26th Edition - BaseballHQ.com

Transcription

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
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
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
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
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
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
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
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
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
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
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
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 $ZD\
3XOO
)LHOG
Opposite field HRs serve as a strong indicator of overall home
run power (AB/HR). Power hitters (smaller AB/HR rates) hit a
far higher percentage of their HR to the opposite field or straight
away (over 30%). Conversely, non-power hitters hit almost 90% of
their home runs to their pull field.
20
version) may be used as a leading indicator for stolen base output,
SB attempts are controlled by managerial strategy which makes
speed score somewhat less valuable.
Speed score is calculated as the mean value of the following
four elements:
1. Stolen base efficiency = 6%6%&6[
2. Stolen base freq. = 6TXDUHURRWRI6%&66LQJOHV%%
3. Triples rating = %$%+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
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
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
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
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
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
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
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
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
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
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
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
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
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
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
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$/
There is some potential talent in older players—age 26, 27 or
higher—who, for many reasons (untimely injury, circumstance,
bad luck, etc.), don’t reach the majors until they have already been
downgraded from prospect to suspect. Equating potential with
age is an economic reality for major league clubs, but not necessarily a skills reality.
Skills growth and decline is universal, whether it occurs at the
major league level or in the minors. So a high-skills journeyman
36
<HDU
0/(FW
<HDU
0/%FW
There is very little difference between the median MLE BA in
Year 1 and the median MLB BA in Year 2:
0/(FW
727$/
<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
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
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
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
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
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 37.5%.
42
$'3
$OH[5RGULJXH]
$OEHUW3XMROV
&DUORV%HOWUDQ
7RGG+HOWRQ
9ODG*XHUUHUR
$OIRQVR6RULDQR
1*DUFLDSDUUD
%DUU\%RQGV
3HGUR0DUWLQH]
0DUN3ULRU
0DQQ\5DPLUH]
5R\+DOODGD\
0DJJOLR2UGRQH]
(GJDU5HQWHULD
6DPP\6RVD
$'3
$OH[5RGULJXH]
&DUORV%HOWUDQ
$OEHUW3XMROV
9ODG*XHUUHUR
0DQQ\5DPLUH]
%REE\$EUHX
0LJXHO7HMDGD
-RKDQ6DQWDQD
7RGG+HOWRQ
-DVRQ6FKPLGW
5DQG\-RKQVRQ
&DUO&UDZIRUG
