospect will fare in his first 5 starts upon call-up. 0/(%39 Pitchers who demonstrate sub-par skills in the minors (sub-50 BPV) tend to fare poorly in their first big league starts. Threefifths of these pitchers register a PQS average below 2.0, while only 8% average over 3.0. Fewer than 1 out of 5 minor leaguers with a 100+ MLE BPV go on to post a sub-2.0 PQS average in their initial major league starts, but more than half average 3.0 or better. Late season performance of rookie starting pitchers (Ray Murphy) Given that a rookie’s second tour of the league provides insight as to future success, do rookie pitchers typically run out of gas? We studied 2002-2005, identified 56 rookies who threw at least 75 IP and analyzed their PQS logs. The group: $OOURRNLHV EHIRUH DIWHU *63 '20 ',6 T(5$ The Japanese major leagues are generally considered to be equivalent to very good Triple-A ball and the pitching may be even better. However, statistics are difficult to convert due to differences in the way the game is played in Japan. 1. While strong on fundamentals, Japanese baseball’s guiding philosophy is risk avoidance. Mistakes are not tolerated. Runners rarely take extra bases, batters focus on making contact rather than driving the ball, and managers play for one run at a time. As a result, offenses score fewer runs than they should given the number of hits. Pitching stats tend to look better than the talent behind them. 2. Stadiums in Japan usually have shorter fences. Normally this would mean more HRs, but given #1 above, it is the American players who make up the majority of Japan’s power elite. Power hitters do not make an equivalent transition to the MLB. 3. There are more artificial turf fields, which increases the number of ground ball singles. Only a few stadiums have infield grass and some still use dirt infields. 4. The quality of umpiring is questionable; there are no sanctioned umpiring schools in Japan. Fewer errors are called, reflecting the cultural philosophy of low tolerance for mistakes and the desire to avoid publicly embarrassing a player. Moreover, umpires are routinely intimidated. 5. Teams have smaller pitching staffs, sometimes no more than about seven deep. Three-man pitching rotations are not uncommon and the best starters often work out of the pen between starts. Despite superior conditioning, Japanese pitchers tend to burn out early due to overuse. 6. Japan has finally standardized its baseball, to a “noncarrying” ball with less elastic rubber surrounding the cork, which decreases offense and inflates pitching stats. 7. Tie games are allowed. If the score remains even after 12 innings, the game goes into the books as a tie. The percentage of hurlers that were good investments in the year that they were called up varied by the level of their historical minor league BPIs prior to that year. 346$YJ ',6 Comparing MLB and Japanese Besuboru BPIs as a leading indicator for pitching success Minor league BPV as a leading indicator '20 Japanese Baseball (Tom Mulhall) Pitchers )DUHGZHOO *63 While the sample size is small, the degree of flameout by these guys (0.85 runs) is more significant. Excluding the <70% cohort (which was a tiny sample size), there is a positive relationship between MLE ct% and MLB BA. 3LWFKHUVZKRKDG *RRGLQGLFDWRUV 0DUJLQDORUSRRULQGLFDWRUV Japanese players as fantasy farm selections Many fantasy leagues have large reserve or farm teams with rules allowing them to draft foreign players before they sign with a MLB team. With increased coverage by fantasy experts, the internet, and exposure from the World Baseball Classic, anyone T(5$ 37 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Some players that are currently stuck at one of the interim steps, and may or may not ever reach Step 10, include Mat Gamel, Max Ramirez and Travis Snider. willing to do a modicum of research can compile an adequate list of good players. However, the key is not to identify the best Japanese players— the key is to identify impact players who have the desire and opportunity to sign with a MLB team. It is easy to overestimate the value of drafting these players. Since 1995, only about three dozen Japanese players have made a big league roster, and about half of them were middle relievers. But for owners who are allowed to carry a large reserve or farm team at reduced salaries, these players could be a real windfall, especially if your competitors do not do their homework. A list of Japanese League players who could jump to the majors appears in the Prospects section. Developmental Dogmata 1. Defense is what gets a minor league prospect to the majors; offense is what keeps him there. (Deric McKamey) 2. The reason why rapidly promoted minor leaguers often fail is that they are never given the opportunity to master the skill of “adjusting to the competition.” 3. Rookies who are promoted in-season often perform better than those that make the club out of spring training. Inferior March competition can inflate the latter group’s perceived talent level. 4. Young players rarely lose their inherent skills. Pitchers may uncover weaknesses and the players may have difficulty adjusting. These are bumps along the growth curve, but they do not reflect a loss of skill. 5. Late bloomers have smaller windows of opportunity and much less chance for forgiveness. 6. The greatest risk in this game is to pay for performance that a player has never achieved. 7. Some outwardly talented prospects simply have a ceiling that’s spelled AAA. Other Diamonds Age 26 Paradox Age 26 is when a player begins to reach his peak skill, no matter what his address is. If circumstances have him celebrating that birthday in the majors, he is a breakout candidate. If circumstances have him celebrating that birthday in the minors, he is washed up. A-Rod 10-Step Path to Stardom Not all well-hyped prospects hit the ground running. More often they follow an alternative path: 1. Prospect puts up phenomenal minor league numbers. 2. The media machine gets oiled up. 3. Prospect gets called up, but struggles, Year 1. 4. Prospect gets demoted. 5. Prospect tears it up in the minors, Year 2. 6. Prospect gets called up, but struggles, Year 2. 7. Prospect gets demoted. 8. The media turns their backs. Fantasy leaguers reduce their expectations. 9. Prospect tears it up in the minors, Year 3. The public shrugs its collective shoulders. 10. Prospect is promoted in Year 3 and explodes. Some lucky fantasy leaguer lands a franchise player for under $5. Rule 5 Reminder Don’t ignore the Rule 5 draft lest you ignore the possibility of players like Jose Bautista, Josh Hamilton, Johan Santana, Joakim Soria, Dan Uggla, Shane Victorino and Jayson Werth. All were Rule 5 draftees. 38 2012 Baseball Forecaster / Encyclopedia of Fanalytics Gaming (1B or 3B), two additional outfielders and a utility player/designated hitter (which often can be a batter who qualifies anywhere). Nine pitchers, typically holding any starting or relief role. Open: 25 players. One at each defensive position (plus DH), 5-man starting rotation and two relief pitchers. Nine additional players at any position, which may be a part of the active roster or constitute a reserve list. 40-man: Standard 23 plus 17 reserves. Used in many keeper and dynasty leagues. Reapportioned: In recent years, new obstacles are being faced by 12-team A.L. leagues and 13-team N.L. leagues thanks to changes in the real game. The 14/9 split between batters and pitchers no longer reflects how MLB teams structure their rosters. Of the 30 teams, each with 25-man rosters, not one contains 14 batters for any length of time. In fact, many spend a good part of the season with only 12 batters, which means teams often have more pitchers than hitters. For fantasy purposes in AL- and NL-only leagues, that leaves a disproportionate draft penetration into the batter and pitcher pools: Standard Rules and Variations Rotisserie Baseball was invented as an elegant confluence of baseball and economics. Whether by design or accident, the result has lasted for three decades. But what would Rotisserie and fantasy have been like if the Founding Fathers knew then what we know now about statistical analysis and game design? You can be sure things would be different. The world has changed since the original game was introduced yet many leagues use the same rules today. New technologies have opened up opportunities to improve elements of the game that might have been limited by the capabilities of the 1980s. New analytical approaches have revealed areas where the original game falls short. As such, there are good reasons to tinker and experiment; to find ways to enhance the experience. Following are the basic elements of fantasy competition, those that provide opportunities for alternative rules and experimentation. This is by no means an exhaustive list, but at minimum provides some interesting food-for-thought. 2Q$/URVWHUV 'UDIWHG$/ 3FW Player pool Standard: American League-only, National League-only or Mixed League. AL typically drafts 8-12 teams (pool penetration of 52% to 79%). NL typically drafts 8-13 teams (46% to 75% penetration). Mixed leagues draft 10-18 teams (31% to 55% penetration), though 15 teams (46%) is a common number. Drafting of reserve players will increase the penetration percentages. A 12-team AL-only league adding 6 reserves onto 23-man rosters would draft 99% of the available pool of players on all AL teams’ 25-man rosters. The draft penetration level determines which fantasy management skills are most important to your league. The higher the penetration, the more important it is to draft a good team. The lower the penetration, the greater the availability of free agents and the more important in-season roster management becomes. There is no generally-accepted optimal penetration level, but we have often suggested that 75% provides a good balance between the skills required for both draft prep and in-season management. Alternative pools: There is a wide variety of options here. Certain leagues draft from within a small group of major league divisions or teams. Some competitions, like home run leagues, only draft batters. Bottom-tier pool: Drafting from the entire Major League population, the only players available are those who posted a Rotisserie dollar value of $5 or less in the previous season. Intended as a test of an owner’s ability to identify talent with upside. Best used as a pick-a-player contest with any number of teams participating. 2Q1/URVWHUV 'UDIWHG1/ 3FW %$77(56 3,7&+(56 These drafts are now depleting about 30% more batters out of the pool than pitchers. Add in those leagues with reserve lists— perhaps an additional 6 players per team removing another 72-78 players—and post-draft free agent pools are very thin, especially on the batting side. The impact is less in 15-team mixed leagues, though the FA pitching pool is still disproportionately deep. 2QDOOURVWHUV 'UDIWHG 3FW %$77(56 3,7&+(56 One solution is to reapportion the number of batters and pitchers that are rostered. Adding one pitcher slot and eliminating one batter slot may be enough to provide better balance. The batting slot most often removed is the second catcher, since it is the position with the least depth. Selecting players Standard: The three most prevalent methods for stocking fantasy rosters are: Snake/Straight/Serpentine draft: Players are selected in order with seeds reversed in alternating rounds. This method has become the most popular due to its speed, ease of implementation and ease of automation. In these drafts, the underlying assumption is that value can be ranked relative to a linear baseline. Pick #1 is better than pick #2, which is better than pick #3, and the difference between each pick is assumed to be somewhat equivalent. While that is a faulty Positional structure Standard: 23 players. One at each defensive position (though three outfielders may be from any of LF, CF or RF), plus one additional catcher, one middle infielder (2B or SS), one corner infielder 39 Encyclopedia of Fanalytics / 2012 Baseball Forecaster assumption, it is one we must make in order to believe in a level playing field. Auction: Players are sold to the highest bidder from a fixed budget, typically $260. Auctions provide the team owner with the most control over which players will be on his team, but can take twice as long as snake drafts. Here, the baseline is $0 at the beginning of each player put up for bid. The final purchase price for each player is shaped by many wildly variable factors, from roster need to geographic location of the draft. A $30 player can mean different things to different drafters. Pick-a-player / Salary cap: Players are assigned fixed dollar values and owners assemble their roster within a fixed cap. This type of roster-stocking is an individual exercise which results in teams typically having some of the same players. In these leagues, the “value” decision is taken out of the hands of the owners. Each player has a fixed value, pre-assigned based on past season performance. Hybrid snake-auction: Each draft begins as an auction. Each team has to fill its first seven roster slots from a budget of $154. Opening bid for any player is $15. This assures that player values will be close to reality. After each team has filled seven slots, it becomes a snake draft. If you like, you can assign fixed salaries to the snake-drafted players in such a way that each roster will still add up to about $260. 5RXQG In general, the more categories you add, the more complicated it is to isolate individual performance and manage the categorical impact on your roster. There is also the danger of redundancy; with multiple categories measuring like stats, certain skills can get over-valued. For instance, home runs are double-counted when using the categories of both HR and slugging average. (Though note that HRs are actually already triple-counted in standard 5x5—HRs, runs, and RBIs) If the goal is to have categories that create a more encompassing picture of player performance, it is actually possible to accomplish more with less: Retro 4x4: HR, (RBI+R-HR), SB, BA, W, Sv, K, ERA. This provides a better balance between batting and pitching in that each has three counting categories and one ratio category. You may also opt to replace BA with OBA. Advanced 4x4: Extra bases: Enhances the excitement of the big hit by adding doubles and triples. Runs Produced (RBI + R – HR): In 5x5, two categories are devoted to situational stats. This combines them. Net SB: Players should not get credit for steals without incurring the penalty for CS. OBA: Walks are far too valuable to ignore. Innings Pitched: A pitcher has to be good and have stamina to run up innings. Net saves: You can’t get credit for the good stuff without also being responsible for the bad. Strikeouts – (HR x 4): A convoluted statistic but it captures the opposite ends of the spectrum of pitching outcomes. WHIP: Oddly enough, the only original category to stay. Essentially, the pitching equivalent of OBA. 6DODU\ Keeping score Standard: These are the most common scoring methods: Rotisserie: Players are evaluated in several statistical categories. Totals of these statistics are ranked by team. The winner is the team with the highest cumulative ranking. Points: Players receive points for events that they contribute to in each game. Points are totaled for each team and teams are then ranked. Head-to-Head (H2H): Using Rotisserie or points scoring, teams are scheduled in weekly matchups. The winner of that week’s match-up is the team that finishes higher in more categories (Rotisserie) or scores the most points. Hybrid H2H-Rotisserie: Rotisserie’s category ranking system can be converted into a weekly won-loss record. Depending upon where your team finishes for that week’s statistics determines how many games you win for that week. Each week, your team will play seven games. You can also use this chart to decide how deep into the rosters you want to auction. If you want to auction the first 15 players, for instance, you’d use a budget of $238. Though not shown, if you only wanted to auction the first 5 players, your budget would be $121. This method is intended to reduce draft time while still providing an economic component for selecting players. Stat categories Standard: The standard statistical categories for Rotisserie leagues are: 4x4: HR, RBI, SB, BA, W, Sv, ERA, WHIP 5x5: HR, R, RBI, SB, BA, W, Sv, K, ERA, WHIP 6x6: Categories typically added are Holds and OPS. 7x7, etc.: Any number of categories may be added. 40 3ODFH 5HFRUG 3ODFH 5HFRUG VW WK QG WK UG WK WK WK WK WK WK WK %DVHGRQRYHUDOO5RWLVVHULHFDWHJRU\UDQNLQJIRUWKHZHHN 2012 Baseball Forecaster / Encyclopedia of Fanalytics An advantage of this second bid is that it gives owners an opportunity to see who they are going up against, and adjust. If you are bidding against an owner close to you in the standings, you may need to be more aggressive in that second bid. If you see that the tied owner(s) wouldn’t hurt you by acquiring that player, then maybe you resubmit the original bid and be content to potentially lose out on the player. If you’re ahead in the standings, it’s actually a way to potentially opt out on that player completely by resubmitting your original bid and forcing another owner to spend his FAAB. Some leagues will balk at adding another layer to the weekly deadline process; it’s a trade-off to having more control over managing your FAAB. At the end of each week, all the statistics revert to zero and you start over. You never dig a hole in any category that you can’t climb out of, because all categories themselves are incidental to the standings. The regular season lasts for 23 weeks, which equals 161 games. Weeks 24, 25 and 26 are for play-offs. Free agent acquisition Standard: Three methods are the most common for acquiring free agent players during the season. First to the phone: Free agents are awarded to the first owner who claims them. Reverse order of standings: Access to the free agent pool is typically in a snake draft fashion with the last place team getting the first pick, and each successive team higher in the standings picking afterwards. Free agent acquisition budget (FAAB): Teams are given a set budget at the beginning of the season (typically, $100 or $1000) from which they bid on free agents in a closed auction process. Vickrey FAAB: Research has shown that more than 50% of FAAB dollars are lost via overbid on an annual basis. Given that this is a scarce commodity, one would think that a system to better manage these dollars might be desirable. The Vickrey system conducts a closed auction in the same way as standard FAAB, but the price of the winning bid is set at the amount of the second highest bid, plus $1. In some cases, gross overbids (at least $10 over) are reduced to the second highest bid plus $5. This method was designed by William Vickrey, a Professor of Economics at Columbia University. His theory was that this process reveals the true value of the commodity. For his work, Vickrey was awarded the Nobel Prize for Economics (and $1.2 million) in 1996. Double-Bid FAAB: One of the inherent difficulties in the current FAAB system is that we have so many options for setting a bid amount. You can bid $47, or $51, or $23. You might agonize over whether to go $38 or $39. With a $100 budget, there are 100 decision points. And while you may come up with a rough guesstimate of the range in which your opponents might bid, the results for any individual player bidding are typically random within that range. The first part of this process reduces the number of decision points. Owners must categorize their interest by bidding a fixed number of pre-set dollar amounts for each player. In a $100 FAAB league, for instance, those levels might be $1, $5, $10, $15, $20, $30, $40 or $50. All owners would set the general market value for free agents in these eight levels of interest. (This system sets a $50 maximum, but that is not absolutely necessary.) The initial stage of the bidding process serves to screen out those who are not interested in a player at the appropriate market level. That leaves a high potential for tied owners, those who share the same level of interest. The tied owners must then submit a second bid of equal or greater value than their first bid. These bids can be in $1 increments. The winning owner gets the player; if there is still a tie, then the player would go to the owner lower in the standings. The season Standard: Leagues are played out during the course of the entire Major League Baseball season. Split-season: Leagues are conducted from Opening Day through the All-Star break, then re-drafted to play from the All-Star break through the end of the season. 50-game split-season: Leagues are divided into three 50-game split-seasons. There would be a one-week break in between each segment. The advantages: t 8JUIEXJOEMJOHBUUFOUJPOTQBOTPWFSUIFMPOHHBNF season, 50 games is a more accessible time frame to maintain interest. There would be fewer abandoned teams. t ćFSFXPVMECFGPVSTIPUTBUBUJUMFFBDIZFBSUIFĕSTU place team from each split, plus the team with the best overall record for the entire year. t (JWFO UIBU ESBęJOH JT DPOTJEFSFE UIF NPTU GVO BTQFDU of the game, these splits triple the opportunities to participate in some type of draft. Leagues may choose to do complete re-drafts and treat the year as three distinct mini-seasons. Or, leagues might allow teams to drop their five worst players and conduct a restocking draft at each break. Single game (Quint-Inning Lite format): Played with five owners drafting from the active rosters of two Major League teams in a single game. Prior to the game, 5-player rosters are snake-drafted (no positional requirements). Points are awarded based on how players perform during the game. Batters accumulate points for bases gained: 6LQJOH 'RXEOH 7ULSOH +5 %% +%3 6% Pitchers get +1 point for each full inning completed (3 outs) and lose 1 point for each run allowed (both earned and unearned). A deck of standard playing cards may be used as an aid for scorekeeping and to break ties. Players may be dropped, added or traded after the first inning. However, an owner must always have 5 players by the beginning of each half inning. Any player can be cut from an owner’s roster. 