Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Dialed In > Discussion
Dialed In
— 

Monday, October 30, 2006

2006 National League Gold Gloves - As I see it

Defensive data has been and is being refined pretty well these days.  With more and more play-by-play data making it to the mainstream, all of us are stretching the boundaries of what we require from black-box analysts.  With the exception of some park factors, we are discovering that Zone Rating provides a pretty good picture of defense.  Taking the zone rating and accounting for league averages, based on tens of thousands of defensive innings played, we can closely assess the number of runs saved by a defensive players as compared to his peers.

To be sure, even this data could be refined to account for parks better - Fenway’s Green Monster is a tremendous issue - and handedness of batters - NOT handedness of pitchers - to tune the picture a bit better, but the data you will read will be very close to any refined data.  Very close.  The basic methodology for this work is here.  You can also read more on where we are headed with Park Factors by reading Rally’s latest work.

I have tweaked this for chances per inning from the original data, so the chances assumed here may be slightly higher/lower, but if you did the same work from the referenced article, you’d find your results would be within a run or two of what I post.  And really, the most important thing I do here is provide you with the tools to evaluate defense on your own, without me doing the math.  Please note, after this article, I will post some others’ work that even refines what I have done, with a comparison to what I have done.  It should be exciting for you - it is for me.  Most importantly, it broadens the network of individuals accurately creating the defensive evaluations, as well as allows for everyday updates.  Yes, I said *every day*.

Now on with the show.  Here are the leaders and trailers at every position for the National League, with some commentary where necessary.  The American League results are here.  In general I draw the Gold Glove qualification line at a significant number of innings - usually around 650.  It would be unusual for someone playing only 650 innings to lead the league in anything, but I’m willing to give it a look. 

Catcher

First	LAST		TEAM	GP	INN	RSpt	RS/150
Yadier	Molina		StL	127	1037.3	8	10
Henry	Blanco		ChC	69	526.0	5	13
Miguel	Olivo		Fla	124	971.3	5	7
Ronny	Paulino		Pit	124	1047.0	4	6
Chris	Snyder		Ari	60	495.0	4	11
Yorvit	Torrealba	Col	63	530.0	4	10
David	Ross		Cin	75	621.7	3	7
Damian	Miller		Mil	98	840.0	3	5
Mike	Lieberthal	Phi	60	484.0	3	8
Russell	Martin		LA	117	1015.0	3	3
Jason	LaRue		Cin	63	512.3	2	6
Brian	Schneider	Was	123	990.3	2	3
Johnny	Estrada		Ari	108	925.7	2	3
Brad	Ausmus		Hou	138	1125.7	0	0
Eliezer	Alfonzo		SF	84	700.3	0	-1
Josh	Bard		SD	71	495.7	-2	-5
Brian	McCann		Atl	124	1016.3	-3	-4
Paul	LoDuca		NYM	118	1027.0	-4	-5
Michael	Barrett		ChC	102	852.0	-7	-11
Mike	Piazza		SD	99	718.0	-13	-24

Yadier Molina was a good defensive catcher, and amazingly, the good hitting catchers - McCann, LoDuca, Barrett and Piazza are all bringing up the rear.  Maybe the old cliche rings true.  Molina led in CS%, and will likely win the conventional GG award. 

First Base

First	LastName	TEAM	GP	INN	RSpt	RS/150
Scott	Hatteberg	Cin	131	1089.7	9	11
Adrian	Gonzalez	SD	155	1341.0	5	5
Nomar	Garciprra	LA	118	1017.0	3	5
Carlos	Delgado		NYM	141	1246.3	3	3
Adam	LaRoche		Atl	142	1153.3	2	3
Albert	Pujols		StL	143	1244.3	0	1
Conor	Jackson		Ari	129	1079.7	0	0
Mike	Jacobs		Fla	124	972.0	-1	-1
Nick	Johnson		Was	147	1252.3	-1	-1
Todd	Helton		Col	145	1272.3	-2	-2
Ryan	Howard		Phi	159	1412.0	-3	-3
Lance	Berkman		Hou	112	923.0	-4	-5
Prince	Fielder		Mil	152	1319.3	-9	-9

Scott Hatteberg must be able to “really pick it”.  There’s very little variation here.  Nine runs seperate #2 from #11.  That says to me, skill-wise, that the difference could easily be in the chances.  I don’t know who will win the real GG, but I would give it to Hatteberg.  Yes, I know PUjols is a really good fielder - sure, but this season, he just didn’t make a bunch of plays above and beyond. 