$OIRQVR6RULDQR
%HQ6KHHWV
&XUW6FKLOOLQJ
$'3
$OEHUW3XMROV
$OH[5RGULJXH]
9ODG*XHUUHUR
0DUN7HL[HLUD
0DQQ\5DPLUH]
0LJXHO&DEUHUD
'HUUHN/HH
%REE\$EUHX
0LJXHO7HMDGD
&DUO&UDZIRUG
0LFKDHO<RXQJ
&DUORV%HOWUDQ
-DVRQ%D\
'DYLG2UWL]
'DYLG:ULJKW
$'3
$OEHUW3XMROV
$OIRQVR6RULDQR
-RVH5H\HV
$OH[5RGULJXH]
5\DQ+RZDUG
-RKDQ6DQWDQD
&DUO&UDZIRUG
&KDVH8WOH\
&DUORV%HOWUDQ
'DYLG2UWL]
9ODG*XHUUHUR
'DYLG:ULJKW
0LJXHO&DEUHUD
/DQFH%HUNPDQ
&DUORV/HH
$&78$/ ,FKLUR6X]XNL
9ODG*XHUUHUR
5DQG\-RKQVRQ
$OEHUW3XMROV
-RKDQ6DQWDQD
%REE\$EUHX
$GULDQ%HOWUH
%DUU\%RQGV
&DUORV%HOWUDQ
%HQ6KHHWV
0HOYLQ0RUD
&DUO&UDZIRUG
0DQQ\5DPLUH]
0LJXHO7HMDGD
7RGG+HOWRQ
$&78$/ 'HUUHN/HH
$OH[5RGULJXH]
$OEHUW3XMROV
'DYLG2UWL]
0DUN7HL[HLUD
&DUO&UDZIRUG
&KRQH)LJJLQV
-DVRQ%D\
0LJXHO&DEUHUD
0DQQ\5DPLUH]
0LFKDHO<RXQJ
9ODG*XHUUHUR
,FKLUR6X]XNL
%REE\$EUHX
-RKDQ6DQWDQD
$&78$/ -RVH5H\HV
'HUHN-HWHU
$OEHUW3XMROV
5\DQ+RZDUG
-RKDQ6DQWDQD
$OIRQVR6RULDQR
&DUO&UDZIRUG
0DWW+ROOLGD\
9ODG*XHUUHUR
0LJXHO&DEUHUD
,FKLUR6X]XNL
&KDVH8WOH\
*DUUHWW$WNLQV
-HUPDLQH'\H
/DQFH%HUNPDQ
$&78$/ $OH[5RGULJXH]
+DQOH\5DPLUH]
0DWW+ROOLGD\
0DJJOLR2UGRQH]
'DYLG:ULJKW
-LPP\5ROOLQV
,FKLUR6X]XNL
-RVH5H\HV
-DNH3HDY\
'DYLG2UWL]
&DUO&UDZIRUG
(ULF%\UQHV
%UDQGRQ3KLOOLSV
&KLSSHU-RQHV
3ULQFH)LHOGHU
$'3
$OH[5RGULJXH]
+DQOH\5DPLUH]
'DYLG:ULJKW
-RVH5H\HV
0DWW+ROOLGD\
-LPP\5ROOLQV
0LJXHO&DEUHUD
&KDVH8WOH\
5\DQ+RZDUG
$OEHUW3XMROV
3ULQFH)LHOGHU
5\DQ%UDXQ
-RKDQ6DQWDQD
&DUO&UDZIRUG
$OIRQVR6RULDQR
$'3
+DQOH\5DPLUH]
$OEHUW3XMROV
-RVH5H\HV
'DYLG:ULJKW
*UDG\6L]HPRUH
0LJXHO&DEUHUD
5\DQ%UDXQ
-LPP\5ROOLQV
,DQ.LQVOHU
-RVK+DPLOWRQ
5\DQ+RZDUG
0DUN7HL[HLUD
$OH[5RGULJXH]
0DWW+ROOLGD\
&KDVH8WOH\
$'3
$OEHUW3XMROV
+DQOH\5DPLUH]
$OH[5RGULJXH]
&KDVH8WOH\
5\DQ%UDXQ
0DUN7HL[HLUD
0DWW.HPS
3ULQFH)LHOGHU
0LJXHO&DEUHUD
5\DQ+RZDUG
(YDQ/RQJRULD
7RP/LQFHFXP
-RH0DXHU
'DYLG:ULJKW
&DUO&UDZIRUG
$'3
$OEHUW3XMROV
+DQOH\5DPLUH]
0LJXHO&DEUHUD
7UR\7XORZLW]NL
(YDQ/RQJRULD
&DUORV*RQ]DOH]
-RH\9RWWR
$GULDQ*RQ]DOH]
5RELQVRQ&DQR
5\DQ%UDXQ
'DYLG:ULJKW
0DUN7HL[HLUD
&DUO&UDZIRUG
-RVK+DPLOWRQ
5R\+DOODGD\
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:
$&78$/ $OEHUW3XMROV
-RVH5H\HV
+DQOH\5DPLUH]
0DQQ\5DPLUH]
0DWW+ROOLGD\
'DYLG:ULJKW
/DQFH%HUNPDQ
'XVWLQ3HGURLD
5R\+DOODGD\
-RVK+DPLOWRQ
$OH[5RGULJXH]
&&6DEDWKLD
&DUORV%HOWUDQ
*UDG\6L]HPRUH
&KDVH8WOH\
$&78$/ $OEHUW3XMROV
+DQOH\5DPLUH]
7LP/LQFHFXP
'DQ+DUHQ
&DUO&UDZIRUG
0DWW.HPS
-RH0DXHU
'HUHN-HWHU
=DFK*UHLQNH
5\DQ%UDXQ
-DFRE\(OOVEXU\
0DUN5H\QROGV
3ULQFH)LHOGHU
&KDVH8WOH\
0LJXHO&DEUHUD
$&78$/ &DUORV*RQ]DOH]
$OEHUW3XMROV
-RH\9RWWR
5R\+DOODGD\
&DUO&UDZIRUG
0LJXHO&DEUHUD
-RVK+DPLOWRQ
$GDP:DLQZULJKW
)HOL[+HUQDQGH]
5RELQVRQ&DQR
-RVH%DXWLVWD
3DXO.RQHUNR
0DWW+ROOLGD\
5\DQ%UDXQ
+DQOH\5DPLUH]
$&78$/ 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
<HDU
3FWRIWRSUDQNHGSOD\HUVZKRORVW37
When 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
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
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
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
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
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
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
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%$SSUR[
XSWR
XSWR
XSWR
XSWR
XSWR
Integrating Mayberry and Portfolio3
Mayberry scores can be used as a proxy for the BPI filters in
the Portfolio3 chart. Pretty much all you need to remember is the
number “3.”
The third character measures expected batting average.
[%$
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$VFRUH[.VFRUH6DYHVVFRUH,3VFRUH
[,3VFRUH6DYHVVFRUH
51