41 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Snake Drafting Free agents (players not rostered by any of the owners) can be claimed by any owner, between half-innings, in reverse order of the current standings. If two owners are tied and both want to place a claim, the tie is broken by drawing high card from the scorekeeping deck. Trades can be consummated at any time, between any two or more owners. At the beginning of the 5th inning, each owner has the option of doubling the points (positive and negative) for any one player on his roster for the remainder of the game (the “Quint”). Should that player be traded, or dropped and then re-acquired, his “Quint” status remains for the game. Beginning in the 9th inning, all batting points are doubled. Quint-Inning can be played as a low stakes, moderate or higher stakes competition. t *UDPTUT UPHFUJOUIFHBNF t *UDPTUTDFOUT QFSJOOJOHUPTUBZJOUIFHBNF for the first four innings. t #FHJOOJOHXJUIUIFUIJOOJOHUIFTUBLFTHPVQUP cents/$2/$10) per inning to stay in the game. t 4IPVMEUIFHBNFHPJOUPFYUSBJOOJOHTUIFTUBLFTSJTFUP ($1/$5/$25) to stay in the game until its conclusion. Each owner has to decide whether he is still in the game at the end of each full inning. Owners can drop out at the end of any inning, thus forfeiting any monies they’ve already contributed to the pot. When an owner drops, his players go back into the pool and can be acquired as free agents by the other owners. The winner is the owner who finishes the game with the most points. Post-season league: Some leagues re-draft teams from among the MLB post-season contenders and play out a separate competition. It is possible, however, to make a post-season competition that is an extension of the regular season. Start by designating a set number of regular season finishers as qualifying for the post-season. The top four teams in a league is a good number. These four teams would designate a fixed 23-man roster for all post-season games. First, they would freeze all of their currentlyowned players who are on MLB post-season teams. In order to fill the roster holes that will likely exist, these four teams would then pick players from their league’s non-playoff teams (for the sake of the post-season only). This would be in the form of a snake draft done on the day following the end of the regular season. Draft order would be regular season finish, so the play-off team with the most regular season points would get first pick. Picks would continue until all four rosters are filled with 23 men. Regular scoring would be used for all games during October. The team with the best play-off stats at the end of the World Series is the overall champ. Snake draft first round history The following tables record the comparison between pre-season projected player rankings (using Average Draft Position data from Mock Draft Central) and actual end-of-season results. The 8-year success rate of identifying each season’s top talent is only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aseball Forecaster / Encyclopedia of Fanalytics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being the safest investments. However, there are many variables affecting where players finish. Earlier, it was shown that players spend an inordinate number of days on the disabled list. In fact, of the players projected to finish in the top 300 coming into each of the past three seasons, the number who lost playing time due to injuries, demotions and suspensions has been extreme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hen you consider that about half of each season’s very best players had fewer at-bats or innings pitched than we projected, it shows how tough it is to rank players each year. The fallout? Consider: it is nearly a foregone conclusion that Jacoby Ellsbury and Curtis Granderson—players who finished in the Top 15 for the first time last year—will rank as first round picks in 2012. The above data provide a strong argument against them returning first-round value. Yes, they are excellent players, two of the best in the game, in 2011 anyway. But the issue is not their skills profile. The issue is the profile of what makes a worthy first rounder. Note: t 0WFSUIFQBTUTFWFOZFBSTUXPUIJSETPGQMBZFSTĕOJTIJOH in the Top 15 were not in the Top 15 the previous year. There is a great deal of turnover in the first round, year-to-year. t 0G UIPTF XIP XFSF ĕSTUUJNFST POMZ SFQFBUFE JO the first round the following year. t &TUBCMJTIFETVQFSTUBSTXIPĕOJTIFEJOUIF5PQXFSF no guarantee to repeat. These were players like David Ortiz (twice in 1st round, twice unable to repeat), Mark Teixeira (#5, 2005), Ryan Howard (#4, 2006) and Chase Utley (3 times in 1st round, only repeated once). Firsttime stars like Josh Hamilton (2008), Joe Mauer (2009) and Carlos Gonzalez (2010) are even less likely to repeat. t 'SPNUPPGUIFUPQQMBZFSTXFSFCBUUFST In 2008, that dropped to 13. In 2009 and 2010, it dropped again to 12. This past year, only 10 of 15 were batters. As player value shifts toward pitching, more arms are appearing in the first round, leaving fewer spots for the top batters. As such, the odds of Jacoby Ellsbury and Curtis Granderson repeating in the first round, regardless of skill, are steep. What is the best seed to draft from? Most drafters like mid-round so they never have to wait too long for their next player. Some like the swing pick, suggesting that getting two players at 15 and 16 is better than a 1 and a 30. Many drafters assume that the swing pick means you’d be getting something like two $30 players instead of a $40 and $20. Equivalent auction dollar values reveal the following facts about the first two snake draft rounds: In an AL-only league, the top seed would get a $44 player (at #1) and a $24 player (at #24) for a total of $68; the 12th seed would get two $29s (at #12 and #13) for $58. In an NL-only league, the top seed would get a $48 and a $28 ($76); the 13th seed would get two $30s ($60). ADP attrition Why is our success rate so low in identifying what should be the most easy-to-project players each year? We rank and draft players based on the expectation that those ranked higher will return greater value in terms of productivity and playing time, as well as 43 Encyclopedia of Fanalytics / 2012 Baseball Forecaster In a mixed league, the top seed would get a $47 and a $24 ($71); the 15th seed would get two $28s ($56). Since the talent level flattens out after the 2nd round, low seeds never get a chance to catch up: that drafters may be more likely to reach for a batter than for a pitcher. Using the ADP and corresponding standard deviation, we can to estimate the likelihood that a given player will be available at a certain draft pick. We estimate the predicted standard deviation for each player as follows: 'ROODUYDOXHGLIIHUHQFHEHWZHHQ ILUVWSOD\HUVHOHFWHGDQGODVWSOD\HUVHOHFWHG 5RXQG WHDP WHDP 6WGHY $'3(DUOLHVW3LFN (That the figure 0.42 appears twice is pure coincidence; the numbers are not equal past two decimal points.) If we assume that the picks are normally distributed, we can use a player’s ADP and estimated standard deviation to estimate the likelihood that the player is available with a certain pick (MS Excel formula): The total value each seed accumulates at the end of the draft is hardly equitable: 6HHG $/ 1/ QRUPGLVW[$'36WDQGDUG'HYLDWLRQ7UXH ZKHUHv[wUHSUHVHQWVWKHSLFNQXPEHUWREHHYDOXDWHG 0L[HG We can use this information to prepare for a snake draft by determining how early we may need to reach in order to roster a player. Suppose you have the 8th pick in a 15-team league draft and your target is 2009 sleeper candidate Nelson Cruz. His ADP is 128.9 and his earliest selection was with the 94th pick. This yields an estimated standard deviation of 14.2. You can then enter these values into the formula above to estimate the likelihood that he is still available at each of the following picks: /LNHOLKRRG 3LFN $YDLODEOH Of course, the draft is just the starting point for managing your roster and player values are variable. Still, it’s tough to imagine a scenario where the top seed wouldn’t have an advantage over the bottom seed. Using ADPs to determine when to select players ADPs and scarcity (Bill Macey) Most players are selected within a round or two of their ADP with tight clustering around the average. But every draft is unique and every pick in the draft seemingly affects the ordering of subsequent picks. In fact, deviations from “expected” sequences can sometimes start a chain reaction at that position. This is most often seen in runs at scarce positions such as the closer; once the first one goes, the next seems sure to closely follow. Research also suggests that within each position, there is a correlation within tiers of players. The sooner players within a generally accepted tier are selected, the sooner other players within the same tier will be taken. However, once that tier is exhausted, draft order reverts to normal. How can we use this information? If you notice a reach pick, you can expect that other drafters may follow suit. If your draft plan is to get a similar player within that tier, you’ll need to adjust your picks accordingly. (Bill Macey) Although average draft position (ADP) data gives us a good idea of where in the draft each player is selected, it can be misleading when trying to determine how early to target a player. This chart summarizes the percentage of players drafted within 15 picks of his ADP as well as the average standard deviation by grouping of players. $'3 5DQN ZLWKLQ SLFNV 6WDQGDUG 'HYLDWLRQ As the draft progresses, the picks for each player become more widely dispersed and less clustered around the average. Most top 100 players will go within one round of their ADP-converted round. However, as you reach the mid-to-late rounds, there is much more uncertainty as to when a player will be selected. Pitchers have slightly smaller standard deviations than do batters (i.e. they tend to be drafted in a narrower range). This suggests Mapping ADPs to auction value (Bill Macey) Reliable average auction values (AAV) are often tougher to come by than ADP data for snake drafts. However, we can estimate predicted auction prices as a function of ADP, arriving at the following equation: 44 2012 Baseball Forecaster / Encyclopedia of Fanalytics does in a given week, you earn. (In leagues that allow daily transactions, the unit is the player-day. The point stays.) When you draft a player, what have you bought? “You have bought the stats generated by this player.” No. You have bought the stats generated by his slot. Initially, the drafted player fills the slot, but he need not fill the slot for the season, and he need not contribute from Day One. If you trade the player during the season, then your bid on Draft Day paid for the stats of the original player plus the stats of the new player. If the player misses time due to injury or demotion, then you bought the stats of whomever fills the weeks while the drafted player is missing. At season’s end, there will be more players providing positive value than there are roster slots. Before the season, the number of players projected for positive value has to equal the total number of roster slots—after all, we can’t order owners to draft more players than can fit on their rosters. However, the projected productivity should be adjusted by the potential to capture extra value in the slot. This is especially important for injury-rehab cases and late-season call-ups. For example, if we think that a player will miss half the season, then we would augment his projected stats with a half-year of stats from a replacement-level player at his position. Only then would we calculate prices. Essentially, we want to apportion $260 per team among the slots, not the players. \ OQ[ ZKHUHOQ[LVWKHQDWXUDOORJIXQFWLRQ[UHSUHVHQWVWKHDFWXDO$'3DQG \UHSUHVHQWVWKHSUHGLFWHG$$9 This equation does an excellent job estimating auction prices (r2=0.93), though deviations are unavoidable. The asymptotic nature of the logarithmic function, however, causes the model to predict overly high prices for the top players. So be aware of that, and adjust. Auction Value Analysis Auction values (R$) in perspective R$ is the dollar value placed on a player’s statistical performance in a Rotisserie league, and designed to measure the impact that player has on the standings. There are several methods to calculate a player’s value from his projected (or actual) statistics. One method is Standings Gain Points, described in the book, How to Value Players for Rotisserie Baseball, by Art McGee (2nd edition available at BaseballHQ.com). SGP converts a player’s statistics in each Rotisserie category into the number of points those stats will allow you to gain in the standings. These are then converted back into dollars. Another popular method is the Percentage Valuation Method. In PVM, a least valuable, or replacement performance level is set for each category (in a given league size) and then values are calculated representing the incremental improvement from that base. A player is then awarded value in direct proportion to the level he contributes to each category. As much as these methods serve to attach a firm number to projected performance, the winning bid for any player is still highly variable depending upon many factors: t UIFTBMBSZDBQMJNJU t UIFOVNCFSPGUFBNTJOUIFMFBHVF t FBDIUFBNTSPTUFSTJ[F t UIFJNQBDUPGBOZQSPUFDUFEQMBZFST t FBDIUFBNTQPTJUJPOBMEFNBOETBUUIFUJNFPGCJEEJOH t UIFTUBUJTUJDBMDBUFHPSZEFNBOETBUUIFUJNFPGCJEEJOH t FYUFSOBM GBDUPST FH NFEJB JOĘBUJPO PS EFĘBUJPO PG value In other words, a $30 player is only a $30 player if someone in your draft pays $30 for him. Average player value by draft round Roster slot valuation (John Burnson) Tenets of player valuation say that the number of ballplayers with positive value—either positive projected value (before the season) or positive actual value (after the season)—must equal the total number of roster spots, and that, before the season, the value of a player must match his expected production. These propositions are wrong. The unit of production in Rotisserie is not “the player” or “the statistic” but the player-week. If you own a player, you must own him for at least one week, and if you own him for more than one week, you must own him for multiples of one week. Moreover, you cannot break down his production—everything that a player 5G $/ 1/ 0[G $YJ Benchmarks for auction players: t "MMQMBZFSTXJMMHPJOUIFĕSTUSPVOE t "MMQMVTQMBZFSTXJMMHPJOUIFĕSTUGPVSSPVOET t %PVCMFEJHJUWBMVFFOETQSFUUZNVDIBęFS3PVOE t ćFFOEHBNFTUBSUTBUBCPVU3PVOE 45 Encyclopedia of Fanalytics / 2012 Baseball Forecaster Dollar values by lineup position (Michael Roy) Still, a player with two consecutive seasons at $30 in value is barely a 50/50 proposition. And three consecutive seasons is only a 2/3 shot. Small sample sizes aside, this does illustrate the nature of the beast. Even the most consistent, reliable players fail 1/3 of the time. Of course, this is true whether they are kept or drafted anew, so this alone shouldn’t prevent you from keeping a player. How much value is derived from batting order position? 3RV 3$ 5 5%, 5 How well do elite pitchers retain their value? (Michael Weddell) An elite pitcher (one who earns at least $24 in a season) on average keeps 80% of his R$ value from year 1 to year 2. This compares to the baseline case of only 52%. Historically, 36% of elite pitchers improve, returning a greater R$ in the second year than they did the first year. That is an impressive performance considering they already were at an elite level. 17% collapse, returning less than a third of their R$ in the second year. The remaining 47% experience a middling outcome, keeping more than a third but less than all of their R$ from one year to the next. So, a batter moving from the bottom of the order to the cleanup spot, with no change in performance, would gain nearly $4 in value from runs and RBIs alone. Dollar values: expected projective accuracy There is a 65% chance that a player projected for a certain dollar value will finish the season with a final value within plus-orminus $5 of that projection. That means, if you value a player at $25, you only have about a 2-in-3 shot of him finishing between $20 and $30. If you want to get your odds up to 80%, the range now becomes +/- $9. You have an 80% shot that your $25 player will finish somewhere between $16 and $34. Valuing closers Given the high risk associated with the closer’s role, it is difficult to determine a fair draft value. Typically, those who have successfully held the role for several seasons will earn the highest draft price, but valuing less stable commodities is troublesome. A very rough, but workable, rule of thumb is to start by paying $10 for the role alone. Any pitcher slated to be the closer on draft day should merit at least that bid amount. Then add anywhere from $0 to $15 for support skills. In this way, the top level talents will draw upwards of $20-$25. Those with moderate skill will draw $15-$20, and those with more questionable skill in the $10-$15 range. How likely is it that a $30 player will repeat? (Matt Cederholm) From 2003-2008, there were 205 players who earned $30 or more (using single-league 5x5 values). Only 70 of them (34%) earned $30 or more in the next season. In fact, the odds of repeating a $30 season aren’t good no matter how you slice it. As seen below, the best odds during that period were 42%. And as we would expect, pitchers fare far worse than hitters. +LWWHUV 3LWFKHUV 7RWDO 7RWDO! 5HSHDW 5HSHDW +LJK5HOLDELOLW\ +LWWHUV 3LWFKHUV 7RWDO %39 +LWWHUV 3LWFKHUV 7RWDO +LJK5HOLDELOLW\DQG%39 +LWWHUV 3LWFKHUV 7RWDO 5HOLDELOLW\ILJXUHVDUHIURP Profiling the end game What types of players are typically the most profitable in the endgame? First, our overall track record on $1 picks: $YJ5HWXUQ 7RWDO! 5HSHDW $YJ3URI $YJ/RVV On aggregate, the hundreds of players drafted in the end-game earned $1.89 on our $1 investments. While they were profitable overall, only 51% of them actually turned a profit. Those that did cleared more than $10 on average. Those that didn’t—the other 49%—lost about $7 apiece. 3RV 3FWRIWRW $YJ9DO 3URILW &$ &2 0, 2) 63 53 $YJ3URI $YJ/RVV These results bear out the danger of leaving catchers to the end game. They were the only position that returned negative value. Corner infielder returns beg the tactic of leaving open a 1B or 3B spot for the end. If we look at players with multiple seasons of $30 or more in value, the numbers do get better. For players with consecutive $30 seasons, 2003-2008: 7ZR<HDUV 7KUHH<HDUV 3URILWDEOH $JH 3FWRIWRW $YJ9DO 3URILW 5HSHDW 46 $YJ3URI $YJ/RVV 2012 Baseball Forecaster / Encyclopedia of Fanalytics Mound Aces, and also pays tribute to Jose Lima, a $1 pitcher in 1998 who exemplified the power of the strategy. In a $260 league: 1. Budget a maximum of $60 for your pitching staff. 2. Allot no more than $30 of that budget for acquiring saves. In 5x5 leagues, it is reasonable to forego saves at the draft (and acquire them during the season) and re-allocate this $30 to starters ($20) and offense ($10). 3. Ignore ERA. Draft only pitchers with: t$PNNBOESBUJP,## PGPSCFUUFS t4USJLFPVUSBUFPGPSCFUUFS t&YQFDUFEIPNFSVOSBUFPGPSMFTT 4. Draft as few innings as your league rules will allow. This is intended to manage risk. For some game formats, this should be a secondary consideration. 5. Maximize your batting slots. Target batters with: t$POUBDUSBUFPGBUMFBTU t8BMLSBUFPGBUMFBTU t19PS4QEMFWFMPGBUMFBTU Spend no more than $29 for any player and try to keep the $1 picks to a minimum. The goal is to ace the batting categories and carefully pick your pitching staff so that it will finish in the upper third in ERA, WHIP and saves (and Ks in 5x5), and an upside of perhaps 9th in wins. In a competitive league, that should be enough to win, and definitely enough to finish in the money. Worst case, you should have an excess of offense available that you can deal for pitching. The strategy works because it better allocates resources. Fantasy leaguers who spend a lot for pitching are not only paying for expected performance, they are also paying for better defined roles—#1 and #2 rotation starters, ace closers, etc.—which are expected to translate into more IP, wins and saves. But roles are highly variable. A pitcher’s role will usually come down to his skill and performance; if he doesn’t perform, he’ll lose the role. The LIMA Plan says, let’s invest in skill and let the roles fall where they may. In the long run, better skills should translate into more innings, wins and saves. And as it turns out, pitching skill costs less than pitching roles do. In snake draft leagues, don’t start drafting starting pitchers until Round 10. In shallow mixed leagues, the LIMA Plan may not be necessary; just focus on the BPI benchmarks. In simulation leagues, build your staff around BPI benchmarks. The practice of speculating on younger players—mostly rookies—in the end game was a washout. Part of the reason was that those that even made it to the end game were often the longterm or fringe type. The better prospects were typically drafted earlier. ,QMXU\UHKDEV 3FWRIWRW $YJ9DO 3URILW $YJ3URI $YJ/RVV One in five end-gamers were players coming back from injury. While only 36% of them were profitable, the healthy ones returned a healthy profit. The group’s losses were small, likely because they weren’t healthy enough to play. Advanced Draft Strategies Stars & Scrubs v. Spread the Risk Stars & Scrubs (S&S): A Rotisserie auction strategy in which a roster is anchored by a core of high priced stars and the remaining positions filled with low-cost players. Spread the Risk (STR): An auction strategy in which available dollars are spread evenly among all roster slots. Both approaches have benefits and risks. An experiment was conducted in 2004 whereby a league was stocked with four teams assembled as S&S, four as STR and four as a control group. Rosters were then frozen for the season. The Stars & Scrubs teams won all three ratio categories. Those deep investments ensured stability in the categories that are typically most difficult to manage. On the batting side, however, S&S teams amassed the least amount of playing time, which in turn led to bottom-rung finishes in HRs, RBIs and Runs. One of the arguments for the S&S approach is that it is easier to replace end-game losers (which, in turn, may help resolve the playing time issues). Not only is this true, but the results of this experiment show that replacing those bottom players is critical to success. The Spread the Risk teams stockpiled playing time, which in turn led to strong finishes in many of the counting stats, including clear victories in RBIs, wins and strikeouts. This is a key tenet in drafting philosophy; we often say that the team that compiles the most ABs will undoubtedly be among the top teams in RBI and Runs. The danger is on the pitching side. More innings did yield more wins and Ks, but also destroyed ERA/WHIP. So, what approach makes the most sense? The optimal strategy might be to STR on offense and go S&S with your pitching staff. STR buys more ABs, so you immediately position yourself well in four of the five batting categories. On pitching, it might be more advisable to roster a few core arms, though that immediately elevates your risk exposure. Admittedly, it’s a balancing act, which is why we need to pay more attention to risk analysis and look closer at strategies like the RIMA Plan and Portfolio3. Variations on the LIMA Plan LIMA Extrema: Limit your total pitching budget to only $30, or less. This can be particularly effective in shallow leagues where LIMA-caliber starting pitcher free agents are plentiful during the season. SANTANA Plan: Instead of spending $30 on saves, you spend it on a starting pitcher anchor. In 5x5 leagues where you can reasonably punt saves at the draft table, allocating those dollars to a high-end LIMA-caliber starting pitcher can work well as long as you pick the right anchor. One way to approach that selection is... RIMA Plan: LIMA is based on optimal resource allocation. These days, however, no matter how good of a team you draft, The LIMA Plan The LIMA Plan is a strategy for Rotisserie leagues (though the underlying concept can be used in other formats) that allows you to target high skills pitchers at very low cost, thereby freeing up dollars for offense. LIMA is an acronym for Low Investment 47 Encyclopedia of Fanalytics / 2012 Baseball Forecaster do happen, but in general, other owners will not provide that much resistance. player inconsistency, injuries and unexpected risk factors can wreak havoc with your season. The RIMA Plan adds the element of RIsk MAnagement. Players are not risks by virtue of their price tags alone. A $35 Roy Halladay, for example, might be a very good buy since he is a healthy, stable commodity. But most LIMA drafters would not consider him because of the price. The RIMA Plan involves setting up two pools of players. The first pool consists of those who meet the LIMA criteria. The second pool includes players with high Reliability grades. The set of players who appear in both pools are our prime draft targets. We then evaluate the two pools further, integrating different levels of skill and risk. RIMA was introduced in 2004; the specifics of the plan have since been assimilated into Portfolio3. How it’s done 1. Create your optimal draft pool. 2. Get those players. Start by identifying which players will be draftable based on the LIMA or Portfolio3 