Second Base

First	LastName	TEAM	GP	INN	RSpt	RS/150
Jose	Valentin	NYM	94	782.3	12	21
Jamey	Carroll		Col	109	895.7	11	17
Craig	Biggio		Hou	129	1062.0	5	6
Chase	Utley		Phi	156	1367.3	3	3
Brandon	Phillips	Cin	142	1216.3	2	2
Josh	Barfield	SD	147	1259.0	1	1
Aaron	Miles		StL	88	650.7	1	2
Ray	Durham		SF	133	1139.7	-2	-3
Orlando	Hudson		Ari	157	1349.0	-4	-4
Dan	Uggla		Fla	151	1304.3	-4	-4
Jose	Vidro		Was	107	902.7	-5	-8
Jeff	Kent		LA	108	888.7	-8	-12
Rickie	Weeks		Mil	92	794.0	-8	-14
Marcus	Giles		Atl	134	1150.7	-9	-11
Jose	Castillo	Pit	145	1235.0	-18	-20

I really hope Harvey shows up to tell us how much Weeks improved over the last couple of months.  He was at -8 RSpt at teh ASB.  He held his ground, so he may have turned a corner for next year.  As for the GG, I am certainly wary of awarding it to someone with less than 800 innings, so I would probably go with Jamey Carroll - plus he has the Colorado BIP issue to deal with.

Third Base

First	LastName	TEAM	GP	INN	RSpt	RS/150
Pedro	Feliz		SF	159	1372.3	11	11
Morgan	Ensberg		Hou	117	975.0	9	12
Scott	Rolen		StL	142	1216.7	5	6
Freddy	Sanchez		Pit	99	822.7	5	8
Ryan	Zimmrmn		Was	157	1368.3	1	1
David	Bell		Mil/Phi	143	1216.0	0	0
Chad	Tracy		Ari	147	1278.0	0	0
Aramis	Ramirez		ChC	156	1353.0	-4	-4
Chipper	Jones		Atl	105	888.3	-4	-6
Garrett	Atkins		Col	157	1381.3	-5	-5
Edwin	Encarnc'n	Cin	111	931.3	-10	-14
David	Wright		NYM	153	1365.3	-10	-10
Miguel	Cabrera		Fla	157	1334.0	-12	-12

Pero Feliz played as many innings as anyone else, and played them better than anyone else.  He should win the GG.  Given the history of the award, Rolen will win.  Not on this list, but really good in 600 innings was Corey Koskie for Milwaukee.  He was a +9 RSpt, and a +19 for RS/150.

Shortstop

First	LastName	TEAM	GP	INN	RSpt	RS/150
Adam	Everett		Hou	149	1292.3	27	28
Omar	Vizquel		SF	152	1281.3	11	12
Jose	Reyes		NYM	149	1320.3	10	10
David	Eckstein	StL	120	1029.0	7	9
Khalil	Greene		SD	113	998.7	5	7
Craig	Counsell	Ari	88	737.7	4	8
Clint	Barmes		Col	125	1073.7	4	5
Ronny	Cedeno		ChC	134	1130.7	2	2
Jack	Wilson		Pit	131	1130.0	1	2
Bill	Hall		Mil	127	1090.3	-1	-1
Rafael	Furcal		LA	156	1371.0	-4	-4
Royce	Clayton		Cin/Was	129	1055.0	-5	-7
Edgar	Renteria	Atl	146	1265.3	-6	-6
Jimmy	Rollins		Phi	157	1378.0	-6	-5
Felipe	Lopez		Cin/Was	155	1337.0	-17	-17
Hanley	Ramirez		Fla	154	1323.3	-17	-18

If Ozzie Smith was as good as Adam Everett, he was incredible.  Everett is on the verge of saving hte most runs on defense over the last 20 years.  He’s truly incredible at outpacing his peers.  Hanley Ramirez can be forgiven - hopefully he’ll learn and develop.  There’s Eckstein, ugly arm and all, performing well.  He did look very good in the NLCS, making several stops on balls I was sure would be hits.

