2008 Defense Behind Pitchers By PMR

by Dan Turkenkopf on Dec 1, 2008 8:00 AM EST in Stat-Colored Glasses

Each season, David Pinto releases his Probablistic Measure of Range (PMR) ratings based on play-by-play data from Baseball Info Solutions. PMR measures how many plays above or below expected each team or player made based on batted ball characteristics like velocity and location. Here at Beyond the Box Score, we're translating the play numbers into runs (following a method developed by LA Black Hawk of Waterloo). More information can be found in this post.

Along with the position-by-position numbers, David shares how each pitcher was affected by his defense during the season. He follows a very similar methodology, determining the number of outs that should have been converted, and comparing those to the outs that were actually converted.

Player	Team	IP	In Play	Actual Outs	Predicted Outs	Outs Delta	Runs Delta	ERA Diff
Fausto Carmona	Cle	120.67	405	273	288.06	-15.06	-14.194251	1.06
Adam Eaton	Phi	107	356	236	248.23	-12.23	-11.175789	0.94
Darrell Rasner	NYY	113.33	387	257	269.56	-12.56	-11.837967	0.94
Andy Pettitte	NYY	204	641	420	439.26	-19.26	-18.152807	0.8
Ian Snell	Pit	164.33	522	335	347.6	-12.6	-11.513895	0.63
Livan Hernandez	Min	139.67	525	339	349.78	-10.78	-9.850777	0.63
Carlos Villanueva	Mil	108.33	320	220	228.02	-8.02	-7.3286857	0.61
Garrett Olson	Bal	132.67	451	295	304.47	-9.47	-8.9256013	0.61
Tom Gorzelanny	Pit	105.33	332	225	231.65	-6.65	-6.076778	0.52
Andrew Miller	Fla	107.33	336	216	222.12	-6.12	-5.5924634	0.47
Jonathan Sanchez	SF	158	442	297	306.08	-9.08	-8.297315	0.47
Odalis Perez	Was	159.67	507	337	345.85	-8.85	-8.0871407	0.46
Mark Buehrle	CWS	218.67	699	466	477.66	-11.66	-10.989705	0.45
Cha Seung Baek	SD	111	353	238	243.4	-5.4	-4.9345265	0.4
Luke Hochevar	KC	129	430	291	297.11	-6.11	-5.7587565	0.4
Barry Zito	SF	180	576	393	401.59	-8.59	-7.8495524	0.39
Brett Myers	Phi	190	554	379	388.08	-9.08	-8.297315	0.39
Brandon Backe	Hou	166.67	512	341	348.41	-7.41	-6.771267	0.37
Brian Burres	Bal	129.67	460	309	314.46	-5.46	-5.1461228	0.36
Brian Bannister	KC	182.67	603	408	415.51	-7.51	-7.0782751	0.35
Randy Wolf	SD	119.67	348	237	242.07	-5.07	-4.6329721	0.35
Kevin Millwood	Tex	168.67	569	360	366.45	-6.45	-6.079211	0.32
Mark Hendrickson	Fla	133.67	439	302	307.04	-5.04	-4.6055581	0.31
Nick Blackburn	Min	193.33	658	445	451.38	-6.38	-5.8300517	0.27
Pedro Martinez	NYM	109	337	225	228.38	-3.38	-3.0886481	0.26
Jorge de la Rosa	Col	130	361	240	243.99	-3.99	-3.6460668	0.25
A.J. Burnett	Tor	221.33	613	405	411.37	-6.37	-6.0038099	0.24
John Maine	NYM	140	399	286	290.11	-4.11	-3.755723	0.24
Kyle Davies	KC	113	361	248	251.2	-3.2	-3.0160427	0.24
Mike Pelfrey	NYM	200.67	652	446	450.98	-4.98	-4.55073	0.2
Edinson Volquez	Cin	196	511	350	354.63	-4.63	-4.2308996	0.19
James Shields	TB	215	641	448	452.73	-4.73	-4.4580881	0.19
Justin Verlander	Det	201	598	415	419.47	-4.47	-4.2130346	0.19
Randy Johnson	Ari	184	531	359	363.22	-4.22	-3.8562411	0.19
Ricky Nolasco	Fla	212.33	606	432	435.96	-3.96	-3.6186528	0.15
Carlos Silva	Sea	153.33	564	365	367.36	-2.36	-2.2243315	0.13
Clayton Kershaw	LAD	107.67	306	204	205.65	-1.65	-1.507772	0.13
Jo-Jo Reyes	Atl	113	361	241	242.79	-1.79	-1.6357042	0.13
Nate Robertson	Det	168.67	563	365	367.59	-2.59	-2.4411095	0.13
Kenny Rogers	Det	173.67	598	400	402.42	-2.42	-2.2808823	0.12
Zach Duke	Pit	185	669	445	447.62	-2.62	-2.3941592	0.12
Ubaldo Jimenez	Col	198.67	572	395	397.49	-2.49	-2.275365	0.1
Brandon Webb	Ari	226.67	671	458	460.45	-2.45	-2.238813	0.09
Ervin Santana	LAA	219	605	422	424.34	-2.34	-2.2054812	0.09
Jason Bergmann	Was	139.67	445	310	311.45	-1.45	-1.3250118	0.09
Joe Blanton	Oak	127	440	303	304.31	-1.31	-1.2346925	0.09
Daniel Cabrera	Bal	180	594	409	410.64	-1.64	-1.5457219	0.08
Bronson Arroyo	Cin	200	605	408	409.76	-1.76	-1.6082901	0.07
Chris Sampson	Hou	117.33	383	267	267.83	-0.83	-0.758455	0.06
Andy Sonnanstine	TB	193.33	632	432	433.24	-1.24	-1.1687165	0.05
Tim Redding	Was	182	572	397	397.98	-0.98	-0.8955252	0.04
Wandy Rodriguez	Hou	137.33	393	266	266.71	-0.71	-0.6487989	0.04
Ted Lilly	ChC	204.67	574	412	412.67	-0.67	-0.6122468	0.03
Zack Greinke	KC	202.33	587	399	399.62	-0.62	-0.5843583	0.03
Cliff Lee	Cle	223.33	670	462	462.5	-0.5	-0.4712567	0.02
Chad Billingsley	LAD	200.67	556	370	370.36	-0.36	-0.3289684	0.01
Jarrod Washburn	Sea	153.67	512	350	350.13	-0.13	-0.1225267	0.01
Mike Mussina	NYY	200.33	613	409	409.01	-0.01	-0.0094251	0
Gil Meche	KC	210.33	611	420	419.79	0.21	0.1979278	-0.01
Manny Parra	Mil	166	499	322	321.73	0.27	0.24672633	-0.01
Aaron Laffey	Cle	93.67	316	217	216.67	0.33	0.3110294	-0.03
Boof Bonser	Min	118.33	382	249	248.24	0.76	0.71631013	-0.05
Greg Maddux	SD	153.33	511	360	359.16	0.84	0.76759302	-0.05
Jon Lester	Bos	210.33	632	438	436.2	1.8	1.696524	-0.07
Brad Penny	LAD	94.67	311	212	211.06	0.94	0.85897314	-0.08
Johnny Cueto	Cin	174	500	344	341.79	2.21	2.01950067	-0.1
Tim Lincecum	SF	227	562	385	382.12	2.88	2.63174748	-0.1
Aaron Harang	Cin	184.33	552	379	376.37	2.63	2.40329718	-0.12
Joel Pineiro	StL	148.67	505	342	339.9	2.1	1.91898254	-0.12
Javier Vazquez	CWS	208.33	598	405	401.92	3.08	2.90294107	-0.13
Jose Contreras	CWS	121	402	280	278.09	1.91	1.80020047	-0.13
Gavin Floyd	CWS	206.33	625	450	446.51	3.49	3.28937153	-0.14
Jason Marquis	ChC	167	554	390	387.25	2.75	2.51295333	-0.14
Matt Cain	SF	217.67	630	436	431.93	4.07	3.71917092	-0.15
Jair Jurrjens	Atl	188.33	589	401	397.07	3.93	3.59123876	-0.17
Kevin Correia	SF	110	382	248	245.75	2.25	2.05605272	-0.17
Miguel Batista	Sea	115	379	257	254.22	2.78	2.62018707	-0.21
Braden Looper	StL	199	653	453	447.51	5.49	5.01676864	-0.23
Carlos Zambrano	ChC	188.67	570	404	398.66	5.34	4.87969846	-0.23
Dan Haren	Ari	216	610	421	414.6	6.4	5.84832774	-0.24
Jamie Moyer	Phi	196.33	625	437	431.36	5.64	5.15383882	-0.24
Jered Weaver	LAA	176.67	513	355	350.05	4.95	4.665441	-0.24
Todd Wellemeyer	StL	191.67	579	419	413.37	5.63	5.14470081	-0.24
Dana Eveland	Oak	168	519	351	346.09	4.91	4.62774047	-0.25
Jon Garland	LAA	196.67	684	462	456.1	5.9	5.56082867	-0.25
Doug Davis	Ari	146	457	303	298.15	4.85	4.43193587	-0.27
Hiroki Kuroda	LAD	183.33	598	418	412.03	5.97	5.45539322	-0.27
Jeff Suppan	Mil	177.67	589	407	401.18	5.82	5.31832304	-0.27