criteria. Then, at the draft, your focus has to be on your roster only. When it’s your bid opener, toss a player you need at about 50%-75% of your projected value. Bid aggressively. Forget about bargain-hunting; just pay what you need to pay. Of course, don’t spend $40 for a $25 player, but it’s okay to exceed your projected value within reason. Mix up the caliber of openers. Instead of tossing out an Albert Pujols at $35 in the first round, toss out a Ben Revere at $8. Wise Guy Baseball’s Gene McCaffrey suggests tossing all lower-end players early, which makes sense. It helps you bottom-fill your roster with players most others won’t chase early, and you can always build the top end of your roster with players others toss out. Another good early tactic is to gauge the market value of scarce commodities with a $19 opener for Heath Bell (saves) or a $29 opener for Michael Bourn (stolen bases). At the end of the draft, you may have rostered 23 players who could have been purchased at somewhat lower cost. It’s tough to say. Those extra dollars likely won’t mean much anyway; in fact, you might have just left them on the table. TCD almost ensures that you spend all your money. In the end, it’s okay to pay a slight premium to make sure you get the players with the highest potential to provide a good return on your investment. It’s no different than the premium you’d pay to get the last valuable shortstop, or for the position flexibility a player like Emilio Bonafacio provides. With TCD, you’re just spending those extra dollars up front on players with high skill and low risk. The best part is that you take more control of your destiny. You build your roster with what you consider are the best assemblage of players. You keep the focus on your team. And you don’t just roster whatever bargains the rest of the table leaves for you, because a bargain is just a fleeting perception of value we have in March. Total Control Drafting (TCD) Part of the reason we play this game is the aura of “control,” our ability to create a team of players we want and manage them to a title. We make every effort to control as many elements as possible, but in reality, the players that end up on our teams are largely controlled by the other owners. Their bidding affects your ability to roster the players you want. In a snake draft, the other owners control your roster even more. We are really only able to get the players we want within the limitations set by others. However, an optimal roster can be constructed from a fanalytic assessment of skill and risk. We can create our teams from that “perfect player pool” and not be forced to roster players that don’t fit our criteria. It’s now possible. It’s just a matter of taking Total Control. Why this makes sense 1. Our obsession with projected player values is holding us back. If a player on your draft list is valued at $20 and you agonize when the bidding hits $23, odds are about two chances in three that he could really earn anywhere from $15 to $25. What this means is, in some cases, and within reason, you should just pay what it takes to get the players you want. 2. There is no such thing as a bargain. Most of us don’t just pay what it takes because we are always on the lookout for players who go under value. But we really don’t know which players will cost less than they will earn because prices are still driven by the draft table. The concept of “bargain” assumes that we even know what a player’s true value is. To wit: If we target a player at $23 and land him for $20, we might think we got a bargain. In reality, this player might earn anywhere from $19 to $26, making that $3 in perceived savings virtually irrelevant. The point is, a “bargain” is defined by your particular marketplace at the time of your particular draft, not by any list of canned values, or an “expectation” of what the market value of any player might be. So any contention that TCD forces you to overpay for your players is false. 3. “Control” is there for the taking. Most owners are so focused on their own team that they really don’t pay much attention to what you’re doing. There are some exceptions, and bidding wars The Portfolio3 Plan The previously discussed strategies have had important roles in furthering our potential for success. The problem is that they all take a broad-stroke approach to the draft. The $35 first round player is evaluated and integrated into the plan in the same way that the end-gamer is. But each player has a different role on your team by virtue of his skill set, dollar value, position and risk profile. When it comes to a strategy for how to approach a specific player, one size does not fit all. We need some players to return fair value more than others. When you spend $40 on a player, you are buying the promise of putting more than 15% of your budget in the hands of 4% of your roster. By contrast, the $1 players are easily replaceable. If you’re 48 2012 Baseball Forecaster / Encyclopedia of Fanalytics in a snake draft league, you know that a first-rounder going bellyup is going to hurt you far more than a 23rd round bust. We rely on some players for profit more than others. Those firstrounders are not where we are likely going to see the most profit potential. The $10-$20 players are likely to return more pure dollar profit; the end-gamers are most likely to return the highest profit percentage. We can afford to weather more risk with some players than with others. Since those high-priced early-rounders need to return at least fair value, we cannot afford to take on excessive risk. Since we need more profit potential from the lower priced, later-round picks, that means opening up our tolerance for risk more with those players. Players have different risk profiles based solely on what roster spot they are going to fill. Catchers are more injury prone. A closer’s value is highly dependent on fickle managerial whim. These types of players are high risk even if they have the best skills on the planet. That needs to affect their draft price or draft round. For some players, the promise of providing a scarce skill, or productivity at a scarce position, may trump risk. Not always, but sometimes. At minimum, we need to be open to the possibility. The determining factor is usually price. A $5, 20th round Daniel Murphy is not something you pass up, even with a Reliability Grade of FDD. In the end, we need a way to integrate all these different types of players, roles and needs. We need to put some form to the concept of a diversified draft approach. Thus: The Portfolio3 Plan provides a three-tiered approach to the draft. Just like most folks prefer to diversify their stock portfolio, P3 advises to diversify your roster with three different types of players. Depending upon the stage of the draft (and budget constraints in auction leagues), P3 uses a different set of rules for each tier that you’ll draft from. The three tiers are: 1. Core Players 2. Mid-Game Players 3. End-Game Players 7,(5&25(3/$<(56 5RVWHU 6ORWV %XGJHW 0D[ %$77(56 5HO &W 3;RU6SG %%% pitcher in your core, it will be one with high-level skill. For both batters and pitchers, minimum reliability grades of BBB cover risk. Since these are going to be the most important players on your roster, the above guidelines help provide a report card, of sorts, for your draft. For instance, if you leave the table with only three Tier 1 players, then you know you have likely rostered too much risk or not enough skill. If you manage to draft nine Tier 1 players, that doesn’t necessarily mean you’ve got a better roster, just a better core. There still may be more work to do in the other tiers. Tier 1 remains the most important group of players as they are the blue chips that allow you to take chances elsewhere. However, there can be some play within this group on the batting side. The 80% contact rate is important to help protect the batting average category. However, with some care, you can roster a few BA studs to allow you the flexibility to take on some low-contact hitters who excel in other areas (typically power). The tactic would work like this... If you are short on Tier 1 players and have exhausted the pool of those who meet all the filters, you can work your way down the following list... 7,(5%$77(56 3ULPDU\JURXS 6HFRQGDU\ 7HUWLDU\ 5HO %%% %%% %%% &W 3;RU 6SG ...knowing full well that, for every player you roster from these lower groups, you are putting your batting average at greater risk. You should only do this if you think the power/speed gains will sufficiently offset any BA shortfalls. These two sub-groups are not fixed filters; they form a continuum. So if you have a player with a 78% contact rate, your PX/Spd requirement would probably be somewhere around 105. I would not go anywhere near a player with a contact rate less than 70%. 7,(50,'*$0(3/$<(56 5RVWHU %$77(56 6ORWV%XGJHW 5HO &WRU 3;RU6SG %%% $OOSOD\HUVPXVWEHOHVVWKDQ %DWWHUVPXVWEHSURMHFWHGIRUDWOHDVW$% 3,7&+(56 5HO %39 %%% 3,7&+(56 5HO %39 %%% In an early 2008 column, I noted how fellow Tout combatant, Jason Grey, was consistently able to assemble offensive juggernauts by the singular tactic of accumulating massive amounts of often-cheap playing time. Intrinsic skill was irrelevant. On the offensive side, this makes sense. Runs and RBI are only tangentially related to skill. If a player is getting 500 AB, he is likely going to provide positive value in those categories just from opportunity alone. And given that his team is seeing fit to give him those AB, he is probably also contributing somewhere else. These players have value to us. And we can further filter this pool of full-timers who miss the P3 skills criteria by skimming off those with high REL grades. There are two dangers in this line of thinking. First, it potentially puts the batting average category at risk. You don’t want to accumulate bad AB; when you dig yourself into a hole in BA, it is These are the players who will provide the foundation to your roster. These are your prime stat contributors and where you will invest the largest percentage of your budget. In snake drafts, these are the names you pick in the early rounds. There is no room for risk here. Given their price tags, there is usually little potential for profit. The majority of your core players should be batters. The above chart shows general roster goals. In a snake draft, you need to select core-caliber players in the first 5-8 rounds. In an auction, any player purchased for $20 or more should meet the Tier 1 filters. The filters are not strict, but they are important, so you should stick to them as best as possible. An 80% contact rate ensures that your batting average category is covered. PX and Spd ensure that you draft players with a minimum league average power or speed. On the pitching side, a BPV of 75 ensures that, if you must draft a 49 Encyclopedia of Fanalytics / 2012 Baseball Forecaster the one category that is nearly impossible to dig out of. However, this just means we need to approach it tactically; if we decide to roster a Mark Reynolds, we must also roster a Michael Young. Second, this line of thinking assumes we can accurately project playing time. But if we focus on those players who are locked into firm roles, there is still a decent-sized pool to draw from. Tier 2 is often where the biggest auction bargains tend to be found as the blue-chippers are already gone and owners are reassessing their finances. It is in that mid-draft lull where you can scoop up tons of profit. In a snake draft, these players should take you down to about round 16-18. 7,(5(1'*$0(3/$<(56 5RVWHU %$77(56 6ORWV%XGJHW 5HO &W 3;RU6SG 8SWR QD $OOSOD\HUVPXVWEHOHVVWKDQ to hold off on starting pitchers until as late as Round 10 or 11; however, if it’s Round 7 and Cole Hamels is still sitting out there, by all means jump. LIMA Plan: Although LIMA says no starting pitchers over $20, Tier 1 provides a few options where it would be okay to break the rules. You can adjust your $60 pitching budget up to accommodate, or downgrade saves targets. Punting saves: Still viable, unless a Tier 1 closer falls into your lap. These are extremely rare commodities anyway. Keeper leagues: When you decide upon your freeze list, you should be looking for two types of keepers—the best values and the most valuable players. Freezing a $15 Tim Lincecum is a no-brainer; where some drafters struggle is with the $25 Corey Hart. Given that TCD says that we should be willing to pay a premium for Tier 1 players, any name on that list should be a freeze consideration. Adding in the variable of potential draft inflation, you should be more flexible with the prices you’d be willing to freeze players at. For instance, if you currently own a $40 Miguel Cabrera, you might be tempted to throw him back. However, between draft inflation and the premium you should be willing to pay for a Tier 1 commodity, his real value could be well over $50. 3,7&+(56 5HO %39 QD For some fantasy leaguers, the end game is when the beer is gone and you have to complete your roster with any warm body. In P3, these are gambling chips, but every end-gamer must provide the promise of upside. For that reason, the focus must remain on skill and conditional opportunity. P3 drafters should fill the majority of their pitching slots from this group. By definition, end-gamers are typically high risk players, but risk is something you’ll want to embrace here. You probably don’t want a Kosuke Fukudome-type player at the end of the draft. His ABB reliability grade would provide stability, but there is no upside, so there is little profit potential. This is where you need to look for profit so it is better here to ignore reliability; instead, take a few chances in your quest for those pockets of possible profit. If the player does not pan out, he can be easily replaced. As such, a Tier 3 end-gamer should possess the BPI skill levels noted above, and... t QMBZJOHUJNFVQTJEFBTBCBDLVQUPBSJTLZGSPOUMJOFS t BOJOKVSZIJTUPSZUIBUIBTEFQSFTTFEIJTWBMVF t TPMJETLJMMTEFNPOTUSBUFEBUTPNFQPJOUJOUIFQBTU t NJOPSMFBHVFQPUFOUJBMFWFOJGIFIBTCFFONPSFSFDFOUMZ a major league bust Notes on draft implementation... Auction leagues: Tier 1 player acquisition should be via the Total Control Drafting method. Simply, pay whatever it takes, within reason. Be willing to pay a small premium for the low risk and high skills combination. Snake drafters will have choices in the first six rounds or so. There are no guarantees—a swing-pick seed might negate any chance you have for rostering some players—but at least there are some options. If you miss out on the cream, you can either drop down and select a lower round player early, or relax the filters a bit to grab someone who might have higher value but perhaps greater risk. Position scarcity: While we still promote the use of position scarcity in snake drafts, it may be more important to have solid foundation players in the early rounds. Drafting pitchers early is still something we advise against. However, if you are going to grab a pitcher in the first six rounds, at least make sure it’s a Tier 1 name. It is still a viable strategy The Mayberry Method The Mayberry Method—named after the fictional TV village where life was simpler—is a player evaluation method that embraces the imprecision of the forecasting process and projects performance in broad strokes rather than with precise statistics. MM reduces every player to a 7-character code. The format of the code is 5555AAA, where the first four characters describe elements of a player’s skill on a scale of 0 to 5. The three alpha characters are our reliability grades (health, experience and consistency). The skills numerics are forward-looking; the alpha characters grade reliability based on past history. Batting The first character in the MM code measures a batter’s power skills. It is assigned using the following table: 3; 00 5RXJK+5$SSUR[ XSWR XSWR XSWR XSWR XSWR The second character measures a batter’s speed skills. 6SG 50 00 5RXJK6%$SSUR[ XSWR XSWR XSWR XSWR XSWR 2012 Baseball Forecaster / Encyclopedia of Fanalytics Integrating Mayberry and Portfolio3 Mayberry scores can be used as a proxy for the BPI filters in the Portfolio3 chart. Pretty much all you need to remember is the number “3.” The third character measures expected batting average. [%$ 00 7U The fourth character measures playing time. 5ROH 3RWHQWLDOIXOOWLPHUV 0LGWLPHUV )ULQJHEHQFK 1RQIDFWRUV 3$ 006FRUH 3;VFRUH6SGVFRUH[%$VFRUH3$VFRUH [3$VFRUH Pitching The first character in the pitching MM code measures xERA, which captures a pitcher’s overall ability and is a proxy for ERA, and even WHIP. %$77,1* WHDPPL[HG WHDPPL[HG WHDP$/1/ In-Season Analyses The second character measures strikeout ability. The efficacy of streaming 00 6DYHVHVW 00 The fourth character measures playing time. ,3 (John Burnson) In leagues that allow weekly or daily transactions, many owners flit from hot player to hot player. But published dollar values don’t capture this traffic—they assume that players are owned from April to October. For many leagues, this may be unrealistic. We decided to calculate these “investor returns.” For each week, we identified the top players by one statistic—BA for hitters, ERA for pitchers—and took the top 100 hitters and top 50 pitchers. We then said that, at the end of the week, the #1 player was picked up (or already owned) by 100% of teams, the #2 player was picked up or owned by 99% of teams, and so on, down to the 100th player, who was on 1% of teams. (For pitchers, we stepped by 2%.) Last, we tracked each player’s performance in the next week, when ownership matters. We ran this process anew for every week of the season, tabulating each player’s “investor returns” along the way. If a player was owned by 100% of teams, then we awarded him 100% of his performance. If the player was owned by half the teams, we gave him half his performance. If he was owned by no one (that is, he was not among the top players in the prior week), his performance was ignored. A player’s cumulative stats over the season was his investor return. The results... t PG QJUDIFST IBE QPPSFS JOWFTUPS SFUVSOT XJUI BO aggregate ERA 0.40 higher than their true ERA. The third character measures saves potential. 5ROH 3RWHQWLDOVWDUWHUV 3RWHQWLDOVWDUWHUV VWDUWHUVVZLQJPHQ 5HOLHYHUV 3; 6SG [%$ 3$ 00 3,7&+,1* (5$ . 6Y ,3 00 WHDPPL[HG WHDPPL[HG WHDP$/1/ 0DNHVXUHWKHPDMRULW\RIWKHVHSRLQWVFRPHIURPVWDUWLQJSLWFKHUV 00 'HVFULSWLRQ 1RKRSHIRUVDYHVVWDUWLQJSLWFKHUV 6SHFXODWLYHFORVHU &ORVHULQDSHQZLWKDOWHUQDWLYHV )URQWOLQHFORVHUZLWKILUPEXOOSHQUROH 3,7&+(56 5HO [(5$ %%% %%% QD Targets For roster budgeting purposes, here are targets for several standard leagues: The highest score you can get is 100, so this becomes an easy scale to evaluate. . %$77(56 5HO 3$[%$ 3;RU6SG %%% QD %%% RU QD QD The conversion to Mayberry is pretty straightforward. On the batting side, xBA provides a good proxy for contact rate (since ct% drives xBA), and PX and Spd stay. For pitching, xERA easily replaces BPV. Mayberry’s broad-stroke approach significantly opens up the draftable player pool, particularly for Tier 3 sleeper types. About 25% more players will be draft-worthy using these filters. 00 An overall MM batting score is calculated as: [(5$ 5RVWHU 6ORWV %XGJHW 0D[ 8SWR 00 An overall MM pitching score is calculated as: 006FRUH [(5$VFRUH[.VFRUH6DYHVVFRUH,3VFRUH [,3VFRUH6DYHVVFRUH 51 Encyclopedia of Fanalytics / 2012 Baseball Forecaster very powerful, as a single week might impact the standings by over 1.5 points in ERA/WHIP, positively or negatively. t PG CBUUFST IBE QPPSFS JOWFTUPS SFUVSOT CVU XJUI BO aggregate batting average virtually identical to the true BA. Sitting stars and starting scrubs Consistency (Ed DeCaria) In setting your pitching rotation, conventional wisdom suggests sticking with trusted stars despite difficult matchups. But does this hold up? And can you carefully start inferior pitchers against weaker opponents? Here are the ERA’s posted by varying skilled pitchers facing a range of different strength offenses: 23326,1*2))(16(5&* 3LWFKHU(5$ Recommendations: 1. Never start below replacement-level pitchers. 2. Always start elite pitchers. 3. Other than that, never say never or always. Playing matchups can pay off when the difference in opposing offense is severe. Two-start pitcher weeks (Ed DeCaria) Gaming Demographics, etc. A two-start pitcher is a prized possession. But those starts can mean two DOMinant outings, two DISasters, or anything else in between, as shown by these results: 3463DLU :HHNV '20'20 '20$9* $9*$9* '20',6 $9*',6 ',6',6 (5$ (Dylan Hedges) Few things are as valuable to head-to-head league success as filling your roster with players who can produce a solid baseline of stats, week in and week out. In traditional leagues, while consistency is not as important—all we care about are aggregate numbers— filling your team with consistent players can make roster management easier. Consistent batters have good plate discipline, walk rates and on base percentages. These are foundation skills. Those who add power to the mix are obviously more valuable, however, the ability to hit home runs consistently is rare. Consistent pitchers demonstrate similar skills in each outing; if they also produce similar results, they are even more valuable. We can track consistency but predicting it is difficult. Many fantasy leaguers try to predict a batter’s hot or cold streaks, or individual pitcher starts, but that is typically a fool’s errand. The best we can do is find players who demonstrate seasonal consistency over time; in-season, we want to manage players and consistency tactically. Each year at BaseballHQ.com, we keep tabs on ongoing fantasy baseball industry trends by means of our weekly HQ Poll, which we’ve been running since December 1998. We’ve been tracking dozens of trends. Since these questions are asked at Baseball HQ, they only represent the opinions of folks who a) visit Baseball HQ, and b) respond to online polls. So these are clearly not scientific representations of the industry as a whole. However, even this group can reveal some interesting tidbits about how fantasy leaguers play the game. Note that all poll results had at least 500 responses. Here is a profile of these respondents—who very well might be an accurate reflection of you—by the numbers: :+,3 :LQ:N .:N Weeks that include even one DISaster start produce terrible results. Unfortunately, avoiding such disasters is much easier in hindsight. But what is the actual impact of this decision on the stat categories? ERA and WHIP: When the difference between opponents is extreme, inferior pitchers can actually be a better percentage play. This is true both for one-start pitchers and two-start pitchers, and for choosing inferior one-start pitchers over superior two-start pitchers. Strikeouts per Week: Unlike the two rate stats, there is a massive shift in the balance of power between one-start and two-start pitchers in the strikeout category. Even stars with easy one-start matchups can only barely keep pace with two-start replacementlevel arms in strikeouts per week. Wins per week are also dominated by the two-start pitchers. Even the very worst two-start pitchers will earn a half of a win on average, which is the same rate as the very best one-start pitchers. The bottom line: If strikeouts and wins are the strategic priority, use as many two-start weeks as the rules allow, even if it means using a replacement-level pitcher with two tough starts instead of a mid-level arm with a single easy start. But if ERA and/ or WHIP management are the priority, two-start pitchers can be GENERAL/OFFSEASON 31% play fantasy for the competition and the camaraderie. Cash winnings are irrelevant. 17% say there are certain owners they dislike and work harder to beat. 73% say their primary league is a keeper league. 17% still play 4x4 Rotisserie (down from 23% in 2003) 37% start prepping for fantasy baseball following the end of the fantasy football season. 