Friday, August 27, 2010

Baseball Study: Can a few terrible starts conceal a gem?

I'm no mathematician, but I would love it if someone out there shares my curiosity for baseball and can (dis)prove some of my theories.

Basically, I would be the idea (people) person; you will be the brains.

Thesis #1: If we remove a starting pitcher's worst two starts (as defined by lowest Game Score), we may uncover a few hidden gems - pitchers that perform very well most times out but have their stats - sabermetric or otherwise - crushed by these terrible performances. [Quality Start % does something similar, but I think this could yield results that give us more insight; if it helps, make up a fancy moniker like xQS%.]

ESPN* attempts to quantify each player's contribution to a particular game by assigning a Game Score. For pitchers, the formula for reads:

50 + 1*Outs + 2*IP after the 4th + 1*K - 2*H - 4*ER - 2*UnearnedRuns - 1*BB

The average Game Score this season for starting pitchers is approximately 50.8.

Doing some very rudimentary math, I filtered for any pitcher that made 10 or more starts through 8/19 (157 in total). I then removed his lowest two Game Scores and re-sorted by average Game Score.
w/o 2 = average Games Score rank after removing two worst starts
Pure = average Games Score rank including all starts

Games Startedw/o 2PureGmSc Diff
Felipe Paulino1476103-27
Scott Olsen1295122-27
Luke Hochevar1381106-25
Homer Bailey1078102-24
Brett Anderson101437-23
Jair Jurrjens144464-20
Joel Pineiro205878-20
Todd Wellemeyer11109129-20
Hisanori Takahashi1294113-19
Tim Wakefield16112131-19
Vicente Padilla152441-17
Madison Bumgarner104257-15
Carlos Zambrano12117132-15
Thomas Hanson253044-14
Tommy Hunter146679-13
Dana Eveland10141153-12
Jake Peavy175667-11
Tom Gorzelanny18938211
Randy Wells25978611
Mark Buehrle251018912
Jonathan Sanchez25473413
Livan Hernandez25604713
Kevin Correia231109614
Clayton Richard25877215
Jake Westbrook241069115
Kris Medlen14917318
Jon Garland25715120

Takeaways: If not for a couple brutal starts, Tommy Hanson would be getting talked about like an ace. Padres pitchers have been incredibly consistent.

w/o 2 = average ERA rank after removing two worst starts
Pure = average ERA rank including all starts
Games Startedw/o 2PureERA Diff
Scott Olsen1279128-49
Carlos Zambrano1282131-49
Dana Eveland10112151-39
Vicente Padilla153571-36
Tommy Hunter142356-33
Homer Bailey1088117-29
Felipe Paulino146289-27
Joel Pineiro205681-25
Jair Jurrjens144468-24
Luke Hochevar1395119-24
Hisanori Takahashi1298122-24
Todd Wellemeyer11123146-23
Thomas Hanson252040-20
Jhoulys Chacin135777-20
Vince Mazzaro143452-18
Brett Aarion Cecil205472-18
John Ely1487104-17
Ben Sheets208096-16
Justin Verlander25745816
Jeff Karstens171069016
Kevin Correia2311910316
Ian Kennedy241058817
Joe Saunders251169719
Dan Haren261078720
Colby Lewis24593623
Ryan Dempster25755124
Jake Westbrook241189424
Kris Medlen14906030

Takeaways: I think this tells me players with negative differences have made two starts that were MUCH worse than players with positive differences.

General takeaway: I've been sitting on this data for over a week and wanted to get something out before it got too stale, but as I thought more about it, I realize I should have removed each pitcher's worst two games and compared those numbers against the rest of the pitchers without two starts removed.

Other things I came across during the study which I found interesting:

-Since 2007, no team failed to make the playoffs when they won 90 games (.556 winning percentage).

-King Felix leads MLB by a wide margin with a QS% sit around 89%. Great pitchers sit around 80%.
-One more chart you might find interesting:
Pure ERAStdDevGS
Josh Beckett6.6719.68079
Kyle Lohse6.7918.91334
A.J. Burnett4.6618.58501
Bradley Bergesen5.8918.50852
Joe Saunders4.5418.32362
Dana Eveland6.3418.04706
Mike Pelfrey3.9717.75738
Scott Baker4.8517.5601
Brandon Morrow4.4517.08355
Hisanori Takahashi5.0117.03232
Jake Peavy4.6316.96551
Gavin Floyd3.9716.93827
Matt Garza3.7616.89335
Javier Vazquez4.9016.78431
Aaron Cook5.3416.67158
Justin Masterson5.2316.50587
Joel Pineiro4.1816.32833
Wandy Rodriguez4.0416.24488
Francisco Liriano3.4516.20305
Brett Anderson2.8910.50085
Kevin Correia4.6310.3978
Wade LeBlanc3.4610.38217
Brett Myers3.1110.26853
Justin Verlander3.7710.23054
Andy Pettitte2.8810.13575
David Price2.859.905987
Jeff Karstens4.419.852242
Jason Hammel4.369.531665
Madison Bumgarner3.279.47082
Derek Lowe4.329.463794
Mike Leake3.789.194392
Manny Parra5.869.031285
Stephen Strasburg2.978.902247
Carlos Zambrano5.168.167687
Kris Medlen3.868.125437
Vince Mazzaro3.637.191641
Jair Jurrjens3.927.141428

*ESPN may not have invented this stat, but this is the only place I have seen it.

For anyone interested, I would be happy to provide any of the source data. Just email me - Special thanks to David Pinto ( for providing the Game Scores.

No comments:

Post a Comment