Left Field

First	LastName    	TEAM	GP	INN	RSpt	RS/150
Dave	Roberts		SD	116	970.0	14	19
Ryan	Langrhns	Atl	104	706.0	11	20
Matt	Murton		ChC	133	1049.0	6	8
Matt	Diaz		Atl	95	587.3	6	14
Alfonso	Soriano		Was	158	1374.7	5	5
Matt	Holliday	Col	153	1334.3	3	3
Andre	Ethier		LA	109	896.7	2	4
Luis	Gonzalez	Ari	150	1315.0	2	2
Cliff	Floyd		NYM	92	768.3	1	1
Barry	Bonds		SF	116	875.0	0	0
Pat	Burrell		Phi	126	988.7	0	0
Carlos	Lee		Mil	98	835.3	-3	-5
Jason	Bay		Pit	157	1373.0	-6	-5
Adam	Dunn		Cin	156	1321.0	-12	-12
Josh	Willinghm	Fla	132	1070.7	-15	-19
Preston	Wilson		Hou/StL	102	873.0	-20	-31

Dave Roberts?  I didn’t see that coming.  He is a CF playing LF, so I can buy it.  We’re expecting a park factor for Wilson in Houston, and tehre is some effect in Florida, but Willingham is a catcher, so he probably isn’t very good.  Then Adam Dunn… 

Center Field

First	LastName	TEAM	GP	INN	RSpt	RS/150
Juan	Pierre		ChC	162	1426.0	16	15
Carlos	Beltran		NYM	136	1184.0	9	10
Eric	Byrnes		Ari	123	1051.0	8	11
Mike	Cameron		SD	141	1244.0	8	9
Chris	Duffy		Pit	77	672.7	6	12
Jim	Edmonds		StL	99	792.3	6	9
Willy	Taveras		Hou	138	1117.7	5	6
Steve	Finley		SF	130	973.3	3	5
Cory	Sullivan	Col	114	841.0	-1	-2
Kenny	Lofton		LA	120	961.0	-2	-3
Brady	Clark		Mil	114	911.7	-3	-4
Aaron	Rowand		Phi	107	901.7	-3	-5
Andruw	Jones		Atl	153	1317.3	-9	-9
Ken	GriffeyJr	Cin	100	870.3	-11	-18

Juan Pierre?  That’s a good-sized lead as well.  Beltran, with this defense, is the best selection for MVP.  Griffey hasn’t played well in CF for a few years, and while not as bad as last year, he still isn’t good.  I’d play him in LF.  And if BIS still has Andruw as the top CF, they should re-consider their algorithms.

Right Field

First	LastName	TEAM	GP	INN	RSpt	RS/150
Brian	Giles		SD	158	1399.7	12	11
Juan	Encrncion	StL	125	983.3	9	13
J.D.	Drew		LA	135	1118.0	5	6
Austin	Kearns		Cin/Was	144	1229.7	5	6
Jeremy	Hermida		Fla	85	684.7	4	8
Jacque	Jones		ChC	143	1205.0	4	4
Randy	Winn		SF	89	653.7	2	4
Bobby	Abreu		Phi	97	846.0	1	1
Xavier	Nady		NYM/Pit	99	855.7	0	1
Jeff	Francoeur	Atl	162	1422.7	-1	-1
Geoff	Jenkins		Mil	133	1101.0	-2	-2
Shawn	Green		NYM/Ari	131	1121.3	-5	-6
Moises	Alou		SF	81	647.7	-6	-12
Jason	Lane		Hou	89	679.3	-8	-16
Brad	Hawpe		Col	145	1198.7	-9	-10
Jeromy	Burnitz		Pit	84	643.0	-12	-25

Brian Giles was a good CF a few years back.  With an OF of Cameron, Roberts and Giles, I am not surprised the Padres OF defense is this good.  Jeromy Burnitz should retire.

I have to talk about Endy Chavez.  His defense this year was incredible, and most people got to see it in Game 7 of the NLCS.  Fortunately you are going to see that play a few million times over the rest of your life.  He played incredibly all season, and in all three outfield positions.  He totaled 264.3 innings in CF at +5 RSpt, 239.3 innings at +6 in LF, and +5 RSpt in 312.6 innings in right field.  That’s +16 in 816.3 innings.  His RS/150 is upwards of 25 runs.  He deserves the Gold Glove over Dave Roberts. 