20% are in leagues that use Holds as a category; 3% don’t like it. 60% are in mixed leagues with 12 or fewer teams. 48% are in weekly leagues and 41% of those have a Monday noon transaction deadline. 21% say it is impossible to contend without a constant pulse on all levels of the minors. 16% of those in keeper leagues fill up all allowable slots, even if some players are marginal keepers. 52 2012 Baseball Forecaster / Encyclopedia of Fanalytics End-game wasteland Home for players undraftable in the deepest of leagues, who stay in the free agent pool all year. It’s the place where even crickets keep quiet when a name is called at the draft. DRAFT DAY 66% prefer a live auction over any other roster-stocking method. 30% like to draft a balanced team all around. 69% rarely or never try to roster at least one player from your favorite major league team. 35% don’t bother calculating draft inflation. FAAB Forewarnings 1. Spend early and often. 2. Emptying your budget for one prime league-crosser is a tactic that should be reserved for the desperate. 3. If you chase two rabbits, you will lose them both. IN-SEASON 47% try to devote an equal amount of time to all their teams. 50% say their preferred method for free agent acquisition is a Free Agent Acquisition Budget. 35% cite May 1 as the date they start taking the standings seriously to start working their roster. 48% say that, when it comes to dump trades, “all’s fair in fantasy baseball.” 58% do not have an in-season salary cap. 65% of keeper leaguers have an August trading deadline. 49% say it is possible to overcome a 20-point deficit in July to win a league. 26% said they would never trade Matt Kemp for Roy Halladay even up. 34% will quit on a struggling team before the season is over. 26% have at least some interest in post-season fantasy competitions. Fantasy Economics 101 The market value for a player is generally based on the aura of past performance, not the promise of future potential. Your greatest advantage is to leverage the variance between market value and real value. Fantasy Economics 102 The variance between market value and real value is far more important than the absolute accuracy of any individual player projection. Hope A commodity that routinely goes for $5 over value at the draft table. JA$G Just Another Dollar Guy. Professional Free Agent (PFA) Player whose name will never come up on draft day but will always end up on a roster at some point during the season as an injury replacement. Formula for consistent success Anyone can win a league in any given season. Winning once proves very little, especially in redraft leagues. True success has to be defined as the ability to win consistently. It is a feat in itself to reach the mountaintop, but the battle isn’t truly won unless you can stay atop that peak while others keep trying to knock you off. What does it take to win that battle? We surveyed 12 of the most prolific fantasy champions in national experts league play. Here is how they rated six variables: %HWWHULQGUDIWVWUDWHJ\WDFWLFV %HWWHUVHQVHRISOD\HUYDOXH %HWWHUOXFN %HWWHUJUDVSRIFRQWH[WXDO HOHPHQWVWKDWDIIHFWSOD\HUV %HWWHULQVHDVRQURVWHUPDQDJHPHQW %HWWHUSOD\HUSURMHFWLRQV RUM pick A player who is rosterable only as a Reserve, Ultra or Minors pick. Standings Vantage Points First Place: It’s lonely at the top, but it’s comforting to look down upon everyone else. Sixth Place: The toughest position to be in is mid-pack at dump time. Last Place in April: The sooner you fall behind, the more time you will have to catch up. Last Place, Yet Again: If you can’t learn to do something well, learn to enjoy doing it badly. 3HUFHQWUDQNHG 6FRUH Mike Timlin List Players who you are unable to resist drafting even though they have burned you multiple times in the past. Seasonal Assessment Standard If you still have reason to be reading the boxscores during the last weekend of the season, then your year has to be considered a success. Other Diamonds Cellar value The dollar value at which a player cannot help but earn more than he costs. Always profit here. The Three Cardinal Rules for Winners If you cherish this hobby, you will live by them or die by them... 1. Revel in your success; fame is fleeting. 2. Exercise excruciating humility. 3. 100% of winnings must be spent on significant others. Crickets The sound heard when someone’s opening draft bid on a player is also the only bid. Scott Elarton List Players you drop out on when the bidding reaches $1. 53 S TAT I S T I C A L R E S E A R C H A B S T R AC T S MLB’s Changing Statistical Landscape It may be that five straight years of BABIP decline can be attributed to the recent emphasis on defense in MLB team construction, and/or the improvement of advanced defensive metrics over that period. But the longer-term data still suggests that the next point on the BABIP graph is more likely to be up than down. As for the decline in HRs, that trend is more likely to be sustainable in this (officially) post-PED era. by Ray Murphy Over the past few seasons, we have seen a noteworthy shift in the run-scoring environment in Major League Baseball. These changes may seem at first glance to be a zero-sum game (“if ERA drops across MLB, then everyone’s ERA in my league should drop accordingly”). However, full awareness of these developments may create opportunities to better exploit them in our own league context, and gain an advantage on our opponents. These changes date back to the 2010 season, a year labeled by the mainstream media as the “Year of the Pitcher” (though that was due largely to the spike in no-hitters). Overall, pitchers did make significant progress that year in rolling back the offensive spike of the last decade-plus, and they followed that up with further gains in 2011: <HDU (5$ :+,3 +5 Conclusions Instinctively when it comes to targeting and managing pitchers, we live in fear of the home run. But, in the current power-sapped MLB environment, that fear is not necessarily well-founded. If we shake our fear of the HR, it opens up several doors for us: t *UFYQBOETPVSQPPMPGEFTJSBCMFQJUDIFSTBMXBZTBHPPE thing when we are competing against better-informed and like-minded competitors. t *GXFEPOUGFBSUIF)3XFEPOUIBWFUPGFBSUIFĘZCBMM either. The combination of high strikeout rate and high groundball rate has emerged in recent years as the gold standard in pitcher profiles, perhaps to an excessive degree. When fly balls don’t clear the fence, they are turned into outs with a high degree of frequency. And now that they are clearing fences less frequently, the risk/reward ratio on fly balls is shifting more toward the reward: easy outs. When you see strong skills among our other LIMA criteria combined with a lofty fly-ball rate, that is becoming a skill set to target, rather than avoid. t 1BSLGBDUPSTNBUUFSQFSIBQTNPSFUIBOFWFSćFNPSF a pitcher’s home park is helping him, the more liberties you can take with a pitcher’s skills. This doesn’t just go for San Diego’s PETCO Park; there are plenty of home parks where a fly-ball bias can be an asset to the pitcher. A friendly home park and good team defense can make a pitcher with a low k/9 rate more successful, especially given the league-wide combination of lower BABIPs and fewer HRs. %$%,3 However, some of these 2011 gains may be a mirage. Much was made of the lack of run-scoring in the early part of the season, and the monthly splits bear that out: 0RQWK 0DU$SU 0D\ -XQH -XO\ $XJXVW 6HSWHPEHU (5$ :+,3 There was definitely an early-season dip in run production, but it did not sustain into the second half of the season. In fact, the 2nd half aggregate numbers (4.07 ERA, 1.33 WHIP) were almost identical to 2010 levels. It is also worth noting that the month-over-month trend in 2010 was the exact opposite. That year, pitching improved as the season went on. This further suggests that what we saw early in 2011 was an anomaly, likely fueled by unseasonable weather. The data above also shows that the decrease in run-scoring was driven by two factors: fewer hits on balls in play, and fewer HRs. Addressing the BABIP element first, we can see that oscillation of BABIP around the expected .300 standard is not uncommon: 6HDVRQ %$%,3 55 Statistical Research Abstracts / 2012 Baseball Forecaster Trending Into Trouble, Part 2 # -‐3YT Avg Statistic Players Y1 IP 222 145 xERA 245 4.00 BABIP 155 27 S% 164 78 GB% 74 49 BB/9 158 2.6 K/9 184 8.1 CMD 151 3.1 BPV 203 79 $15 285 4 by Ed DeCaria In 2011 Baseball Forecaster, we studied hitters and pitchers from 2003-2009 riding a positive three-year trend (+3YT) in any of ten statistics. Though tempting to treat such a trend as a sign of sustainable growth, our analysis revealed that +3YT players are actually far more likely to reverse trend than to continue growth into the next season (Y4), and that they typically “give back” most or all of their recent gains in the season following a positive threeyear trend. What about players coming off of a negative three-year trend (-3YT)? Do they show similar trend continuation odds and rebound percentages, or are downward trends more indicative of a true decline in skill than positive trends are of a reaching a new level? The following tables show the average performance trend for all -3YT players in each of ten hitting and pitching categories. “Odds of Trend Continuation” counts the number of -3YT players who rebound in Y4 for every one who continues a falling trend; odds are not weighted by playing time because we are calculating the likelihood of trend reversion on a per-player basis. “Typical Rebound” shows the average percentage of Y2 to Y3 losses recouped from Y3 to Y4 (among all -3YT players regardless of trend continuation). Here are the Year 4 results for -3YT hitters: # -‐3YT Statistic Players AB 253 BB% 131 CT% 134 XBA 137 FB% 50 PX 247 HR/F 147 SX 276 SBO 230 $15 289 Avg Y1 503 11 84 288 41 123 14 92 12 9 Avg Y2 432 9 81 270 38 104 10 75 8 6 Avg Y2 129 4.38 30 73 45 3.1 7.1 2.4 60 1 Avg Odds of Trend Result Typical Y3 Continuation Y4 Rebound 99 1.5 : 1 103 14% 4.85 1.0 : 1 4.84 2% 33 3.0 : 1 31 65% 68 3.0 : 1 74 128% 42 2.0 : 1 44 52% 3.8 2.0 : 1 3.4 53% 6.1 2.0 : 1 6.8 71% 1.9 2.5 : 1 2.3 80% 39 1.5 : 1 53 65% 0 1.5 : 1 2 157% xERA is the only statistic among either +3YT or -3YT hitters or pitchers in which players are just as likely to continue a trend (in this case deliver an even worse xERA) as they are to reverse it. On average, the Y4 xERA is almost identical to the Y3 xERA; combined with a 35% attrition rate, it seems that -3YT xERA pitchers may in fact be reaching bottom, potentially in line for a demotion or involuntary retirement. Conversely, what we think of as mainly (though not entirely) “luck” statistics like strand rate (S%) and hit rate (BABIP or H%) do tend to bounce back in Year 4 following a negative three-year trend. As with hitters, -3YT pitcher rebounds are neither as frequent nor as strong as the +3YT pitcher regressions. The odds are actually decent (on average 2:1) that a three-year decline will continue into a fourth year. Still, we once again see that regression to established performance levels (weighting the most recent season most heavily, but without real regard for trend) can be powerful. When it comes to +3YT and -3YT players, fantasy GMs cannot let their brain’s pattern recognition tendencies get in the way of solid evidence. If Player A is riding a three-year PX trend of 90-100-110 and Player B is riding a three-year PX trend of 130-120-110, then all else equal, Player B is actually the better power play. Your brain may argue “But Player A’s power is on the rise and Player B’s power is in decline!”, but baseball history tells us that A’s true PX skill is likely 100-105 while B’s PX skill may be 115-120. Many examples of misleading multi-year “trends” littered MLB player profiles heading into 2011: t 1PXFSMPTTFTPG-BODF#FSLNBOBOE3VTTFMM.BSUJO t (#JODSFBTFTPG3ZBO#SBVOBOE"TESVCBM$BCSFSB t ##TQJLFTGPS+PTI#FDLFUUBOE.BUU)BSSJTPO t '#SJTFTGPS.BUU(BS[BBOE%BO)BSFO Anyone who wrote off these players passed up on terrific buying opportunities for 2011. Keep this data-driven analysis of trends in mind when scouting players for your 2012 roster this off-season. Challenge any analysis that hints at a player’s demise coming off a negative trend or that suggests an imminent breakout following a positive trend. Before you believe it, dig one level deeper to determine if there is any underlying change (injury, trade, batting stance, arm angle, new pitch, etc.) that might make that particular trend the rare one that continues into a fourth year. Otherwise, more often than not, such predictions simply do not pan out. Avg Odds of Trend Result Typical Y3 Continuation Y4 Rebound 328 1.5 : 1 371 41% 7 4.5 : 1 9 87% 78 4.5 : 1 80 68% 253 2.5 : 1 264 65% 34 5.5 : 1 37 80% 87 1.5 : 1 95 45% 8 2.5 : 1 10 71% 58 1.5 : 1 64 37% 6 2.0 : 1 7 63% 3 2.0 : 1 6 97% Similar to the original +3YT results, we see a clear trend disruption (i.e., a rebound) in all ten metrics among -3YT players. However, the -3YT rebounds are neither as frequent nor as strong as the +3YT regressions. In particular, declining trends in at bats (AB), power (PX), speed (SX), and stolen base attempt rate (SBO) seem to be relatively “sticky” compared to rising trends in those statistics. Rebounds are more likely and/or more complete in walk rate (BB%), contact rate (CT%), fly ball rate (FB%), and total dollar value ($15). Not shown on the chart, but significant, is that the attrition rate (i.e., not playing at all in Y4) for -3YT hitters is twice that of +3YT hitters. Real-life GMs and managers are less likely to give hitters who appear to be in decline a chance to redeem themselves. Here are the Year 4 results for -3YT pitchers: 56 2012 Baseball Forecaster / Statistical Research Abstracts Using PAs to Identify Growth Batters Ramirez (SS, CHW), Denard Span (OF, MIN) and Jay Bruce (OF, CIN)—were over 800 PAs in their first two seasons. The other two were Carlos Gonzalez (OF, COL), who had 621 PAs, eighth in the cohort, and Brett Gardner (OF, NYY), who had 409 PAs but added fantasy value with SBs. What about batters not in their third years? We followed up with a study of every batter whose MLB career began from 200309, seeking a sweet spot irrespective of MLB tenure. We found that batters to target, in addition to those who have two seasons and 800+ PAs, are those who reach 400 PAs early in their careers. We totaled each batter’s cumulative PAs year-over-year. Then we looked at each hitter’s 4x4 R$ performance in the season he reached 400, 600 and 800 PAs. Finally, we compared R$ values in the years before and after a player reached a PA threshold. Our hypothesis was that batters reaching the 800-PA threshold would show the greatest and most consistent gains. And in fact 800-PA batters did show the highest average value in the year they reached the threshold: by Patrick Davitt There’s a truism that batters to target are “age 26 with experience,” but BaseballHQ.com studies suggest we shouldn’t be looking at the candles on the cake, but at the times at the plate—especially in the first two years of their careers. We looked at every MLB hitter from 2003-2007 who: t )BE BU MFBTU QMBUF BQQFBSBODFT UIF TJNQMF "### model) in his rookie year t )BEBUMFBTU1"JOFBDIPGUIFOFYUUXPZFBST t 8BTOPUBOJNQPSUGSPN+BQBO That gave a universe of 204 hitters. We then correlated Year-3 production (using 4x4 Roto dollar values as a proxy) against various factors in the batters’ first two seasons. Age was only modestly correlated, at -0.29, indicating that younger was better. This makes sense because older rookies are not usually top batters. The highest correlate, at +0.53, was combined Year-1 and Year-2 plate appearances. First, here is Year-3 production by age: $JH Q $YH 0D[ 0LQ <HDUUHDFKLQJ $YH5 Q $YH 0D[ 0LQ 3$ 3$ But the 400-PA cohort showed both the largest average dollar gains and number of gainers in the threshold year, and was the best investment for the longer haul: 0HG 3UHWRSRVW 3$ *DLQHUV 'HFOLQHUV $YH*DLQ IURPSUH3$\HDU 3$ 3$ Again, age did not correlate with R$ value in any season at any PA level. The 800-PA cohort fared so poorly in threshold-year-plus-one seasons because nearly one-fifth reached the threshold in their fifth season or later. These are clearly lower-skill and/or part-time batters. As a separate group, these batters averaged just $1.52 (and batters who reached 800 in Years 6-7 returned negative value on average), while batters who reached 800 PAs in Years 2 and 3 had $13.68 average value. Clearly the cohorts to target are batters who reach 800 PAs in their second season, and 400-PA second-year batters for potential sleepers. Remember, these conclusions are based on some small sample sizes, and you should not rely on PA thresholds or other playing time gauges alone. They’re just another tool in your belt. Average outcome declines steadily with age, and age 26 is not a treasure chest. Now the same batters, sorted instead into sexiles (of about 35) by combined Year 1-2 PAs: 3$V 3$ 0HG Average and median value decline steadily with fewer PAs. We had thought batters became more productive at age 26 because they were maturing physically. But most batters arrive at MLB around age 24, so they reach 800 PA in their age-26 seasons. We might have assumed age was the key when it was playing time all along. Also, younger batters getting lots of PAs shows they have earned the trust of team decision-makers, who see real ability. And amassing PAs means young batters are honing their skills against the toughest pitching, while their peers are still learning at AA or AAA. This obvious advantage carries forward; we checked all the rookies’ five-year career total value, and of the 16 who totaled $90 or more, 11 had amassed 800+ PA in their first two years (three more were high-SB specialists). We back-tested with the 2008 crop of rookies, to see if their Year-3 (2010) seasons followed these expectations. Sure enough, 61 rookies met the criteria, but only six earned double-digit Year-3 value, and four of them—Evan Longoria (3B, TAM), Alexei Conclusion In sum: t 8IFO QSPKFDUJOH QMBZFST .-# FYQFSJFODF JT NPSF important than age. t 1MBZFSTXIPBNBTT1"TJOUIFJSĕSTUUXPTFBTPOT are highly likely to have double-digit value in Year 3. t "MTP UBSHFU ZPVOH QMBZFST JO UIF TFBTPO XIFSF UIFZ attain 400 PAs, as they are twice as likely as other players to grow significantly in value. 57 Statistical Research Abstracts / 2012 Baseball Forecaster Skill-Specific Aging Patterns represents the weighted decline in contact rate for each set of players as they aged from 27-to-28, 28-to-29, etc. through age 39. Again, it is critical to establish that even though the age-byage results are chained together to give the appearance of a single curve, relatively few players will follow this exact aging curve from age 22 through 39. Players with careers of such duration are rare and more likely to “age well.” Other aging studies have clearly demonstrated flatter curves for longer-tenured players. Following are the aging patterns for other batting skills: Walk Rate (bb%) trends the opposite way with age compared to contact rate, as batters tend to peak at age 30 and largely remain there until they turn 38. Stolen Base Opportunity (SBO) seems to match speed (the 2011 Forecaster showed the aging curve for Statistically Scouted Speed, which declines fairly linearly from age 23 onward). Players typically maintain their SBO through age 27, but then reduce their attempts steadily in each remaining year of their careers. Success rate also reveals an interesting insight: Stolen Base Success Rate (SB%) has two patterns: one for players who exhibited lower SB attempt rates (<=14% SBO) in their prior three seasons and one for players who exhibited higher attempt rates (>14% SBO). Aggressive runners (solid line) tend to lose about 2 points per year as they age. However, less aggressive runners (dotted line) actually improve their SB% by about 2 points per year until age 28, after which they reverse course and give back 1-2 pts every year as they age. Ground Ball Rate (GB%), Line Drive Rate (LD%), and Fly Ball Rate (FB%) go hand-in-hand. Both GB% and LD% peak at the start of a player’s career and then decline as many hitters seemingly learn to elevate the ball more. But at about age 30, hitter GB% ascends toward a second late-career peak while LD% continues to plummet and FB% continues to rise through age 38. Hit Rate (h%) declines linearly with age. This is a natural result of a loss of speed and change in batted ball trajectory, as line drives have a far higher average hit rate (70-75%) than ground balls (25-30%), which themselves have a higher hit rate than non-HR fly balls (15-20%). Isolated Power (ISO), calculated as slugging percentage minus batting average, typically peaks from age 24-26. Similarly, home runs per fly ball, opposite field HR %, and Hard Hit % all peak by age 25 and decline somewhat linearly from that point on. It is now clear that hitters don’t just magically “peak” at a certain age. Success in baseball requires a variety of skills— patience, pitch recognition, bat control, speed, and power—and growth or decline in any one of these skills can happen at different points in a player’s career. For pitchers, prior research has shown that pitcher value peaks somewhere in the late 20’s to early 30’s. But how does aging affect each demonstrable pitching skill? Strikeout Rate (K/9) declines fairly linearly beginning at age 25. This seems in part due to declines in fastball velocity, but also because that decline in velocity is not matched by an equal drop in changeup velocity: by Ed DeCaria Baseball forecasters obsess over “peak age” of player performance because we must understand player ascent toward and decline from that peak to predict future value. Most published aging analyses are done using composite estimates of value such as OPS, linear weights, or WAR (Wins Above Replacement). By contrast, fantasy GMs are typically more concerned with category-specific player value (HR, SB, AVG, K, WHIP, etc.). We can better forecast what matters most to us by analyzing peak age of individual baseball skills rather than overall player value. In this analysis, we quantify the skill changes of MLB players as they pass from one age to the next. Preceding each written explanation, a micro-chart visually illustrates the aging pattern, where the highest point on each chart represents an average peak age for that skill. The points surrounding that peak on either side show the typical path toward and away from that peak, noting that not all players age the same way (early vs. late development, short vs. long peak, steep vs. flat growth/decline, etc.). Methodologically, each specific point along the path is calculated as the aggregate rate change in skill for all players who played at that age compared to the skill level that those same players exhibited the previous season. Each player-age pair is weighted by the harmonic mean of the plate appearances earned or innings pitched in each of the two seasons; metrics have been normalized to account for season-to-season variance and, in the case of batted ball data, scorer bias. This calculation is done for all player-age pairs for all ages from 22 to 39, and the aggregate rate changes are then chained together on the chart to visualize the peak. We’ll begin with batters. The generally recognized peak age for overall batting value is a player’s late 20s. But individual skills do not peak uniformly at the same time: Contact Rate (ct%) typically ascends modestly by about a half point of contact per year from age 22 to 26, then holds steady within a half point of peak until age 35, after which players lose a half point of contact per year. To better connect the explanation to the micro-chart for this and other statistics, let’s zoom in on the full ct% chart: Contact Rate vs. Peak Ϭ͘ϬϬϬ ͲϬ͘ϬϬϱ ͲϬ͘ϬϭϬ ͲϬ͘Ϭϭϱ ͲϬ͘ϬϮϬ ͲϬ͘ϬϮϱ ϮϮ Ϯϯ Ϯϰ Ϯϱ Ϯϲ Ϯϳ Ϯϴ Ϯϵ ϯϬ ϯϭ ϯϮ ϯϯ ϯϰ ϯϱ ϯϲ ϯϳ ϯϴ ϯϵ Player Age From left to right, you see ages 22 through 39. The peak in this case is age 27, marked as 0.00 on the vertical axis, meaning that the distance between peak ct% and age 27 ct% is 0.00 points of contact. The slope up toward age 27 represents the weighted rise in contact rate for each set of players as they aged from 22-to23, 23-to-24, 24-to-25, and so on. The slope down after age 27 58 2012 Baseball Forecaster / Statistical Research Abstracts Strand Rate (S%) is very similar to hit rate, except since strand rate is a positive for pitchers, the curve moves in the opposite direction. Prior research has shown that strikeouts and walks can affect expected strand rate. The strand rate aging curve reflects that fact, and can be seen visually when aligned to the K/9 and BB/9 patterns above. Strand rate and hit rate are also affected by one other key skill: the ability to control batted ball trajectories. However, the news here is also bad for aging pitchers. Pitch Velocity vs. Peak Ϭ͘ϬϬ Ͳϭ͘ϬϬ ͲϮ͘ϬϬ Ͳϯ͘ϬϬ Ͳϰ͘ϬϬ Ͳϱ͘ϬϬ Ͳϲ͘ϬϬ Ͳϳ͘ϬϬ &ĂƐƚďĂůůsĞůŽĐŝƚLJ ŚĂŶŐĞƵƉsĞůŽĐŝƚLJ &ĂƐƚďĂůůͲŚĂŶŐĞƵƉŝĨĨĞƌĞŶƚŝĂů ϮϮ Ϯϯ Ϯϰ Ϯϱ Ϯϲ Ϯϳ Ϯϴ Ϯϵ ϯϬ ϯϭ ϯϮ ϯϯ ϯϰ ϯϱ ϯϲ ϯϳ ϯϴ ϯϵ Player Age This reduces the gap between the two, and may erode a pitcher’s critical ability to disrupt batter timing. Some may try to compensate by altering their pitch mix to include more breaking balls, but doing so can breed other problems beyond a decline in strikeout rate: Walk Rate (BB/9) improves until age 25 and holds somewhat steady until age 29, at which point it begins to steadily worsen. Whether due to a loss of feel, more movement on breaking balls, or simple fear of the strike zone brought on by declining velocity, more walks and fewer strikeouts equate to markedly weaker command (K/BB ratio) once past the mid-20s. Younger pitchers can compensate for still-developing control with better overall stuff, but pitchers at age 30+ have little fallback. Deteriorating K/9 and BB/9 rates result in inefficiency, as it requires far more pitches to get an out. For starting pitchers, this affects the ability to pitch deep into games: Innings Pitched per Game (IP/G) among starters improves slightly until about age 27 and then tails off considerably with age, costing pitchers nearly a full inning per game by age 33 and another by age 39. This is costly for fantasy GMs, as efficiency and IP/G are two critical drivers of starting pitcher wins. Still, fewer innings per start is not only caused by lack of efficiency, but also a loss of effectiveness. Does aging have any effect on what we widely consider to be “luck” indicators such as hit rate and strand rate? Hit Rate (H%) among pitchers appears to increase slowly but steadily as pitchers age, to the tune of .002-.003 points per year. Some of this is a mirage due to what is called survivor bias, where pitchers with unlucky hit rates one year are given less playing time or not hired at all in the following season. We can readily remove some of that bias by making assumptions about how players would have performed had they gotten more playing time in the year following a poor performance. As a general rule, such players were likely playing below their “true talent” level and therefore were likely to regress back toward some sort of league average level of performance. When compared to their actual recorded prior season, accounting for this hypothetically positive regression is reflected as a delay or flattening of an aging curve such as this, resulting in a “true” H% rise about .001-.002 per year. % of Batted Balls vs. Peak Ϭ͘ϬϬϬ ͲϬ͘ϬϭϬ ͲϬ͘ϬϮϬ ͲϬ͘ϬϯϬ ͲϬ͘ϬϰϬ ͲϬ͘ϬϱϬ 'ƌŽƵŶĚĂůů >ŝŶĞƌŝǀĞ KƵƚĨŝĞůĚ&ůLJ /ŶĨŝĞůĚ&ůLJ ͲϬ͘ϬϲϬ ϮϮ Ϯϯ Ϯϰ Ϯϱ Ϯϲ Ϯϳ Ϯϴ Ϯϵ ϯϬ ϯϭ ϯϮ ϯϯ ϯϰ ϯϱ ϯϲ ϯϳ ϯϴ ϯϵ Player Age Ground Ball Rate (GB%), Line Drive Rate (LD%), and Fly Ball Rate (FB%) again combine to tell a story. Line drives increase slightly beginning at age 24, and because line drives are 2-3 times as likely to fall for hits as ground balls and 4-5 times as likely to fall for hits as fly balls, this spells trouble. From a hit rate perspective, the increase of fly balls with age is a positive, but the obvious consequence is the potential for a home run, yet one last problem that is exacerbated with age: Home Runs per Fly Ball (hr/f) allowed increases with age. A good deal of this perceived impact of aging is removed after accounting for survivor bias (as giving up an abundance of home runs would leave a pitcher especially vulnerable to job loss), but some degradation still remains due to age. Indeed, the calendar is an enemy to major league pitchers. Many skills peak early, and efforts to compensate can lead to other troubles that chip away at efficiency and effectiveness. Though some pitchers are seemingly unaffected by age until well into what should statistically be their decline phase, fantasy GMs looking for breakouts or those just looking for some roster stability should give added weight to a pitcher’s age when comparing alternatives. Resources: Batted ball (e.g., G/L/F, hr/f) analyses use data from Baseball Info Solutions for the 2002-2010 seasons. Fastball and changeup velocity analyses use data from Fangraphs.com for the 2002-2010 seasons. All other analyses are based on the period 1993-2009. 59 Statistical Research Abstracts / 2012 Baseball Forecaster Sustainability of Batter FB% Increases If we turn our attention to the group of players that most successfully added to their power (as measured by slugging percentage and home runs) by increasing their fly ball rates, some interesting patterns emerge: by Brandon Wilson We often think of hitters as fitting into certain categories. There are slap-hitting singles hitters, power-hitting sluggers, and players in between. With several years of balls-in-play data being categorized by groundball, fly ball and line drive, we can now begin to measure whether players shift from one type to another, and what impact it has on their results. Specifically, can a hitter become more of a slugger by increasing the percentage of fly balls he hits? Since we are studying the change in hitter fly ball percentage, the population was filtered down to those with 250 at bats in three consecutive seasons (a base year, the year of increased fly ball percentage and the next year). The idea being to not only determine if hitting more fly balls significantly increases home runs, but to also determine if that increased fly ball percentage (and hopefully the home runs) is maintained in the following year. Once that group of players was defined, we measured only those seasons where the fly ball percentage was increased by at least 5% over the base year. The result was 186 seasons from 161 players in the set of data from 2002-2010 saw an increase at least that large. First, let’s take a look at some basic facts from that group of player-seasons: $OO)%JDLQHUV %DVH\HDU <HDURI)%LQFUHDVH )ROORZLQJ\HDU )% $YJ 6OJ 7RSSRZHUJDLQHUV %DVH\HDU <HDURI)%LQFUHDVH )ROORZLQJ\HDU )% $YJ 6OJ +5 First, in the base year, all of the players that saw an increase of at least 75 points in slugging percentage had a fly ball percentage under 40%. In the year of their jump in fly ball rate, they more than doubled their home run total on average from less than 13 home runs to more than 26 home runs. Like the other groups, these hitters regressed some the next season, but they still saw an increase of more than eight home runs over the base year on average. The success of this group is undoubtedly due in part to growth in other skills and fly ball percentage in isolation is probably not enough to create sustained home run number improvement. In fact, at a certain level, continued growth in fly ball rates becomes counterproductive to overall power production: %DVH\HDU)%! )% %DVH\HDU <HDURI)%LQFUHDVH )ROORZLQJ\HDU +5 $% $YJ 6OJ +5 Players that began with a fly ball rate already over 40% in the base year tended to see a decrease in production as the hit to batting average (which likely played a part in the second year diminished playing time) resulted in lower slugging percentage and fewer home runs. Despite the increase in home runs that resulted from the higher fly ball rate, slugging percentage followed batting average lower. Ultimately the lower total production resulted in playing time reduction and continued decline. A number of the players associated with these numbers were aging players or declining power hitters who appear to have been trying to keep the home run numbers inflated despite declining skills. The average player in the group increased their fly ball rate by 7.7%, starting with an average fly ball percentage in the base year of 33.1% and increasing it to nearly 41%. Since we know that fly balls result in more outs on average than ground balls and line drives, it was no surprise that batting average fell by about 6 points, but that was offset by an increase in slugging percentage of about 20 points. On average, the number of home runs increased by 25%. That increase is partially attributable to playing time, but for some of these players that is part of the benefit of increased FB%, so that net increase of five home runs per season is promising. As usual, however, this group experienced some regression in the following season. While roughly 24% went on to further increase their fly ball percentage the following year and 64% managed to maintain at least half of the increased rate, the group as a whole saw their slugging percentage fall back to near the base year levels. In fact, on average the second year slugging percentage was just two points higher than the base year. The group on average, however, did manage to keep some of the home run gain as the average number of home runs in the second year was more than two higher than the base year in spite of the regression in slugging percentage (the home runs were fly ball driven, not playing time driven). Conclusion In general, increasing fly ball percentage does not lead to long term sustained success. Two useful trends did emerge, however. First, players that start with lower fly ball rates and demonstrate large increases in their fly ball percentage (and in turn slugging percentage) are more likely to hold on to those gains. Second, established power hitters who already have fly ball rates north of 40% that demonstrate another significant increase in fly ball rate often experience falling production in the following year. Otherwise, especially in the absence of other skill indicators, expect regression the year following big increases in fly ball rates for hitters. 60 2012 Baseball Forecaster / Statistical Research Abstracts Hard Hit Balls LWPwr has been down with it. So BHQ invented Power Index (PX), comparing a batter’s power in a particular season with the league-wide power level, normalized to 100. We did the same with Hct%. We divided each batter’s result by the league-wide Hct% average (about 20%), then multiplied the result by 100 to normalize each season to a standard level. Finally, we stacked ‘em up. First, the 2010 results (min 95 AB), with the 135+ HctX: by Patrick Davitt The best way to evaluate batters is by assessing their core skills: the ability to make contact, draw walks and thus control the strike zone. Another core batting skill that is important: the ability to hit the ball hard. This is not the same as “power,” which we measure using results like doubles and home runs. These outcomes depend on the trajectory of the batted ball, opposing defense and ballpark dimensions. They are not necessarily a measure of the actual energy transferred from the swinging bat into the pitched ball. Fortunately, we are now getting data on how hard every player is hitting the ball, including breakdowns of how hard they’re hitting grounders, liners and fly balls. (There are also in-between trajectories being measured, called “fliners,” combining “flies” and “liners” but no such in-between for ground balls, probably because they’d have to be called “flounders.”) We looked at batters’ hard-hit (hh) data for the last few seasons and just before the end of this season, and applied a new metric, hctX, to identify batters with what might the most important skill combination: the ability make hard contact. %DWWHU 0DXHU-RH 3XMROV$OEHUW &DEUHUD0LJXHO 0RUQHDX-XVWLQ %HOWUH$GULDQ 6DQWDQD&DUORV /RZULH-HG <RXNLOLV.HYLQ :HOOV&DVSHU :HOOV9HUQRQ *RQ]DOH]&DUORV 6RIW 0HG +FW; +FW +FW; And under 55 HctX: %DWWHU 2MHGD$XJLH ,]WXULV&HVDU (YHUHWW$GDP $WNLQV*DUUHWW (OOVEXU\-DFRE\ 3HQD5DPLUR $ULDV-RDTXLQ :LOOLWV5HJJLH )LJJLQV&KRQH *DUGQHU%UHWW :KLWHVLGH(OL &URVE\%REE\ 0RRUH6FRWW Hard contact and hit rates MLB-wide: t PGCBUUFECBMMTBSFSBUFEUPIBWFCFFOIJUiTPęwBOE they have the lowest overall hit rate, at 19%; t BSFSBUFEiNFEJVNwXIJDIBSFIJUTPGUIFUJNFBOE t BSFSBUFEiIBSEwXIJDIBSFIJUTPGUIFUJNF Hit rates also vary by trajectory. GB and FB are close in overall percentage, at 26% and 23%, respectively. But when we add hardness of batted ball data, we see some differences: 7UDM *% /' )% +FW Repeating for 2011 (min 80 AB), the highest... +DUG %DWWHU 2UWL]'DYLG %UDXQ5\DQ %HOWUDQ&DUORV %DXWLVWD-RVH &DEUHUD0LJXHO 'RXPLW5\DQ %HOWUH$GULDQ 3XMROV$OEHUW 0F&XWFKHQ$QGUHZ %HUNPDQ/DQFH 3HQD&DUORV *LDPEL-DVRQ 6WDQWRQ0LNH /D5RFKH$QG\ )LHOGHU3ULQFH 0RUVH0LFKDHO 9RWWR-RH\ %XWOHU%LOO\ +DUW&RUH\ So clearly it’s very advantageous for a batter to hit the ball hard. There’s no difference between FB and GB when the ball is hit hard—both become hits at 56%. Ditto for soft. But a medium GB is three times more likely (23% to 8%) to be a hit than a mediumhit flyball, also known as a “can o’ corn.” Something to remember when you look at the G/L/F profiles of batters. HctX So we know that hitting the ball hard means good results. And we already knew that making a lot of contact is also an excellent indicator of batting ability. So we combined them. Midway through 2011, BaseballHQ.com rolled out the metric “HctX.” First, we multiplied hh% by ct% to get a “hard contact percentage” (Hct%), a raw indication of the ability to combine good contact with hard hitting. We found top hitters had Hct%s above 25%, and a handful of really elite guys (Miguel Cabrera, Jose Bautista) were at or above 30%. But, as we’ve seen in the past with power, as measured by the established Linear Weighted Power (LWPwr) metric, batter performance can vary season-to-season because of general trends. For the last couple of years, power has been down, and batters’ 61 +FW +FW; Statistical Research Abstracts / 2012 Baseball Forecaster Line Drives and Luck ... and the lowest: %DWWHU &DPSDQD7RQ\ &RUD$OH[ &RXQVHOO&UDLJ *HW]&KULV %ULJQDF5HLG &DUUROO-DPH\ *RUGRQ'HH &RQVWDQ]D-RVH 6DQFKH]$QJHO 7HMDGD5XEHQ 5HYHUH%HQ %XUULVV0DQQ\ %DUQH\'DUZLQ 6X]XNL,FKLUR *Z\QQ7RQ\ +FW +FW; by Patrick Davitt A few years ago, research at BaseballHQ.com showed that batters’ hit rates did not regress to 30%, a level widely accepted as the de facto standard for pitchers. Each batter sets his own baseline. That finding has long inspired us to ask: Why? It’s obvious that some hitters just hit the ball better than others. Manny Ramirez’ hit rate (h%) was always five or six points higher than Willie Bloomquist’s, despite Bloomquist’s speed advantage. That real-world observation led us to think about Line Drives (LD). Not surprisingly, LDs have a much higher h% than other batted balls. And different batters hit different numbers of LDs. So, since Miguel Cabrera and Matt Holliday hit more LDs than Chris Getz and Aaron Miles, and LDs are hits more often than other batted balls, we would expect them to have higher hit rates and therefore higher batting averages. We checked this idea by looking at LD and h% data from 200710, and partial-year data from 2011. Cumulative data included all batters (at least 1 AB in a season), while individual batter comparisons included only batters with at least 100 AB in a season. Our aim was to look at variance in hit rates on LDs (LDh%), to find standards to identify lucky or unlucky outliers: batters hitting LDs, but over- or underperforming their expected LDh%. Overall, LDh% has been stable over the last four years, from a low of 70% in 2008 to a high 74% in 2007, and averaging 72%. In essence, seven of every 10 line drives fall for safe hits. Among individual batters, LDh% has been relatively consistent around 72-73%, with 95% of all batters falling into an LDh% range of 60%-86%. We’ll consider batters outside those boundaries as our “outliers.” Our 2011 study looked at batters from Opening Day through July 7, and from July 8 through September 9. In the first period, we looked for batters with at least 35 AB and 10 LD, and LDh%s below that 60-86% normal range. We then checked how these outliers did in the second period. All nine batters with at least 10 second-period LDs brought their LDH% up to the normal range, and eight of the nine were within normal range for the year. (The last was one point short.) Six of the seven batters with LDH% higher than the normal range also corrected, some drastically, though only one fell all the way out of the normal range. Note that batters’ BAs did not always follow their LD% up or down, because some of them enjoyed higher hit rates on other batted balls, improved their contact rates, or both.nRemember that the periods studied, at two or three months apiece, are short runs, and regressions don’t always happen over short runs. Still, it’s justifiable to bet that players hitting the ball with authority but getting fewer hits than they should will correct over time. Such a bet certainly makes more sense than hoping for a batting average gain from a bunch of fly balls. Conclusions t )JU3BUFTPO-%TBSFNVDIIJHIFSUIBOPO'#TPS(#T and are generally stable in a range from 70% to 74%. t "NPOH JOEJWJEVBM CBUUFST -%I IBT CFFO SFMBUJWFMZ consistent around 72-73%, with 95% of all batters falling into an LDh% range of 60%-86%. t #BUUFST PVUTJEF UIJT SBOHF SFHSFTT WFSZ RVJDLMZ PęFO within the season. In all four cases, we see pretty decent evidence that the HctX metric is capturing good hitters and slap hitters, which could be very useful in future as we try to project growth players. To that end, we also tracked various actual outcomes with HctX both in-season and season-to-season to check how solidly connected HctX is to the stats that count in fantasy leagues. HctX correlated very strongly with BA, and at higher BA levels, with often eerie accuracy. For HR, we assumed that the only batted balls that could become HRs are hard-hit fly balls (hhFB) and hard-hit line-drives (hhLD). We calculated what percentage of each player’s combined hhFB and hhLD became HRs, then compared 2011 outcomes with 2010 levels to identify outliers but success was more limited. Another obvious area is pitching. If it’s good for a batter to have a high HctX, surely it must be just as good for a pitcher to have a low one. In 2010, the leaders (minimum 90% starts/150 IP) were Jon Lester (68), Adam Wainwright (76), Clayton Kershaw, Jered Weaver and Roy Halladay (79 each). Also boasting good numbers: Ricky Romero, John Danks, Colby Lewis and Jaime Garcia. At the other end: Jeremy Bonderman (123), Nick Blackburn (122), Carl Pavano (119) and A.J. Burnett (114). 62 2012 Baseball Forecaster / Statistical Research Abstracts Evaluating a Potential Refinement to xBA In all but one year, the revised xBA formula predicted current year BA with higher accuracy than the current formula, although the 5-year aggregate difference was not overwhelming. However, we found significant differences between the accuracy of the xBA formulas when we looked at players with different Spd ratings: by Bill Macey The expected batting average (xBA) metric estimates what a player’s batting average might be after stripping out various elements of luck. The formula uses as its inputs a batter’s GB, LD, and FB rates as well as his linear-weighted power and speed (PX and SX, respectively). This formula works very well and is intuitive— players that hit for more power and that have greater speed can be expected to hit for a higher average. Hard-hit ball data provides another measure of a batter’s fundamental underlying skill—the ability to hit the ball hard. Our research has shown that batters who hit the ball harder have higher hit rates and hit for more power. This article introduces and evaluates a potential update to the xBA formula that makes use of this hard-hit data. 6SG 7RWDO 6SG $OO where GBS%, for example, represents the number of soft-hit ground balls as a percentage of all balls put into play, compared to the league average for that year. For example, if 10% of a batter’s balls-in-play were classified as soft-hit ground balls and the league average was 8%, then GBS% would equal 1.25. _ 1HZ _ [%$!%$ [%$%$ _ _ _ _ _ _ &XUUHQW [%$ 1HZ [%$ In general both measures fare equally, and each do less well predicting future BA than they do current-year BA. However, the revised xBA formula appears to introduce additional bias when it comes to projecting future performance. The current xBA overpredicted future BA 48% of the time whereas the new formula rate was 58%. Evaluation and Comparison To measure whether the revised formula is an improvement, we looked at the average absolute difference between xBA and actual BA by year (100 AB min). &XUUHQW [%$ &XUUHQW [%$!