Kyle Lohse	StL	200	650	453	446.33	6.67	6.09505407	-0.27
Aaron Cook	Col	211.33	725	489	481.78	7.22	6.59764474	-0.28
Edwin Jackson	TB	183.33	582	401	394.74	6.26	5.90013347	-0.29
Felix Hernandez	Sea	200.67	577	391	384.12	6.88	6.48449173	-0.29
Oliver Perez	NYM	194	527	380	373.11	6.89	6.29609034	-0.29
Johan Santana	NYM	234.33	668	476	467.22	8.78	8.02317462	-0.31
Derek Lowe	LAD	211	644	453	444.85	8.15	7.44747986	-0.32
John Danks	CWS	195	569	396	388.39	7.61	7.17252647	-0.33
Micah Owings	Ari	104.67	312	220	215.84	4.16	3.80141303	-0.33
Roy Halladay	Tor	246	713	501	491.48	9.52	8.97272693	-0.33
Scott Baker	Min	172.33	497	354	347.21	6.79	6.39966553	-0.33
Scott Olsen	Fla	201.67	640	463	454.69	8.31	7.59368806	-0.34
Jeff Francis	Col	143.67	469	321	314.83	6.17	5.63815347	-0.35
Matt Garza	TB	184.67	560	399	391.43	7.57	7.13482593	-0.35
Paul Maholm	Pit	206.33	621	437	428.22	8.78	8.02317462	-0.35
Scott Feldman	Tex	151.33	488	344	337.73	6.27	5.9095586	-0.35
Cole Hamels	Phi	227.33	635	464	453.13	10.87	9.93301915	-0.39
David Bush	Mil	185	567	422	413.33	8.67	7.92265649	-0.39
Jake Peavy	SD	173.67	459	331	322.53	8.47	7.73989625	-0.4
Jeremy Guthrie	Bal	190.67	587	428	417.45	10.55	9.94351567	-0.47
Josh Beckett	Bos	174.33	492	333	322.69	10.31	9.71731247	-0.5
Vicente Padilla	Tex	171	524	356	345.93	10.07	9.49110927	-0.5
Kevin Slowey	Min	160.33	480	340	330.37	9.63	9.0764034	-0.51
Ben Sheets	Mil	198.33	589	417	404.45	12.55	11.4682052	-0.52
John Lannan	Was	182	560	401	389.21	11.79	10.7737163	-0.53
Jorge Campillo	Atl	158.67	490	347	336.39	10.61	9.69543084	-0.55
Zach Miner	Det	118	385	276	268.08	7.92	7.4647056	-0.57
Adam Wainwright	StL	132	404	289	279.63	9.37	8.56231734	-0.58
Dustin McGowan	Tor	111.33	337	228	220.26	7.74	7.2950532	-0.59
Ryan Dempster	ChC	206.67	572	406	391.16	14.84	13.56081	-0.59
John Lackey	LAA	163.33	469	329	316.59	12.41	11.6965905	-0.64
Shaun Marcum	Tor	151.33	427	317	305.52	11.48	10.8200531	-0.64
Joe Saunders	LAA	198	623	445	429.72	15.28	14.4016037	-0.65
CC Sabathia	Mil	130.67	353	246	235.49	10.51	9.60405072	-0.66
Brian Moehler	Hou	150	509	356	343.33	12.67	11.5778613	-0.69
R.A. Dickey	Sea	112.33	374	263	253.88	9.12	8.5957216	-0.69
Scott Kazmir	TB	152.33	378	277	264.57	12.43	11.7154407	-0.69
Greg Smith	Oak	190.33	578	423	406.63	16.37	15.4289433	-0.73
Glen Perkins	Min	151	520	357	343.77	13.23	12.4694514	-0.74
Armando Galarraga	Det	178.67	525	391	375.29	15.71	14.8068845	-0.75
CC Sabathia	Cle	122.33	334	228	217.23	10.77	10.1508686	-0.75
Paul Byrd	Cle	131	446	319	307.48	11.52	10.8577536	-0.75
Kyle Kendrick	Phi	155.67	560	380	365.69	14.31	13.0764953	-0.76
Roy Oswalt	Hou	208.67	617	436	416.75	19.25	17.5906733	-0.76
Tim Hudson	Atl	142	435	317	302.69	14.31	13.0764953	-0.83
Jeremy Sowers	Cle	121	409	279	266.73	12.27	11.5646386	-0.86
Justin Duchscherer	Oak	141.67	409	308	292.97	15.03	14.1659754	-0.9
Ryan Rowland-Smith	Sea	118.33	366	262	249.06	12.94	12.1961225	-0.93
Daisuke Matsuzaka	Bos	167.67	449	327	306.07	20.93	19.7268041	-1.06
Tim Wakefield	Bos	181	539	405	382.33	22.67	21.3667773	-1.06
Jesse Litsch	Tor	176	569	407	382.14	24.86	23.4308815	-1.2
Chien-Ming Wang	NYY	95	306	215	200.92	14.08	13.2705877	-1.26