very powerful, as a single week might impact the standings by
over 1.5 points in ERA/WHIP, positively or negatively.
t PG CBUUFST IBE QPPSFS JOWFTUPS SFUVSOT CVU XJUI BO
aggregate batting average virtually identical to the true BA.
Sitting stars and starting scrubs
Consistency
(Ed DeCaria)
In setting your pitching rotation, conventional wisdom suggests
sticking with trusted stars despite difficult matchups. But does
this hold up? And can you carefully start inferior pitchers against
weaker opponents? Here are the ERA’s posted by varying skilled
pitchers facing a range of different strength offenses:
23326,1*2))(16(5&*
3LWFKHU(5$
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
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
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
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
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
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
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
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$QG\
)LHOGHU3ULQFH
0RUVH0LFKDHO
9RWWR-RH\
%XWOHU%LOO\
+DUW&RUH\
So clearly it’s very advantageous for a batter to hit the ball hard.
There’s no difference between FB and GB when the ball is hit
hard—both become hits at 56%. Ditto for soft. But a medium GB
is three times more likely (23% to 8%) to be a hit than a mediumhit flyball, also known as a “can o’ corn.” Something to remember
when you look at the G/L/F profiles of batters.
HctX
So we know that hitting the ball hard means good results. And
we already knew that making a lot of contact is also an excellent
indicator of batting ability. So we combined them.
Midway through 2011, BaseballHQ.com rolled out the metric
“HctX.” First, we multiplied hh% by ct% to get a “hard contact
percentage” (Hct%), a raw indication of the ability to combine
good contact with hard hitting. We found top hitters had Hct%s
above 25%, and a handful of really elite guys (Miguel Cabrera,
Jose Bautista) were at or above 30%.
But, as we’ve seen in the past with power, as measured by
the established Linear Weighted Power (LWPwr) metric, batter
performance can vary season-to-season because of general trends.
For the last couple of years, power has been down, and batters’
61
+FW
+FW;
Line Drives and Luck
... and the lowest:
%DWWHU
&DPSDQD7RQ\
&RUD$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
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
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
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
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
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
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
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
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
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
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
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
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
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

26th Edition - BaseballHQ.com

Transcription

Similar documents

Untitled - West Ham United

October 1st - Registration Opens

GestureXtreme

Step by step instructions for players registering for Pelada FA

Welcome to CARHA Hockey`s online registration. Below you will find

ExamView Player

Land Before Time Make A Match Rules

USA Baseball Host Program

to learn more

Publisher: DMD 2500 Target Rating: Teen Developer: Mytosis Jones