%$ [%$%$ In general, the current xBA formula is more likely to predict an xBA less than actual BA whereas the revised xBA formula is more likely to predict an xBA greater than actual BA. Overall, the revised formula appears to be more balanced. Perhaps the most useful application of xBA is as a predictor of future performance. The following summarizes the average absolute difference between xBA in year 1 and actual BA in year 2 for both the current and new formulas. [%$ &7>*%6OQ6SG *%0OQ6SG*%+OQ6SG)%6 )%+/'6/'0/'+ 6DPSOH 6L]H 1HZ [%$ The current xBA does a better job of predicting current BA for the slowest cohort of players, but the revised version seems to be an improvement for most of the population, particularly for the fastest players. If xBA is truly successful in stripping out luck, then we would expect it to be unbiased in its direction. That is, we would expect half of the players to have an xBA greater than their actual BA and the other half to have an xBA less than their actual BA. Approach This revised approach to xBA estimates batting average as a player’s contact rate multiplied by his batting average on all balls in play, inclusive of home runs. In contrast to the current formula, the revised formula calculates batting average on all balls in play using hard-hit ball data rather than PX and it uses Spd rather than SX. There are two theoretical advantages of this revision. First, in the 2011 Baseball Forecaster, Ed DeCaria introduced the metric Spd, which he found to be closely correlated with home-to-first times. One could reasonably believe that faster home-to-first times are positively correlated with batting average. Second, by parsing different categories of GB, LD and FB based upon how hard balls are hit, we should be able to more accurately predict the outcome of those batted balls. For instance, the difference in likely outcomes between a soft-hit fly ball and a hard-hit fly ball is rather dramatic; the former is almost always an out whereas the latter results in a hit more than half of the time. The revised equation is: <HDU 7RWDO &XUUHQW [%$ Conclusion Early indications are that the revised xBA formula presents some potential advantages over the current formula, most notably for faster players. While it’s too early to declare victory (and as such, xBA as shown in the subsequent player boxes does not consider this revised approach yet) we will continue to monitor the efficacy of the new metric. We conclude by illustrating the potential advantage of the revised approach by highlighting two players with whom xBA 1HZ [%$ 63 Statistical Research Abstracts / 2012 Baseball Forecaster has historically had trouble—Derek Jeter and Ichiro Suzuki. The following tables compare their actual BA with their xBA using both the current and revised xBA formulas. <HDU 'HUHN-HWHU &XUUHQW 1HZ [%$ [%$ $FWXDO %$ <HDU ,FKLUR6X]XNL &XUUHQW 1HZ [%$ [%$ $FWXDO %$ start was 86, compared to the yearly average of 95 for the sample of pitchers analyzed. As a result, of the 80 starts analyzed, 27 (34%) were a PQS-0; in all but one of those starts, the pitcher failed to last 5 innings. The percentage of PQS-0 starts following a DL stint is nearly double the average frequency of 18% over MLB in 2009 and 2010. But all injuries are not the same—it’s not unreasonable to expect that a pitcher returning from a strained elbow might pitch worse than a pitcher placed on the DL due to a strained calf. To evaluate potential differences due to injury type, we divided the sample of pitchers into four groups based on a broad categorization of injuries: 7KURZLQJDUP /HJV %DFNFRUH 2WKHU by Bill Macey Whether we play in daily or weekly transaction leagues, one question that fantasy baseball managers frequently struggle with is whether or not to start a pitcher when he first returns from the disabled list. To help guide decision making, this article analyzes pitcher performance, using Pure Quality Starts, in pitchers’ first starts following a DL trip during the 2009 and 2010 seasons. We began by comparing each pitcher’s PQS score in their first post-DL start against his average PQS score for that year. To make sure that the results were meaningful and not just noise, we limited our analysis to pitchers who had at least 15 starts during the year and to those pitchers whose first post-DL appearance was as a starting pitcher (i.e. they weren’t eased back via the bullpen). Using these criteria, we identified 66 pitchers in 80 starts following a return from the DL (several pitchers made more than one trip to the DL) over 2009 and 2010. In aggregate, the performance of this group in their first post-DL start was worse than both their yearly average, as well as the 2009-2010 MLB average. $YJ346 '20 <HDUO\ $YJ346 VW3RVW '/346 '20 ',6 Pitchers returning from arm injuries tended to pitch much worse than normal (for them) in their first start back, earning an average PQS score nearly one full point below their yearly average. Conversely, pitchers seemed to return from leg injuries (broadly categorized to include anything from a groin injury to the toe) without any noticeable dip in performance; if anything, they pitched better, as the 50% PQS-DOM rate shows. The danger of using these broad categories is that injuries to the same body part aren’t necessarily alike; for example, shoulder injuries can vary from a mild case of tendonitis to a torn labrum. To test whether the severity of an injury had any effect on post-DL performance, we examined pitcher performance over different durations of layoffs, as time spent on the DL might be a good indicator of the severity of the injury. Post-DL Pitching Performance &ULWHULD 0/%DYHUDJH <HDUO\DYHUDJHVDPSOH )LUVWSRVW'/VWDUW &RXQW <HDUO\ VW3RVW 'D\VEHWZHHQVWDUWV &RXQW $YJ346 '/346 '20 ',6 /HVVWKDQ %HWZHHQDQG %HWZHHQDQG *UHDWHUWKDQ Sample size begins to become an issue here, but it’s interesting to note that pitchers who missed near the minimum amount of time tended to pitch just as well upon their return to the rotation. In fact, of those 18 starts following a short DL trip (less than 20 days between starts), the DOM/DIS frequencies are remarkably similar to the 2009-2010 MLB averages. ',6 Conclusion So, what are the issues to be considered when debating whether to activate a pitcher for his first start following a trip to the DL? t &YFSDJTF DBVUJPO JO HFOFSBM QJUDIFST QFSGPSN XPSTF than their yearly average in the first post-DL start, with a high rate of PQS-DIS starts. t "WPJE QJUDIFST SFUVSOJOH GSPN UIF %- EVF UP BO BSN injury as they perform significantly worse than average. t *GUIFSFBSFOPCFUUFSPQUJPOTBWBJMBCMFGFFMDPNGPSUBCMF activating pitchers who spent near the minimum amount of time on the DL and/or suffered a leg injury, as they typically perform at a level consistent with their yearly average. That the yearly average PQS score and DOM/DIS rates for the post-DL pitchers are slightly less than that of the general population is no surprise; poor performance over a stretch of several starts frequently precedes a trip to the DL for injured pitchers. But these differences are so small that it’s imprudent to read too much into them. One reason for the low average PQS in the first post-DL start is that many pitchers struggle to last 5 innings in that start, earning an automatic PQS-0. The average number of innings in the first post-DL start was 5.0, about 2/3 of an inning less than the 5.7 IP yearly average for the pitchers evaluated. Consistent with this fact pattern, the average number of pitches thrown in the first post-DL 64 2012 Baseball Forecaster / Statistical Research Abstracts Closer Hunting Season the reliever who started the season as his team’s closer wound up the #2 reliever or worse, while another reliever went on to become his team’s #1 closer for the year. Presumably, in fantasy leagues where saves is a category, nearly every Opening Day closer is rostered before the season begins. Though 70% of these relievers go on to lead their teams in saves, of those seasons in which the Opening Day closer does not end up as his team’s saves leader, over 75% of transitions take place by Game 40. Other relievers may yet earn saves beyond that point, but the big payoffs tend to arrive very early. To make these findings more concrete, turn to the Closer Volatility History chart in the Pitchers section of the Encyclopedia. This chart tallies new saves sources by year. In recent years, there have been about 12 new closers per season, on average. Typically, two of them were their team’s new closer beginning on Opening Day. Another two became closers within the first three weeks of the season (through Game 20). Three more relievers then became closers between Game 21 and Game 40. It then takes another 40 games (until Game 80) to find another three closers. During the entire second half of the season, only two more relief pitchers typically became meaningful saves sources. In addition to seeing more new closers anointed in the early season, relievers who earn their first save early also accumulate more total saves throughout the year. Games 21-40 are particularly valuable in securing long-duration closers, as these 20 games introduce more than twice as many quarter-season, half-season, and rest-of-season closers than even games 40-60 do. Speculating on setup men and other “closers-in-waiting” is not all upside; holding them too long can have negative consequences as well. Sometimes they simply aren’t very good relievers, and can harm WHIP and ERA. More often they have a hidden opportunity cost, whether it means missing a chance to pick up a high-upside rookie, failure to hold an effective backup for an injury-prone star, or falling back in counting categories like wins or strikeouts. Without question, fantasy GMs should make strategic tradeoffs to put themselves in a position to capitalize on closer turnover during its high season. Early on, it may make sense to dump mediocre starters or clear out reserve lists to take as many relief flyers as possible; but if they don’t pop by mid-May, then the best move may be to purge them and convert those roster slots into players that deliver more production than potential. by Ed DeCaria Talent and opportunity combine to create the conditions for a successful closer. There are many ways to evaluate each, but opportunity analyses tend to be more reactive than strategic. That often leaves fantasy GMs on edge all season long waiting for breaking bullpen news that may never come or, if it does, may be too late to be valuable. So when is the best time to hunt for closers? To find out, we identified the closer for each game in which a save was earned from 1996-2010, and assumed that that day’s closer was also the nominal closer in each subsequent game until another save was earned. From this, we ranked each pitcher as his team’s #1, #2, or #3 reliever measured by number of games as closer in a given year, and plotted all 72,000+ games over the period to see if any patterns emerged: Chronological Game Number 1-20 21-40 41-60 61-80 81-100 101-120 121-140 141+ This is actually a 446 x 163 cell matrix, where each row represents a team-season and each column represents a game, in chronological order from left to right. Each black cell represents the #1 closer for a given team-season. The longer the stretch of black in a given row, the longer an individual reliever served as his team’s #1 closer. Team-seasons are ordered from top to bottom by percent of total games in which the team’s top closer held the role. So, seasons in which a single reliever was closer all year are at the top, while seasons in which a team used many relievers as the closer are at the bottom. Nearly the entire top half of the chart shows seasons in which the #1 closer held that role for at least 80% of his team’s games. But as you look further down the chart, you will notice many long stretches of black that begin after Opening Day. This means that 65 Statistical Research Abstracts / 2012 Baseball Forecaster PITCHf/x Location, Movement and Velocity strike zone. This includes over 71% of swinging strikeouts and amazingly nearly half of strikeouts looking (most of these were very close to the strike zone but slightly outside the rulebook strike zone according to PITCHf/x). Most strikeouts are of the swinging variety so getting a batter to chase a bad pitch is a big key to generating strikeouts. By Brandon Wilson The way a pitch is thrown can make all the difference in the outcome. Some pitchers throw hard, some are known for pinpoint location, some have lots of movement and often the best pitchers can mix all three. The question is, can the impact of these attributes be measured? For decades, the only way to gauge was with the eyes, preferably the eyes of a baseball scout. With the addition of PITCHf/x to all big league parks some four years ago, we now have a new set of data to look at that is providing insight into these questions. We looked at each of these attributes individually. Movement Successful pitchers often make their offerings move in such a way that the batter chases the ball out of the zone or fails to “get good wood” on the pitch. PITCHf/x defines how much vertical and horizontal movement the pitch actually had compared to what would have been expected from a ball thrown with no spin. Since how the ball moves is what distinguishes one kind of pitch from another (e.g. a fastball vs. a curveball), movement was studied by type of pitch. As one might expect, curveballs typically feature the largest vertical movement among the different pitch types. Interestingly, the numbers indicate it is not the vertical movement that drives the curveball results; it is the horizontal “x” movement. Location The data shown below is batting average on balls put into play (including home runs) on pitches that were within the strike zone. This grid is a pure look at the left/right and up/down location of the pitch as it crossed the plate and makes no adjustment for handedness of the batter. +LJK 0LG /RZ /HIW &WU 5LJKW &XUYHEDOOVLQSOD\ ODUJHVW|[}PRYHPHQW DYHUDJHPRYHPHQW VPDOOHVW|[}PRYHPHQW The message is fairly clear. The “meaty” part of the strike zone, right in the middle, is where hitters are generally most successful. The same point is even more obvious when viewed in terms of slugging percentage: +LJK 0LG /RZ /HIW &WU 5LJKW 6OLGHUVZLWKODUJH~[VPDOO~] 6PDOOHVW~]OLIW 0LGGOH~]OLIW /DUJHVW~]OLIW The other interesting note from this study is that the pitchers who are worst at avoiding the middle of the plate as a group walk far more batters than the pitchers that consistently avoid the middle of the plate. These pitchers are ostensibly trying to avoid a walk by throwing a pitch over the heart of the plate, but the numbers show that goal is not being achieved. &WO :RUVW %HVW 6OJ For all curveballs thrown, there is not a significant difference in percentage of balls put into play for the pitches with big vertical breaks versus those with small vertical breaks. As such, it seems clear that the horizontal movement is more important. Similar to curveballs, slider results appear to be somewhat driven by horizontal movement. The trend, however, is not as clear cut all the way across the spectrum. Looking at vertical movement and total movement did nothing to clarify the picture for sliders, but looking at a subset of the pitches with the largest horizontal “x” movement and smallest amount of vertical “z” movement did yield some results. With this data, we can say that pitchers who avoid the middle of the plate with consistency should be expected to have better results. Looking at ERA for pitchers at either end of the location spectrum confirmed this: (5$ :RUVW %HVW $YJ $YJ 6OJ While it is a relative small group of pitches from which to draw conclusions in terms of this study (roughly 1,600 balls in play) it appears that there may be a sweet spot for sliders that have a small amount of drop (0-2 inches) and have large horizontal movement. While fastballs are generally considered straight pitches, they typically have some natural horizontal movement and always “rise” more (actually “drop” less) than a pitch thrown with no spin due to the backspin imparted on the ball as it is released. Horizontal movement varies by pitcher and often results in a pitch being classified as a cutter or sinker instead of a fastball when it occurs. As a result, vertical movement was the focus of the study. Simply put, as a group, pitchers that do not avoid the middle of the plate well just do not have very good control. While this may seem obvious, there was some thought that we might find that there was a difference between “control” and “command within the strike zone” with the latter being more important to avoiding hard hits. The numbers tell us that either the two go hand-inhand, or that there are other factors at work. Location analysis also revealed some interesting trends on strikeouts. Over 65% of strikeout pitches are located out of the )DVWEDOOVLQSOD\ ODUJHVW~]PRYHPHQW DYHUDJH~]PRYHPHQW VPDOOHVW~]PRYHPHQW $YJ 6OJ At best, this data suggests it might be better to have either a lot of movement or very little movement rather than be “average”. But 66 2012 Baseball Forecaster / Statistical Research Abstracts even with these numbers, the differences are small enough that movement is probably not as important as location or velocity. Like the fastball, the changeup is generally considered to be a fairly straight pitch, with most of its effectiveness attributed to the difference in the velocity as compared to the fastball. While the expectation was that movement would not be shown to play a significant role, the data showed the amount of vertical movement affected slugging percentage for changeups. In fact, the pitches thrown with the least vertical movement up or down (smallest variance from what gravity alone would impart) --not more drop-- actually performed the best. In addition to small vertical variance, we find that for changeups the horizontal movement also drives better results: &KDQJHXSVLQSOD\ ODUJHVW|[}PRYHPHQW DYHUDJHPRYHPHQW VPDOOHVW|[}PRYHPHQW $YJ Despite the higher batting average, sliders with higher velocity showed a lower slugging percentage just as we saw with fastballs. Looking at the data in a slightly different way produced another interesting pattern, which carried over from year-to-year. 6OLGHUVLQSOD\ IDVWHVWVOLGHUV QH[WVOLGHUV 6OJ Velocity When we think of velocity, fastballs come to mind immediately. The data below shows the average and slugging percentage just for 4-seam fastballs put in play with the pitches grouped by velocity. $YJ &KDQJHXSVLQSOD\ IDVWHVW&KDQJHXSV QH[W&KDQJHXSV 6OJ 6OJ $YJ 6OJ $YJ 6OJ Conclusion By analyzing the PITCHf/x data we found that location, movement and velocity all have roles to play in the pitch outcome. The individual studies did not provide a “magic bullet” answer to measure pitch attributes, but we do leave with a few conclusions. As it relates to location, a pitcher’s ability to avoid the heart of the play should lead to better results. As it relates to movement, curveballs need horizontal movement as much or more than the vertical drop and for fastballs it does not seem to matter much. As it relates to velocity, it is critically important for fastball effectiveness but it does not seem to matter much for curveballs. Throughout all of the studies it became clear that these factors all work together (rather than individually) to help determine whether or not and to what extent a pitch is likely to be successful. 3HUFHQWRI)DVWEDOO6WULNHV3XWLQWR3OD\ IDVWHVWIDVWEDOOV DOOIDVWEDOOV VORZHVWIDVWEDOOV Not surprisingly, the fastballs with higher velocity were put into play less often. In fact, the data shows that at bats where at least one fastball topped 96 mph were 70% more likely to end in a strikeout than at bats where no fastball topped 87 mph. The next set of data reviewed was for sliders which are typically the harder thrown of the breaking ball varieties. The average velocity for sliders was 83.5 mph while the average velocity for curveballs was 76.5 mph. $YJ $YJ As was found with sliders, the fast, but not fastest, group of change-ups was better than average in both years and better than the fastest change-ups. Unlike with sliders, there was not anything as clear as the vertical movement that seemed to correlate with the slugging difference, but the similarity is still interesting. Looking at batting average on balls in play does not do much to identify which is better, but slugging tells the story. The slowest 5% of fastballs (those under 86.6 mph) were more than twice as likely to result in a triple or home run as those in the fastest 5% (over 96 mph). Not only does velocity affect the results of balls put into play, but it has a big impact on whether or not the ball ever makes it into play. 6OLGHUVLQSOD\ IDVWHVWVOLGHUV DOOVOLGHUV VORZHVWVOLGHUV 6OJ With sliders the fastest pitches (sliders exceeding 88.7 mph) did not produce the best results. While they did produce better results than average speed or slow sliders, especially in terms of slugging, it was actually the next group of sliders (those between 86.3 and 88.7) that performed better. This may have something to do with what scouts call the “flattening out” of a breaking pitch. This “second fastest” group of pitches show about an inch more vertical drop on average than the pitches in the fastest group. Furthermore, the pitches hit for home runs in the group of fastest sliders showed two to three inches less vertical drop than average for the group. The conclusion appears to be faster is better, but not so fast that the pitch loses its shape. For curveballs, velocity proved to be a much less important attribute. Recall that with fastballs movement did not prove to be important, generally speaking. Likewise for curveballs, the fastest and slowest varieties did not show much difference in the batted ball numbers with any consistency. At first glance the changeup data seemed to be as inconclusive as the curveball data. But a second pass at the data showed a pattern similar to sliders in the data. The numbers are not dramatic, but the trend is readily apparent: as with curveballs, more horizontal movement yields better results. As further evidence, when we look at the subset of pitches with the least vertical movement which also fell in the category of greatest horizontal movement, the resulting balls in play had a .277 average and a .321 slugging. )DVWEDOOVLQSOD\ IDVWHVWIDVWEDOOV DOOIDVWEDOOV VORZHVWIDVWEDOOV $YJ 6OJ 67 Statistical Research Abstracts / 2012 Baseball Forecaster Zack Greinke and the Fallacy of Perfect Information a correction, but if those numbers have generated a 5.63 ERA up until now, what’s to say that they won’t continue to do so? Players have gone entire seasons with aberrant variances between their stats and their peripherals; what’s to say that he won’t be one of them? Personality #3: Maybe a 5.63 ERA is not for real, but I’ve been watching him pitch and he looks awful. He may be striking out lots of batters and preventing base-runners, but he keeps giving up hits at the worst times. It’s like he loses complete composure with men on base. That is what his strand rate is really saying and so I’m not sure I’m buying yet. Personality #4: Maybe a 5.63 ERA is not for real, but when will it turn around? I can’t take a chance making a roster move because I don’t know when things will change. It hasn’t in two months; will it happen in July? August? Not until September? I am not going to make a move until I see signs of a turnaround. Personality #5: Maybe a 5.63 ERA is not for real, but how much of a correction will there be? Once he does turn things around, will that correction be a 2.00 ERA? 3.00 ERA? 4.00 ERA? If he turns it around with a 4.00 ERA in late August, that does me no good. It only makes sense to act if we’re talking about a 2.50 ERA in July. Can you guarantee that? Personality #6: Maybe a 5.63 ERA is not for real, but how do I know if an apparent turnaround is for real? What if Greinke goes out and has two great starts? That’s a tiny sample size from which to draw the conclusion that his season has officially turned around. How do we know that he won’t regress into his funk again? How many good starts does it take for us to conclude that a turnaround is real and sustainable? Personality #7: I agree with the peripherals but dispute the conclusion. If the pure skills gauges are so telling, why hasn’t he been able to overcome his high hit rate and low strand rate? Couldn’t those superb skills alone have yielded a better ERA during the first two months? His peripherals are amazing so there has to be some hidden flaw somewhere that we can’t explain. Personality #8: I’m all-in. I agree with the peripherals. I agree with the conclusion. I expect a turnaround any day now and will look to acquire Greinke from Personalities #1-7. Personality #9: I’m all-in and cannot be dissuaded. Not only am I all-in, but I will ignore any other dissenting information that might temper my decision. Those numbers are such a powerful statement that I need consider no other opinion. And I will stop at nothing to acquire Zack Greinke. I am willing to bet that, before you started reading this list, you probably considered that Greinke was a clear BUY candidate. You might have subconsciously considered yourself Personality #8 or #9. After reading about all the shades of grey, perhaps you’ve backed off on that assessment. But “perfect information” is supposed to level the playing field and take away our competitive advantage, right? It’s not so black-or-white. There will always be a few guys in your league who do things their own way. Perhaps one is a Royals fan and is angry that Greinke left Kansas City. Perhaps another still carries scars from by Ron Shandler Many analysts believe that today’s widespread accessibility to advanced data means that we are all using the same information to run our fantasy teams. Access to sabermetrics and ADPs has leveled the playing field at draft time; real-time news feeds have given us the same tools for in-season tactical planning. To some, this means that we are getting close to a situation of play that economic game theorists call “perfect information.” When we reach that point, any competitive advantage has to come from somewhere else. There is much truth in this. However, I think there are still advantages to be gleaned from the information itself. Even if we are all reading the same book, everyone will be interpreting the book differently. Some