My Gold Gloves?  Molina, Hatteberg, Carroll, Everett, Feliz, Chavez, Pierre, Giles.  Any complaints?

Chris Dial Posted: October 30, 2006 at 04:27 AM | 154 comment(s) Login to Bookmark
  Related News:

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

Page 2 of 2 pages  < 1 2
   101. Mike Emeigh Posted: October 30, 2006 at 09:38 PM (#2228243)
So we're suggesting the system is wrong for three plays a year for one 3B?


What I'm suggesting is that the system *may* be wrong by some *unknown* number of plays, based on how closely the optimal positioning for average fielders given the distribution of balls in play against a team matches the optimal positioning for average fielders based on the average distribution of balls in play.

-- MWE
   102. AROM Posted: October 30, 2006 at 09:47 PM (#2228247)
If its three plays per year, then it won't significantly change ratings whether you debit the shortstop on these plays or not. The question is, how many plays does a 3B cut off the shortstop normally, and how many plays would he cut off the shortstop if he plays far off the line?

It probably doesn't make much difference in the infield. Its a bigger deal in the outfield, where many flyballs could be caught be either fielder. If you penalize an outfielder for getting called off on a ball, you wind up with weird ratings.
   103. Fridas Boss Posted: October 30, 2006 at 10:02 PM (#2228256)
If you penalize an outfielder for getting called off on a ball, you wind up with weird ratings.

But your ratings will also be weird if the REASON the fielder didn't catch the ball is because he couldn't get there. That speaks a lot more about a fielders 'ability' than being called off of a ball you can get, yet the data will score you the same. Same result, NOT same measure of fielding ability.
   104. AROM Posted: October 30, 2006 at 10:30 PM (#2228274)
Play A: High fly ball, 2 outfielder can get to it, one calls the other off and makes a catch.

Play B: Difficult fly ball, Fielder Y cannot get there even though it in his zone, Fielder X ranges out of his zone and makes the play.

Which is more common? I don't have the play by play microdata or the time to wade through and identify such plays, but my experience watching games tells me that Play A is far more common.
   105. AROM Posted: October 30, 2006 at 10:37 PM (#2228277)
That speaks a lot more about a fielders 'ability' than being called off of a ball you can get, yet the data will score you the same. Same result, NOT same measure of fielding ability.

I see what you are saying. Lets say Barry Bonds really plays left field in his Lazy-Boy. He catches 5 balls per year hit right at him. Gary Pettis and Willie Mays Hayes in their prime cover the other two spots so well they prevent anything from falling in Bonds' zone. Bonds' ZR is a perfect 5/5, but he's a terrible outfielder.

In reality though, a terrible OF is going to have tons of hits drop in around him, and there's nothing the other OF's can do about it. If they catch a ball in his zone, it is most likely a flyball that hangs up forever and either player could catch.
   106. Mike Green Posted: October 30, 2006 at 11:04 PM (#2228301)
Positioning, and ball in play distribution, is a devilishly difficult issue to address in evaluating particularly ground ball defence. There are a number of ways to do it, but I do agree that one should not take account of balls fielded in the hole by the third baseman and treat those as opportunities for the shortstop.
   107. John Walsh Posted: October 30, 2006 at 11:34 PM (#2228320)
These discussions on defense are great.

Can somebody remind me how the zones of responsibility are determined? I seem to recall that if the
average SS, say, fields 50% of the balls in a particular zone, then that zone is considered
his responsibility. Is that correct?
   108. Scout2u Posted: October 30, 2006 at 11:58 PM (#2228344)
Interesting work Chris. There is a drawback to any metric using ZR as it's baseline, of course, but it is a very good starting point for intelligent discussion.

I did see you are using STATS, Inc's grid (22 vectors). The obvious downfall to that is, the larger the park, the larger the vector (if they are still using the same system they used a few years ago. Do you know?). BIS data is not reliable I often hear, but I don't know to be true.

My last concern would be, and forgive me if you have answered this somewhere else, are you not measuring player-vs-player instead of player-in-a-specific-park-opportunity?

As a side note, perhaps you could include a 3-year historical baseline reference point (league average) and show the player's 3-year value against it as well?

Until we come up with an absolute positioning system on every ball in play, a square footage vector breakdown for every park in baseball, and the ability to measure a host of variables like height and speed of the ball in play, up the middle or spread eagle defensive positioning, and the like, ZR is about as uncomplicated as it gets. I still think Wayne Britton's idea of having the threads coated with a trackable substance makes sense (joking here).