There's a lot of great information here that's just begging for additional analysis. For example, this data might show which pitchers give up the easiest balls in play to field. Or we might look at why a defense performs so much differently for a given pitcher as opposed to his teammates (Chien-Ming Wang vs. Andy Pettite and Darrell Rasner for example). I'm not going to cover those today, and just stick with the same high-level thoughts as the rest of the series.

One additional note; you'll see I went away from the rate stats per 4000 BIP and looking at the runs above average, and added a column describing the effect on ERA (more accurately, RA). I can make those other numbers available if anyone wants, but I find that scaling results for pitchers to the common ERA scale makes the results more easily accessible. A positive number means that the pitchers actual ERA was higher than expected, while a negative indicates the defense helped the pitcher do better than expected.

Fausto Carmona was definitely victimized by his defense this season. He would have had been better than average has his defense performed as expected. Instead he ended up with a 5.44 ERA and an 82 ERA+. I'd expect some bounceback next year from Fausto. On the other end of the scale, the Yankees defense is much better when Chien-Ming Wang is pitching than anyone else. This pattern has repeated itself for a few years now. Boston's defense really helped out Wakefield and Dice-K. That probably helps to explain why Dice-K was able to do so much better than his peripherals would suggest.

0 recs | Comment 12 comments |

Read Related

Story-email Email Printer Print

Comments

This is awesome stuff.

If I remember correctly, Dave Pinto will also publish similar data, but showing expected BABIP based on the BIP allowed by each pitcher, right? I would assume he also has the ability to compute the average run-value of each six-variable vector if the ball isn’t turned into an out. With that, you’ve got a super-duper version of tRA.

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Dec 1, 2008 2:52 PM EST 0 recs

I'm not sure I follow

His original post (which I take bits and pieces of) includes the predicted DER behind each pitcher (which is 1-BABIP, of course). So that might be what you’re looking for.

The suggestion I made, which I’ll probably begin working on, is to look at what the predicted BABIP should be for a pitcher given his actual spread of grounders, line drives and fly balls and compare that to David’s play-by-play based predicted BABIP. That might allow us to determine whether a pitcher gives up easier to field balls in play.

He did do a detailed breakdown for Wang last season (at the end of this post) but didn’t convert it to runs for each vector. Doing that would be more involved than my current approach, which assumes average distribution of balls in play across all vectors. We’d need to calculate the run value for each vector and apply it – which would require a fair amount of detailed data (which I’m pretty sure David could do if he so chose).

by Dan Turkenkopf on Dec 1, 2008 5:56 PM EST up 0 recs

Yes, pitcher-induced expected-BABIP is the basic piece of what I was suggesting. Thanks.

And yes, your second and third paragraphs are the methods I would advocate. Given that David’s already programmed PMR, I wouldn’t think run values would be that big of a next step.

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Dec 1, 2008 6:19 PM EST up 0 recs

I really didn't expect to see Wakefield near the bottom.

by R.J. Anderson on Dec 1, 2008 6:56 PM EST 0 recs

You expected the defense behind him to be worse?

Is this because of the knuckleballer effect to DIPS? (I can never remember which way that goes)

by Dan Turkenkopf on Dec 1, 2008 7:09 PM EST up 0 recs

I'm not sure if I expected them to be worse.

I just didn’t expect them to be THAT good.

by R.J. Anderson on Dec 1, 2008 7:14 PM EST up 0 recs

Gotcha

David seems to think that positioning plays a major role in this.

The Yankees apparently have a very good idea where Wang’s ground balls are going to go and convert them much more often than you’d expect.

by Dan Turkenkopf on Dec 1, 2008 7:22 PM EST 0 recs

One thing to look at with pitchers is groupings of batted balls.

Pitchers who give up wide, seemingly random spreads of batted balls are less valuable than pitchers who give up clustered batted balls (because you can position fielders).

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Dec 1, 2008 7:27 PM EST up 0 recs

Wang doesn't have a heavy slant to either direction.

From FirstInning, his BBchart:

by R.J. Anderson on Dec 1, 2008 7:29 PM EST up 0 recs

Spread within those ranges is probably the key

The effect should show up in smaller slices.

The findings from last year had the left side of the infield performing surprisingly well on balls in the holes (both grounders and line drives), which would be the case if the fielders played a step or two away from their normal positions. (Note: most of this analysis is lifted from David’s post).

by Dan Turkenkopf on Dec 1, 2008 7:35 PM EST up 0 recs

This. You need really fine-grained info.

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Dec 1, 2008 7:45 PM EST up 0 recs

I’m not saying you are wrong but I don’t really trust those charts much. FI has problems keeping their data both up to date and completely accurate (partically because they aren’t up to date) so I’d be careful coming to conclusions based off of that chart.

by jfish26101 on Dec 2, 2008 10:00 AM EST up 0 recs