people will take the information at face value, some will find nuance, some will ask questions, some will draw different conclusions and some will downright reject the data. So I believe that, even if there is such a thing as “perfect information,” it doesn’t matter. People are going to read the information differently and act according to their own belief systems. Some level of competitive advantage is still potentially achievable from the information alone. Take Zack Greinke. Greinke got off to a horrible start after coming off the disabled list early this season. By the end of June, he had a 5.63 ERA. But perfect information told us that this was all a mirage. He had an 11.5 strikeout rate, a 1.7 walk rate and an insane 6.8 command ratio. He had an inflated 37% hit rate and a stark 56% strand rate, both crying for a correction. His BPV was otherworldy at 180. Using advanced—though now mainstream—baseball metrics, this was the perfect storm of leading indicators. Anyone looking at these numbers would be projecting a major improvement in his ERA. Note how easy it is for you to agree with that statement now that you know what happened next. But imagine if it was still late June and you didn’t know for sure. Would you still be 100% all-in? If this was perfect information, then everyone should have interpreted it the same way. There should have been a mad rush to roster him. But more likely, there was a range of reaction to Zack Greinke’s bad start. Call it the “9 Personalities of Perfect Information.” Personality #1: His 5.63 ERA is for real because I don’t look at peripherals. We don’t play this game with strikeout rates and BABIP. As much as those gauges have proven to be predictive, they are not 100% predictive; they just suggest a tendency. For my fantasy team, wins, strikeouts, ERA and WHIP are all that matter and Greinke just hasn’t put up those numbers. Yes, he might improve because he’s a better pitcher than this, but I want to see it for myself first. Personality #2: His 5.63 ERA is for real because, well, that’s what he’s done. Results on the field are far more meaningful and all I see is a 5.63 ERA. Yes, his peripherals are shouting for 68 2012 Baseball Forecaster / Statistical Research Abstracts rate and depressed strand rate still indicated he should be doing much better. Both had leading indicators that pointed to huge upside in 2011. Yet only one met expectation. Shields posted a 2.82 ERA while maintaining an identical 107 BPV. Nolasco’s ERA stayed flat at 4.67; his hit and strand rates remained stuck in neutral. His BPV declined to 87, though still a solid level. The perfect storm of information that so accurately nailed Greinke’s rebound is now somewhat less certain. Now take Brandon Morrow. He has posted 100-plus BPV levels for two consecutive seasons, yet all he has to show for it are ERAs of 4.49 (2010) and 4.72 (2011). His hit rate has averaged 33% for the two seasons; his strand rate 66%. Once again, it’s a perfect storm of predictive information. What do we project for 2012? What’s your personality? the years when Greinke burned him, and won’t dip back into those waters. Perhaps another just hates Brewers. You can find an edge in the most unusual of places, even when everyone has access to the same information. As for Zack Greinke, the rest is history. He posted a 2.80 ERA with 121 strikeouts in 109 innings during the second half of the season. His peripherals were slightly off of those he posted in the first half, but still at an elite level (his BPV slipped to 135). His “luck” factors swung back—31% hit rate, 80% strand rate—but it was clearly the right decision to be all-in on his peripherals... ...in retrospect. This is the same decision you would have had to make last spring for pitchers James Shields and Ricky Nolasco. In 2010, Shields posted a horrible 5.18 ERA but had an elite level BPV of 107, a 35% hit rate and a 68% strand rate. It was the second consecutive season that his ERA had risen from his previous career-best 3.56 mark in 2008. In 2009, Nolasco posted a 5.06 ERA but a BPV of 129. He improved his ERA to 4.51 in 2010, but his 118 BPV, elevated hit 69 GAMING RESEARCH ABSTRACTS The Science of the First Round $'3 $&78$/),1,6+ $OEHUW3XMROV +DQOH\5DPLUH] 0LJXHO&DEUHUD 7UR\7XORZLW]NL (YDQ/RQJRULD &DUORV*RQ]DOH] -RH\9RWWR $GULDQ*RQ]DOH] 5\DQ%UDXQ 5RELQVRQ&DQR 'DYLG:ULJKW 0DUN7HL[HLUD -RVK+DPLOWRQ 5R\+DOODGD\ &DUO&UDZIRUG )LUVWWLPHLQILUVWURXQG by Ron Shandler An inordinate amount of time is spent analyzing the first round in snake drafts (and on $30-plus players in auctions). Who should go first? Should Adrian Gonzalez go before Troy Tulowitzki or after? Should any pitchers even be considered that early? I suppose there is some merit to it; after all, these players represent the foundation of our rosters. And it should be fairly easy to identify who these players are. After all, the best players should be the most projectable. However, history suggests otherwise. From the Encyclopedia: t 5XPUIJSETPGQMBZFSTĕOJTIJOHJOUIF5PQXFSFOPUJO the Top 15 the previous year. t 0GUIPTFXIPXFSFĕSTUUJNFSTGFXFSUIBOSFQFBU in the first round the following year. t &TUBCMJTIFETVQFSTUBSTXIPĕOJTIJOUIF5PQBSFOP guarantee to repeat. Ten of these top 15 players were returnees from the previous year’s top 15 actual finishers. Five players were first-timers to the first round. Only one player (Teixeira) finished lower than the second round in 2010. And if you look at some early ADPs for 2012, you’ll see that 2011 stars Jacoby Ellsbury and Curtis Granderson are new round #1 headliners. In the first high-level draft this past fall, 2011 fantasy MVP Matt Kemp went #1. If we all agree that it’s foolish to project next year’s numbers from last year’s numbers alone, why did 2010 have such a huge impact on our expectations for 2011? And why is 2011 doing the same thing to 2012? The most recent season is a strong point of reference, but drafters tend to use it as an anchor (in the negative sense of the word). We couldn’t imagine that a player like Carlos Gonzalez would regress very far after such an amazing 2010 performance. We said the same thing about Joe Mauer in 2009. We’ll say the same thing about Jacoby Ellsbury and Curtis Granderson this spring. But year-to-year statistical performances are that volatile. When it comes to baseball forecasting, I’ve often said that regression and gravity are the strongest forces known to man. Our first round ADPs do an amazing job of defying these forces. How can we fix it? How can we assemble a more appropriate listing of early round targets? 1. Cast a wider net. Based on the past few seasons, it is reasonable to expect that about 80% of actual first round performances will come from players drafted from the first four rounds. So we need to carefully select these important core players from a pool of about 60. The first step in doing this is to… 2. Remove the shackles of ADPs. ADPs, at best, are deceptive. At worst, they can be downright dangerous. We want ADPs to represent what the market is thinking, but what they have become is a self-fulfilling prophesy. Previous ADP listings create an artificial anchor (there’s that word again) for player value, and each successive mock draft digs that anchor deeper and deeper into the ocean floor. For that reason, we have to stop using ADPs to shape our own draft day decisions. Use them only to get a rough sense on what others may be thinking. Make your player selections independently. Last year’s results continue to bear this out: $&78$/),1,6+ 352-5' 0DWW.HPS -DFRE\(OOVEXU\ 5\DQ%UDXQ -XVWLQ9HUODQGHU &OD\WRQ.HUVKDZ &XUWLV*UDQGHUVRQ $GULDQ*RQ]DOH] 0LJXHO&DEUHUD 5R\+DOODGD\ &OLII/HH -RVH%DXWLVWD 'XVWLQ3HGURLD -HUHG:HDYHU $OEHUW3XMROV 5RELQVRQ&DQR This was actually one of the better years with six of the 15 finishers projected to finish in the first round. Thirteen of the 15 (87%) came from the projected top four rounds, or top 60 players, a rate that has remained steady for several years. This seems acceptable until you realize it creates a wide error bar; essentially, anyone drafted in the first four rounds is fair game for a first round finish. That potentially changes how we look at the early-round drafting process... ...and which is why the next table is a little disturbing. The players who finish in the top rounds one year are very strong bets to be projected to finish in the top rounds the following year. Note the Average Draft Position rankings at the beginning of this past season: 71 Gaming Research Abstracts / 2012 Baseball Forecaster 3. Embrace regression. Roto players often talk about chasing “last year’s bums” as part of their draft strategy. There is wisdom in this, as regression does rule. But it works in reverse as well, and far more often. Avoiding “last year’s heroes” is probably a more prudent strategy. It’s a basic concept that goes back to Bill James’ Plexiglass Principle (“If a player improves markedly in one season, he will likely decline in the next.”). I was heeding this advice when I drafted Ryan Braun with the #2 pick in the Fantasy Sports Trade Association experts league draft last January. This was not intended to thumb my nose at the importance of the first round. It was intended to pick the player I thought had the potential to post the best 2011 statistics, after Albert Pujols. My thought process was outlined in this essay that appeared on Big Lead Sports last spring… 12 Reasons why Ryan Braun should be drafted #2 behind Albert Pujols by Ron Shandler (March 2011) Finally, the seven players currently blocking him from #2: 6. Hanley Ramirez (current ADP #2): His numbers have declined for two years straight. His 51 percent groundball rate will significantly depress his power potential. 7. Miguel Cabrera (#3): Even before the DUI arrest, his career year in 2010 would be bound to regress. Now with the uncertainty of his alcohol problem, he’s way too risky to draft this high. 8. Evan Longoria (#4): His down power year in 2010 looks like a correction from his 2009 HR spike. His declining home run to flyball rate makes it look like he’ll plateau as a 25-HR hitter. That’s not enough for an ADP this high. 9. Troy Tulowitzki (#5): If not for his ridiculous—and unrepeatable—15-HR September, would he even be in this discussion? 10. Carlos Gonzalez (#6): Bill James’ Plexiglass Principle says CarGo’s 2010 season has regression written all over it. Add in his .396 BABIP and the fact that only 14 percent of first-time first-rounders repeat and you can see the odds stacked against him. 11. Joey Votto (#7): See Carlos Gonzalez. If history holds, less than 40 percent of these first round players will finish in the Top 15. Regression and gravity are the two strongest forces known to man. 12. Adrian Gonzalez (#8): He has never finished a season ranked higher than 30th. This year’s ADP surge is based solely on the speculation of huge numbers in Boston. But he’d need a 40-120-.300 season to return this much value—something he’s never done - and his shoulder issues make it uncertain if he’ll even be ready for Opening Day. That leaves Ryan Braun. There is one possible downside to drafting Braun at #2: his position. There are lots of valuable outfielders out there. But years of experience has proven this: the first two rounds in a fantasy snake draft must have the overriding goals of the highest possible productivity and the lowest possible risk, even trumping position scarcity. Next to Pujols, Braun is a darn good fit. Last month, I drafted Ryan Braun with the #2 pick in the Fantasy Sports Trade Association experts league draft. Yes, only Albert Pujols was off the board, and with the world of available players before me, I chose Braun. It created a firestorm of commentary, speculation and condemnation. His current ADP ranking at Mock Draft Central is #9. In the hundreds of drafts tracked by MDC, Braun has never been selected higher than 4th. But Braun is a defensible pick at #2. Here are 12 reasons why: 1. There were explanations for his down performance in 2010. He didn’t get along with former Brewers manager Ken Macha. He was distracted by his outside business interests. This past fall, he supposedly hired a business manager and committed to focus more on baseball in 2011. He backed up that statement by increasing his workout regimen during the winter and showed up to camp with five pounds of additional muscle. 2. As “bad” as his season was, he still posted more than 100 RBI for the fourth year in a row, hit a career-high 45 doubles and batted .364 from August 1 on. 3. Despite his depressed power numbers in 2010, we have to avoid using those numbers as a point of reference. Braun’s peripherals were still very solid. His ADP was in the Top 5 going into last year’s drafts. After a single down year, Braun would be expected to improve from normal regression alone. 4. Milwaukee! What an exciting team this is looking to be in 2011. Manager Ron Roenicke is allegedly going to run more, which could boost Braun’s stolen base totals. This is a potential 30-20 player. The odds of the Brewers being strong contenders would mean more runs scored and more RBIs as well. 5. We can’t get married to the current ADPs. As the number of mock drafts conducted reaches critical mass, the rankings become self-perpetuating. Drafters rarely stray from the established ranges even if a player deserves a significant upgrade, or downgrade. For that reason, Braun’s market value might be anchored at #9 but that is an unreliable measure of his real value. 72 2012 Baseball Forecaster / Gaming Research Abstracts ADP and Pre-Rankings (631 &%6 0'& by Bill Macey Average Draft Position (ADP) data has become an increasingly popular tool in draft preparation. It helps you to develop a general sense of how the market values different players and allows you to identify potential bargains. While useful, over-reliance on ADPs can be dangerous. ADP data typically comes from mock drafts held online and results differ, sometimes substantially, from one source to another. A primary reason why ADP results differ between sources is that each website can have significantly different default rankings for players. This manifests itself in two ways. First, the mock drafts used to generate the ADP lists might not fully fill, leaving “artificial intelligence” (AI) drafters to fill the empty spots. Typically these AI teams will select the top player on the draft board, subject to positional need, but driven by the default rankings. This phenomenon prevents any higher-ranked players from falling too far down in the draft. Also, it is very common for drafters to drop out of a mock early. This often happens as early as rounds 8-10. Subsequent picks then revert to the default rankings, further anchoring the ADPs to a particular site’s defaults. Second, even if the draft is full of active human participants, it’s difficult to not be influenced by the names that are at the top of the draft board, which serve as a constant not-so-subtle reminder that a particular player is still available. Likewise, it’s quite easy to forget about a player you like if you have to scroll far down the pre-rank list to find him. Out of sight, out of mind. To see how strong these effects were, we compared ADP versus default ranking at three major online draft sites: ESPN, CBS, and Mock Draft Central (MDC). Data was collected February 28, 2011. Consistent with expectations, we found that the ADP values closely mirrored each site’s particular default rankings. t "U&41/BNPOHUIFUPQSBOLFEQMBZFSTUIFBWFSBHF difference between default ranking and ADP was just 2.8 places. 163 of the top 175 ranked players were drafted within 15 places of their ADP. t "U$#4UIFBWFSBHFEJČFSFODFCFUXFFOEFGBVMUSBOLJOH and ADP for the top 200 ranked players was 4.6 places. 129 of the top 200 players had an ADP within 15 places of their ADP. t "U.%$UIFBWFSBHFEJČFSFODFCFUXFFOEFGBVMUSBOLJOH and ADP for the top 200 ranked players was the largest at 7.4, but 166 of the 200 had an ADP within 15 places of their ADP. Clearly, the default rankings greatly influence ADPs. Another way to illustrate this is by looking at players with divergent default ranks across the different sites. A good early round example of this is Miguel Cabrera. Reacting to news about his arrest for driving under the influence and the subsequent concerns about its implications for his playing time and performance, ESPN moved him down to #13 on their Top 300 list. By comparison, CBS and MDC made no such adjustment and he remained as the #3 and #4 player per their respective default rankings. 5DQN $'3 We can see that even for a widely known player whose name was been all over the news, the ADP results still closely followed pre-rankings; ESPN drafters moved him down their draft board, but he still went within the first 5 picks at CBS and MDC. The phenomenon continued later in drafts. Consider Adam LaRoche, another player whom the three sites value differently. Again, we see that the mock drafters followed the default ranks closely: (631 &%6 0'& 5DQN $'3 This isn’t always the case, though. Another player similarly slotted by the sites, Ian Desmond, saw his ADP converge. (631 &%6 0'& 5DQN $'3 For both of these players, as well as other players with divergent default rankings, their ADPs trended slightly towards a consensus middle ground. That is, while their ADPs are strongly influenced by their default rank, there is some evidence that they are pulled towards an industry average. How can we best use this information on draft day? t *G ZPVS MFBHVF JT IPMEJOH BO POMJOF ESBę UIF "%1T provided by that service are your most reliable source for identifying likely bargains. Relying on another site’s ADPs may influence you to draft a player earlier than necessarily or miss out on him altogether. (Although, if you really want a player and can’t afford to not get him, it’s best to ignore ADP entirely and take him where you value him.) t 4QFOE UJNF XJUI UIF ESBę TPęXBSF ZPVMM CF VTJOH PO draft day and adjust your pre-rankings to reflect your own preferences such that the players you prefer are closer to the top of the list. You don’t want to forget a player you really like simply because time was tight and you didn’t scroll far enough down the list of available players. t *G UIFSF BSF BOZ QMBZFST ZPV BSF LFFOMZ JOUFSFTUFE JO drafting, consider looking at ADPs from a variety of sources. If there are vastly different ADPs for those select players, continue to use the ADPs provided by the site hosting your draft as a rough guide for their value, but place less confidence in that data. t *GZPVSESBęJTMJWFJUNBZCFXPSUISFWJFXJOHUIF"%1T from multiple sources to see where there is general consensus and where players diverge. Certainly don’t take ADPs from a single source as a reliable estimate as to how your draft will proceed. 73 Gaming Research Abstracts / 2012 Baseball Forecaster Xtreme Regression Drafting (XRD) 5LFN\5RPHUR 'DYLG3ULFH 0DWW7KRUQWRQ &DUORV*RQ]DOH] 1LFN6ZLVKHU &DVH\0F*HKHH &OD\%XFKKRO] -RKQ%XFN '/HGSOD\HUV by Ron Shandler In the essay, “The Science of the First Round,” I talked about how regression and gravity are the two strongest forces known to man. I wrote how chasing “last year’s bums” and avoiding “last year’s heroes” was important in determining an appropriate list of early draft targets. But I don’t think we’ve taken it far enough. Year-to-year player performance does not inherently follow a smooth trend for any player. Rather, surges are followed by declines, and declines by surges, even as overall growth may still occur. These swings are constant and far wilder than we’d care to admit. The greater the movement in one direction, the higher the likelihood of a regression or normalization the following year. Bill James nailed this all years ago. What this means for us is that we might need to downgrade the great majority of players coming off of up years, and carefully target those who are coming off of down years. Perhaps we could build an entire draft strategy around that. Can it work? We have the tools today to help us assess. For starters, we have Average Draft Position (ADP) rankings. While most folks use these to set their own draft plans, we can use annual ADP lists to create our initial pool of “bums” and “heroes.” Then, using the Mayberry Method, we can identify specific players who might be the best bets to improve, and do so at a reduced cost. Let’s look at two lists. The first is a list of the players whose 2011 ADP rank was at least 100 positions higher than their 2010 ADP rank. These are the players who had the most positive change in perceived value coming into this season, or “last year’s heroes.” I’ve screened out all players who were either rookies coming into 2010 or who missed that season with injury. Unranked players from 2010 (-NR-) were assigned a ranking of 300. -RVH%DXWLVWD %XVWHU3RVH\ 7UHYRU&DKLOO &KULV<RXQJ $XEUH\+XII 'HOPRQ<RXQJ %UDQGRQ0RUURZ .HOO\-RKQVRQ -DVRQ+H\ZDUG $OH[*RQ]DOH] 3KLOLS+XJKHV *LR*RQ]DOH] 0DUWLQ3UDGR 'UHZ6WXEEV %UHWW0\HUV )/LULDQR &KULV3HUH] 5LFNLH:HHNV 3DXO.RQHUNR $GULDQ%HOWUH %UHWW*DUGQHU 6FRWW5ROHQ 15 15 15 15 15 15 15 15 15 $'3 *DLQ 15 $YJ ([FOXGLQJ'/ Of these 30 players, a whopping 28 (93%) returned less value in 2011 than they did in 2010, in many cases much less. Their average dollar return was off $14 (still $13 if you exclude the nine players who’ve spent significant time on the DL this season). The second list is the flipside—those whose 2011 ADP rank fell more than 80 positions short of their 2010 ADP rank. These players came into 2011 with the most negative change in perceived value, or “last year’s bums.” I’ve screened out all players who were expected to miss most or all of the 2011 season with injury. /DQFH%HUNPDQ -DNH3HDY\ %ULDQ)XHQWHV 'HUUHN/HH $-%XUQHWW &DUORV%HOWUDQ -DYLHU9D]TXH] (ULFN$\EDU $GDP/LQG $DURQ+LOO 0DUN5H\QROGV *RUGRQ%HFNKDP 0LFKDHO&XGG\HU -DVRQ%D\ -RKQQ\'DPRQ 2UODQGR+XGVRQ -RVH/RSH] 5DXO,EDQH] 5XVV0DUWLQ &DUORV3HQD $GDP-RQHV -RVK%HFNHWW &KULV,DQQHWWD -RQDWKDQ%UR[WRQ &DUORV4XHQWLQ 3DEOR6DQGRYDO %ULDQ5REHUWV '/HGSOD\HUV 5 $'3 /RVV $YJ ([FOXGLQJ'/ 5 Of the 27 players listed, 11 (41%) returned more value in 2011 than they did in 2010. Their average dollar return was down slightly by $2. But injuries to eight of the remaining 16 players dragged down the results significantly. Without those eight, 11 of 14 (79%) are returning more value with an average profit of $5. The first key takeaway is that public ADP lists can be powerful tools, but not in the way that they are intended. By targeting those players who show the largest change in public perception from one year to the next, we can get incredible insight into where the potential for true profit and loss lie. 74 2012 Baseball Forecaster / Gaming Research Abstracts Jordan Walden and Kevin Jepsen. The winner of the Angels Closer Sweepstakes turned out to be the Walden owners, but it made sense to spread the speculations. We need to do that more with position players. Go an extra buck or two for high-skilled players in volatile situations. These players are not only rosterable, but will likely be in demand at some point. And, they will typically come at a discount. Mayberry skills grade “3” or better. 5. Speculate on injured players. Risk-averse drafters will stay away from the chronic hobblers, but these are exactly the types of players that often return big profit. Budget your DL speculations to maybe two or three at most, but the Erik Bedards and Carlos Beltrans have their place. Remember - you are buying roster spots, not players, so any time lost to the DL can still be backfilled. Just try to focus on players at positions where the free agent pool might be a bit more lucrative. Mayberry health grade “C” or worse. 6. Do the best you can with playing time. The challenge with targeting imperfect players is that they are not typically projected to get full-time plate appearances or innings. PAs and IPs are vitally important so you will need to mix in as many regulars as possible. 