Good work Chris.
   109. Robert S. Posted: October 31, 2006 at 12:07 AM (#2228353)
Robert, you are so full of it. Why not throw in Brandon Webb while you're at it?
True, I'd want more for the pair, but the heart's in the right place. Unless, of course, you want to argue that Arizona is better off hanging on to them . . .
   110. Chris Dial Posted: October 31, 2006 at 01:59 AM (#2228451)
Can somebody remind me how the zones of responsibility are determined? I seem to recall that if the
average SS, say, fields 50% of the balls in a particular zone, then that zone is considered
his responsibility. Is that correct?


John,
yes, that is correct. I have written a synopsis here.

Always be sure to read the discussion. I have done a ton of defensive work, but I always miss something, or someone else always has watched and can offer new viewpoints.
   111. Chris Dial Posted: October 31, 2006 at 02:11 AM (#2228460)
Scout,
thanks for your kind words.

My last concern would be, and forgive me if you have answered this somewhere else, are you not measuring player-vs-player instead of player-in-a-specific-park-opportunity?

I'm not sure I can read this clearly. We (and I say we, because I'm not the only one working through this) do need to figure park factors. Someone like Manny Ramirez is clearly portrayed as a worse fielder than he is. I simply haven't had enough granular data to work on that. Joe Arthur has made some interesting obs on Fenway PFs, and we're starting to get the right ideas. Not working with the raw data makes it difficult.

So, the player is half in a specific park. But it is player vs player, and some players have a disadvantage.

As a side note, perhaps you could include a 3-year historical baseline reference point (league average) and show the player's 3-year value against it as well?

this would be a smart thing to do. I am preparing a comparison of my quicker and dirtier method to some better techniques created by SG and Kyle C. We have enough data now that we shold be able to do this.

Have you reviewed the work on 20-year data also in this blog? there's a bunch of data in there, and more to come now that the Mets season is over.

And yes, AFAIK, STATS is using the same system (but is using smaller (5ft) zones). I have also heard that the BIS data is missing some pieces, and when asked I advised the team use MLB's data due to that. there was some question about the professional aspect, but I don't think there is any real difference between scorers. Geeks take this stuff really seriously.
   112. John Walsh Posted: October 31, 2006 at 01:01 PM (#2228619)

The question is, how many plays does a 3B cut off the shortstop normally, and how many plays would he cut off the shortstop if he plays far off the line?


I have looked at some retrosheet data to try to answer the first question. Here's what I did:

I looked at the retrosheet data from 1991-1992: this data has complete batted ball type and hit location data (the data for these years actually comes from STATS). Of course, retrosheet employs the Project Scoresheet zone grid, which has bigger zones. However, I think it can be used to answer this specific question. The 3B zones on the STATS grid are CDEF, then there is a single zone that is a gap (zone G), then the SS is responsible for the next 5 zones (HIJKL).

Now the PS zones of 6S+6+6D correspond to the STATs zones HIJK (approximately). In other words, the right edges of the STATS and PS zones for SS are same. So, if I look at how often a 3B fields a ball in 6+6S+6D on a PS grid, this will be the same as the number of times a 3B fields a ball in the STATs SS zones (neglecting the 5th SS zone (L) way over near 2nd base).

What I find is a total of 15441 ground balls hit in 6+6S+6D (or HIJK). Of those, the SS made outs on 12970, while the 3B made outs on 365 of them. Or, in other words, the 3B is "stealing" balls from the SS about 3% of the time. Doesn't sound like a very big deal to me.

The 3B actually cuts off the slow choppers pretty often: in PS zone 6S, the 3B fields 27% of the outs made, while in zone 6, the 3B fields less than 2% of the outs.

Of course, these numbers represent the average. A rangy 3B playing well of the line could increase the rate of "incursion" into the SS zone. A more careful analysis of the retrosheet data might shed some light on this, but it wouldn't be easy.
   113. Foghorn Leghorn Posted: October 31, 2006 at 01:14 PM (#2228625)
John, very nice work. However, we also have to recognize that over a two year period those 365 balls, a good portion will be plays made under "shifted" conditions. So the 3B isn't "ranging" to his left, but standing in the hole with the SS playing up the middle. This is certainly more prevelant these days, but happened from time to time back then as well.
   114. John Walsh Posted: October 31, 2006 at 02:48 PM (#2228658)
Thanks, Chris.