7. Finally, if you are going to truly practice extreme regression, then you also have to ignore all players coming off peak performances in 2011. For good or bad, their prices are all going to be benched in their 2011 numbers, and very likely inflated, even by a dollar or two. So for 2012, you have to forget that Jacoby Ellsbury had 30-HR power. You have to forget about Curtis Granderson and Lance Berkman and Jered Weaver. Odds are strong they won’t repeat 2011’s numbers. Some will argue that there is regression already built into their draft price; I’ll argue that any adjustment won’t be nearly enough. Ideally, what you’re looking to build is a roster filled with highskilled regulars who are experienced, healthy and erratic, coming off a down year. In Mayberry terms, you’re looking for solid skills indicators, a “5” for playing time, reliability grade of “B” or better for experience, and a “C” or worse for consistency. Budgeting your draft dollars at an auction is important, but it’s not about how much you spend as it is about how you spend it. So, for XRD, the traditional LIMA $200/$60 split will not apply. If you have to spend $150/$110 to get the specific combination of players, then that’s what you spend. So how do we put these revelations into practice? Xtreme Regression Drafting (XRD) is a draft planning framework built on this year-to-year performance volatility. 1. Embrace volatility to the extreme. We are constantly looking for patterns in the data but often latch onto trends that may not be real. Research has shown that the next point in the data series 2,4,6 is more likely to be 3 than it is to be 8. That is so counterintuitive to everything we’ve ever assumed. What this tells us is that a straight line is a terrible percentage play. Of course there are exceptions. When we see a player like Felix Hernandez break out in 2009, then improve in 2010, we want to find more of those. But King Felix is the exception rather than the rule. There are far more Zack Greinkes and Tommy Hansons who surge, then regress, then possibly surge again. That is the trend we should be projecting. And so, those are the trends that we need to look for. The last character of the Mayberry grade measures a player’s consistency. In the past we’ve targeted players with high grades (A or B), particularly for your Portfolio3 Tier 1 foundation commodities. For XRD, look for players with Mayberry grade “C” or worse, along with solid skills indicators, who were coming off a down year. This is intended to isolate northward regression, if you will. 2. Draft a small core. Foundation players don’t need to fill many roster slots, but it’s good to have a few who you can count on. Even with XRD, you still want to find players with a few dollars of potential upside so target anyone who may have had a slightly down year in 2010. The best names for 2012 are players like Carlos Gonzalez, Evan Longoria and even Albert Pujols. 3. Target skills, as always. This is our typical routine, identifying players whose underlying skills show much better potential than their surface stats. In some cases, these players may grade out as an “A” in consistency because their skill level is actually flat. But we’re buying the statistical upside here, so go all in on a few. 4. Keep a very open mind on roles. We often say, “draft skills, not roles.” But that is often tough to do even when you see a good player blocked at a position. We have to come to terms with the fact that roles are fluid over the course of the season. Incumbents will get hurt. Backups will get opportunities. And we have to invest in those backups who have good skills. That’s the only way we can back into this year’s Mark Trumbo. We already do that when it comes to closers. Coming into 2011, we knew that Fernando Rodney was risky, so we threw some dollars (sometimes significant dollars) at Scott Downs, 75 Gaming Research Abstracts / 2012 Baseball Forecaster Black Swans and Ugly Ducks Arredondo, Henry Rodriguez, Craig Stammen, P.J. Walters and Sammy Gervacio. Not an inspiring group, but it only cost me $30. Spending $230 on bats provided a very good offense in a league where hitters are very hard to get via FAAB. The bats drafted were Carlos Ruiz, John Buck, Lance Berkman, Chris Johnson, Kelly Johnson, Rafael Furcal, Matt Kemp, Corey Hart, Nyjer Morgan, Raul Ibanez, Ben Francisco, Carlos Pena, Geoff Blum and Ryan Howard. But a big batter budget doesn’t guarantee a sweep of the hitting categories, of course. Berkman’s rebound was something of a Black Swan event, and along with the super-productive Kemp and Howard, provided a lineup core that stayed largely healthy. That was important. FAAB brought new names aboard, such as Darwin Barney on offense, and pitchers Taylor Buchholz, Scott Linebrink, Jeff Fulchino, Antonio Bastardo and Jeff Karstens. I was dead last in pitching a couple weeks in and had to be active to reconstruct this mess. That’s the rub with Black Swans (or ugly ducks) in general. We expect players to perform in line with historic data—and the historic data for this group projected underwhelming results. But that’s where LIMA came into play. By going with high strikeout rates and command ratios, and low hr/9, I positioned my staff to take advantage of their skills upside. For instance, Salas didn’t even make the Opening Day roster in St. Louis, but the mantra “draft skills, not roles” helped me roster him nearly two full months before he got promoted to the closer role. And then I hit the trade wire. A big early deal sent Raul Ibanez for Ian Kennedy and Daniel Murphy, before it was apparent the type of numbers Kennedy would rack up. By August, the offense was so far ahead of everyone else’s that I was able to deal Ryan Howard for Cole Hamels, and Carlos Pena plus Salas for Wandy Rodriguez. (That was more fortunate timing, as Salas lost his closer role shortly thereafter.) Via trades, aggressive use of FAAB, and some good timing, that bunch of misfits somehow turned into a pitching staff that returned a very respectable 42 of 65 points. The $230 offense nailed 57 of 65 points. This group of “ugly ducks” pulled away in early summer and won LABR by 12.5 points. In the end, it was fun to win with a $30 pitching staff, but the good fortune that accompanied several of these decisions was a key contributor to my success. In redraft leagues with so many intelligent owners, it is the luck of getting the most ugly ducks on a roster (regardless of strategy) that makes one owner a winner one year, and a different owner a winner the next year. Still, the premise behind the plan was to maximize the opportunity to find those outlier performances. In that sense, the plan— both LIMA and Ugly Ducks—did its job. By Doug Dennis On the BaseballHQ.com forums and in the spring First Pitch Forum road show, we discussed the phenomenon of “black swans.” This is a concept written about by philosopher and mathematician Nassim Taleb in his book The Black Swan: A Black Swan is an event with the following three attributes. First, it lies outside the realm of regular expectations, because nothing in the past can convincingly point to its possibility. Second, it carries an extreme impact. Third, in spite of its outlier status, human nature makes us concoct explanations for its occurrence after the fact, making it explainable and predictable. Preparing for this year’s LABR-NL experts draft, it occurred to me that it might be possible to maximize the odds of capturing a Black Swan. At the very least, I could try to collect a handful of ugly ducks with upside potential. The road to the duck pond was a willingness to embrace the inherent league scarcity. LABR is a 13-team redraft league with very deep player penetration. On the offense side, they draft 182 players. In the National League, there are only 128 starters; that leaves 54 holes for parttime bats. Considering the approximately 208 total hitters on NL teams, LABR’s offensive player penetration is a high 88%. Meanwhile, LABR requires 130 pitchers. A usual mix might be 78 starters, 36 middle relievers and 16 closers. The NL uses 80 starters—many of which you might not want—16 closers and about 96 middle relievers. That’s 192 pitchers. LABR’s penetration here is lower 68%. Given that scarcity mismatch, I decided to put my usual LIMA budget on steroids: $230 on offense, a closer for $15, and backfill the rest of the pitching with leftover money. The successful implementation of this approach would require sorting through the bargain bin of pitchers effectively, while reserving enough money to both buy the league’s best offense, plus hopefully provide some surplus offense to trade from. This strategy is hardly a new idea. It is basically a hybrid of the LIMA plan and Larry Labadini’s $9 pitching staff. I was also reminded by BHQ’s Josh Paley of a Roger Anderson observation made over a decade ago: in LABR competitions, the amount of money spent on pitching has had zero correlation to the amount of roto points those teams’ staffs accumulated. At the draft, my team was a laughingstock. By the end of the day—and mind you, this draft is held the first week of March—my pitching staff consisted of Leo Nunez (a closer, check), but then Bud Norris, Jon Niese, Dillon Gee, Jordan Lyles, Ross Detwiler, Esmil Rogers, Joe Thatcher, Chad Qualls and Fernando Salas. My 6-man reserve roster was all pitchers—Nelson Figueroa, Jose 76 2012 Baseball Forecaster / Gaming Research Abstracts Consistency in H2H leagues starting pitcher accumulates in a given week. Next, we calculate the standard deviation of weekly points on a playing-time adjusted basis to measure the raw level of consistency for each player. Finally, we divide each player’s weekly standard deviation by their weekly average, to produce a scaled measure of consistency. The following example underscores the importance of this last step: using the scoring system above, Albert Pujols produced 676 points in 690 plate appearances in 2010, totaling approximately 0.98 points per plate appearance. The standard deviation of his weekly performance was 0.39 pts/PA. By comparison, Mark Reynolds produced 376.5 pts in 582 PA, or approximately 0.65 pts/PA, and the standard deviation of his weekly performance also equaled 0.39. However, compared to Pujols, this same standard deviation represents a larger percentage of his average weekly performance. So while we may measure Pujols’ weekly consistency at 0.39, we measure Reynolds’ consistency at 0.60. Thus, a lower value indicates increased consistency. For batters with at least 100 AB, the league median in 2010 (min 100 AB) was 0.63 and Peter Bourjos led the league with a value of 0.24 (although unfortunately, he was consistently mediocre). Among starting pitchers with at least 100 IP, the league median was 1.02, and David Price led the league with a value of 0.40. Results We then measured each player’s weekly consistency against various measures of skill. Using linear regression, we looked at bb%, ct%, PX, SX and FB%, and found a statistically significant relationship between consistency and each of the variables except for FB%. In other words, an improvement in any of the first four skills reliably correlated with an increase in weekly consistency. Among those four variables, ct% and bb% had the strongest relationship with weekly consistency. This is consistent with the results of our previous research and the reasoning still seems to hold: the more often the batter is able to put a ball into play (or at least not make an out by way of taking a walk), the more opportunities he has to score points. And regardless of a player’s skill level, we expect greater consistency when there are more opportunities if only because luck has more time to even itself out. We performed a similar analysis was for starting pitchers, looking at the relationship between weekly consistency and the skills Dom, Ctl, hr/9 and GB%. Similar to the results for batters, Dom, Ctl and hr/9 were all correlated with consistent weekly performances at a statistically significant level. The strongest relationship was with hr/9, which makes sense given the scoring system outlined above. Depending on the number of runners on base, even a single home run can dramatically affect the outcome of a game, and the 12 point difference between a win and loss is a huge swing, even when measured on a per-inning basis. Conclusion The takeaway from this is that not only do we expect players with better skills to post more overall points, but we also expect more consistent performances on a week to week basis, even after accounting for differences in total points scored and playing-time. Therefore, when drafting your batters in points-based headto-head leagues, ct% and bb% make excellent tiebreakers if you are having trouble deciding between two players with similarly projected point totals. Likewise, when rostering pitchers, favor those who tend not to give up home runs. by Bill Macey As fantasy football’s popularity has continued to spike and newfound fantasy fanatics have crossed over from one sport to another, points-based head-to-head fantasy baseball leagues have also enjoyed an increase in popularity. Previous research has demonstrated that week-to-week statistical consistency is important for head-to-head play. However, such research typically focuses on Rotisserie-based H2H games. This article focuses on points-based games. The table below shows the default points in head-to-head leagues on CBSSports.com, a popular league host; these are also the points used to calculate the subsequent analysis. (YHQW 3RLQWV 6LQJOH 'RXEOH 7ULSOH +RPH5XQ 5XQ 5%, :DON 6WULNHRXW +LW%\3LWFK 6WROHQ%DVH &DXJKW6WHDOLQJ :LQ /RVV 4XDOLW\6WDUW 6DYH (DUQHG5XQ ,QQLQJ3LWFKHG 6WULNHRXW :DON +LW +LW%\3LWFK General Strategy An important difference between the two types of head-tohead leagues arises from a strategic standpoint: in category-based head-to-head leagues, you try to avoid punting a category such as stolen bases or saves as it would give your opponent an easy tactical advantage. But in a points-based league, there is no need for balance as no individual category is more important than another—only total points matter, regardless of how they are accumulated. Generally, in any points-based H2H contest, you should seek to maximize both plate appearances and innings pitched when there is a positive expected value associated with each additional AB or IP. This is certainly true using the scoring system above. However, when trying to decide between two players with similar projected point totals, you should try to maximize consistency for the reasons outlined in the previous research. Therefore, we want to know which skills correlate with weekly consistency, measured using totals points accumulated. Measuring Consistency The first step in measuring week-to-week consistency is to calculate the weekly points scored by each player on a plate appearance or IP basis. This scaling is done to normalize the effects of playing time. This is particularly important for starting pitchers as two-start weeks can drastically affect the number of points a 77 Gaming Research Abstracts / 2012 Baseball Forecaster Realistic Expectations of $1 Endgamers The Formula for Consistent Success, redux by Patrick Davitt by Ron Shandler Many fantasy articles insist leagues are won or lost with $1 batters, because “that’s where the profits are.” But are they? A 2011 analysis shows our main focus in managing $1 endgamers should be more about minimizing losses than fishing for profits in a goldfish bowl. This analysis applies only to batters in deeper leagues, where 90% of batters are taken at draft: standard 13-team NL- and 12-team AL-only leagues, and 24-team mixed leagues. BaseballHQ.com projected 278 batters for $0 to -$5 from 2008-2010, and 372 batters for $1 seasons. We compared those batters’ BHQ projections with their actual outcomes. In the $0 to -$5 cohort, 82% returned losses, based on a $1 bid. Average production was -$2. Worse, 29% returned double-digit losses, while only 9% provided double-digit profit. Two-thirds of the $1 cohort returned losses. The cohort averaged $2 in production, which looks modestly useful, but that average was boosted by the 38 (11%) $1 batters who provided double-digit profits; median production was $0. There was a bit more reason for optimism: While losers outnumbered gainers 2:1, only 12 in the entire group lost in double digits. Clearly, fishing for profit in the sub-zero group is a mug’s game, like hoping to strike it rich playing penny stocks. Nonetheless, when circumstances force us into this part of the pool, as circumstances often will, the best bet is to speculate on speed. Big sub-zero winners in 2010 included batters like Michael Bourn, Juan Pierre, Erick Aybar and Jose Tabata, all of whom created extra value applying their good wheels to added PT. Similarly, double-digit profit-makers in the $1 cohort included burners like Nyjer Morgan, Rajai Davis, Scott Podsednik, Andrew McCutchen, Dexter Fowler and even a generally worthless bat like Willie Bloomquist, who hit 4-29-.247, but swiped 25 bags. Watch out for catchers. In the sub-zero pool, 25 of the 82 double-digit losers, and 16 of the bottom 44, were backstops. Five of the 12 double-digit losers in the $1 cohort were also CAs. If your roster features a few $1 endgamers, it’s not impossible to earn useful or even league-winning profits. But it is unlikely, given the ratios of winners to losers. Draft speed, then monitor for playing time and stolen base opportunity. Is his team running? Is he running? Is he successful? Nothing kills a marginal speed guy’s playing time faster than running into rally-killing outs. Also, monitor the $1 endgamers on your opponents’ rosters for signs of life—a little extra PT here and there, more SBOs, poor performance by batters in front of them on their MLB teams. You might be able to spot a profit opportunity early, and get it for a very low price as a trade negotiation throw-in. In sum, when considering $1 players in deep leagues: t .BOBHJOH FOEHBNFST TIPVME CF NPSF BCPVU minimizing losses than fishing for profits t *OUIFDPIPSUPGCBUUFSTQSPKFDUFEUPSFUVSOFE losses, based on a $1 bid. Two-thirds of the projected $1 cohort returned losses. t 8IFOGBDFEXJUIQMBZFSTTQFDVMBUFPOTQFFE In last year’s book, I reported on the results of study that attempted to determine what it takes to be consistently successful in this game. Anybody can win a title in a given year, but what does it take to win year after year? I identified six variables that I believed were the most important contributors to success, then assembled a group of consistently successful, high-level fantasy players to rank them. Here were the aggregate results from this exercise (best possible score would be 6.0): 5$1.9$5,$%/( 6&25( %HWWHULQGUDIWVWUDWHJ\DQGWDFWLFV %HWWHUVHQVHRISOD\HUYDOXH %HWWHUOXFN %HWWHUJUDVSRIFRQWH[WXDOHOHPHQWVWKDWDIIHFWSOD\HUV %HWWHULQVHDVRQURVWHUPDQDJHPHQW %HWWHUSOD\HUSURMHFWLRQV In short, having the best draft is most important; having the best player projections is least important. And it would have ended there if not for one random comment that was tucked away in the back of the report. A potential seventh variable was brought up by (four-time Tout Wars winner) Larry Schechter: “I would add something to your list, called ‘putting in the work’ or ‘spending time preparing.’ One of the big reasons for my success is taking the time to be very prepared for an auction/draft; and taking time during the year to stay current on news, contextual elements, making roster moves, FAAB, etc. Obviously, just spending a lot of time for the sake of spending time doesn’t do any good. You need to work smart. One of the reasons I have better in-draft strategy, better sense of value, etc., is because I take the time to figure it all out.” At first, I saw this as a throwaway comment. Of course you have to spend the time! But then others starting coming out of the woodwork, similarly trumpeting the importance of “time.” Successful National Fantasy Baseball Championship player Mark Srebro posted the following comment in the BaseballHQ.com subscriber forum: “The single biggest tool to increase your fantasy baseball success is adding more time and using it efficiently. I may or may not be one of the smarter guys in the NFBC, but I’ll be damned if someone outworks me. “Back in my early years in the NFBC, I thought I could get by with how I approached and dominated my local leagues. However, after a couple years of being a slightly above average NFBC owner, I made a decision to do whatever it takes to win. I poured in at least triple the time. It took me a full year to learn the player pool. It took two years to feel 100% confident with the pool. However, once you acquire that knowledge, it puts you in a favorable position from the start.” Srebro’s success in the NFBC is notable so you can’t ignore his assessment. The 2011 Fantasy Baseball Guide magazine had an article by the winner of the inaugural Cardrunners League. This was a league that “pitted five established fantasy baseball experts against 78 2012 Baseball Forecaster / Gaming Research Abstracts So, for 2011 I decided to reclaim the one resource that I’ve let fritter away over the past decade. I took back my time and limited my involvement to just three leagues -- Tout Wars, XFL and FSTA. Time allowed me to do some interesting things. In Tout Wars, I spent $62 on Adam Dunn, John Lackey and Matt Thornton, which was enough to put my team 65 points out of first and 30 points out of 11th by Memorial Day. Ordinarily, I would have been far too stretched out to continue, but this year I had time. I battled back, gaining nearly 40 points over the rest of the season and finishing 8th, but just 4.5 points out of a 5th place money spot. In the XFL keeper league, I built my power game around Adam Dunn and Alex Rodriguez, and settled back into the bowels of the standings by late spring. Ordinarily, I would have started playing for 2012, but this year I had time. I embarked on a mid-season Sweeney Plan, dealing away all HRs and RBIs for pitching and speed. I finished 4th. I was in contention for most of the season in the FSTA, but faded late to finish in 5th. Sometimes, even time won’t help, especially in non-trading leagues. But it was one of the most fun and fulfilling years I’ve had in a decade. five challengers from the worlds of professional poker and Wall Street.” Champion Eric Kesselman was a member of the latter group and summed up his winning strategy thusly: “I decided that the biggest advantage I had over the expert field was the sheer amount of energy I could focus on the league. I played in only two leagues, whereas most of the experts played in many more. Playing 10 or more leagues, the experts had a hard time devoting full attention to any particular league.” Anyone who has the ability to beat you over the head with time is going to have a huge advantage no matter what else he does. In retrospect, that’s why I advise my readers against joining a league where the owners were at different stages of their lives. A college student with abundant discretionary time would have a huge advantage over a 30-something with parental duties. Looking back over my own performance track record in national competition, I am seeing a clear correlation between my success and the amount of time I have devoted to my leagues. During the four year period of 1998 to 2001, I finished in the Top-3 seven times. But back then there was only Tout Wars, LABR and the Fantasy Sports Trade Association experts league, and I was participating in only 2-3 each year. During the next decade, the number of invites to public leagues exploded and I had a tough time turning down opportunities to compete. There have been years in which I participated in upwards of seven leagues. It has taken me 10 years to amass another seven Top-3 finishes. 79