Very good point about shifts. I can try to address the extreme shift issue by only considering balls hit by right-handed batters. The assumption is if there is a rhb at the plate, the 3B will not be playing in the SS zone.

Of course, for the vast majority of (the excluded) left-handed batters, there is no extreme shift, but there may be some subtle shifting away from the 3B line, which is probably not generally the case when a rhb is up.

Anyway, if I remove lhb, I find (still looking in the 6+6S+6D/HIJK zone) 11818 total ground balls. The SS fielded 10181 and the 3B fielded 229 of them. So the fraction of outs made by the 3B is now 2.2% (compare to 2.7% when all batters are included).

The 3B out fraction is down to 21% on the slow balls (fielded in zone 6S) and 1.4% on the balls fielded in zone 6.


In any case, even with the right-handers there are still a bit more than 2% of the outs in the SS zone made by the 3B. I consider 2% a small number. Your metric (or any other) will have other sources of uncertainty that are far larger than 2%.
   115. Mike Emeigh Posted: October 31, 2006 at 03:03 PM (#2228670)
So, if I look at how often a 3B fields a ball in 6+6S+6D on a PS grid, this will be the same as the number of times a 3B fields a ball in the STATs SS zones (neglecting the 5th SS zone (L) way over near 2nd base).


I don't normally ask someone to do a piece of work that I can do myself, but I don't have time to do it right now. So let me ask:

Are there differences in the distribution of these by team? I realize that we are talking small sample sizes here, but if there are some teams that have more of them than others that might provide a clue into the extent to which there are positioning/distributional differences.

-- MWE
   116. Jake Luft Posted: October 31, 2006 at 03:32 PM (#2228709)
I watched nearly every Dontrelle start this season and, all things considered, he pitched extremely well this season. As the numbers show, his infield defense was brutal behind him. Dan Uggla got better as the season went along, but he couldn't turn a DP to save his life in the first half. I have high hopes for Hanley Ramirez. He has tremendous range, a big-time arm and makes the flashy plays, but he vapor locks sometimes on the easy ones. Chalk it up to rookie jitters. Mike Jacobs is awful; not sure if he's as bad as Delgado though. Miguel Cabrera played better in the second half, but still not good overall. Miguel Olivo had a legitimately great season behind the plate.
   117. Jake Luft Posted: October 31, 2006 at 03:36 PM (#2228717)
Re: Delgado
He is in no way "above average" at anything with the glove. He may have been at one point in his career, but not anymore. At best you can get through a game without noticing him, but he won't make any plays that aren't routine.
   118. John Walsh Posted: October 31, 2006 at 03:56 PM (#2228744)
Mike, I've done the breakdown by team. There are only about 500 balls per team in the 6-slice, though,
so there is a large uncertainty -- around 0.5-1% if binomial statistics holds -- on each team's result. Anyway, here is a list of the fraction of outs in the 6-zone (6+6S+6D) fielded by the 3B:
LAN    5.20%
KCA    4.10%
SFN    3.74%
HOU    3.73%
CIN    3.40%
SEA    3.21%
SLN    3.21%
ATL    3.21%
OAK    3.09%
SDN    2.93%
BOS    2.90%
PHI    2.89%
NYN    2.87%
NYA    2.78%
BAL    2.77%
TOR    2.69%
PIT    2.48%
CLE    2.38%
MIL    2.24%
TEX    2.04%
MIN    1.95%
CAL    1.61%
CHN    1.57%
CHA    1.54%
DET    1.49%
MON    1.46
   119. RobertMachemer Posted: October 31, 2006 at 06:06 PM (#2228923)
I have high hopes for Hanley Ramirez. He has tremendous range, a big-time arm and makes the flashy plays, but he vapor locks sometimes on the easy ones.
That sounds a lot like what he was like throughout the minors for Boston. I'm not sure you should expect much in the way of improvement -- just as he becomes more settled and comfortable and focused, he may also be losing range.
   120. North Side Chicago Expatriate Giants Fan Posted: October 31, 2006 at 07:35 PM (#2229035)
This is an excellent thread. I learned a lot today. Thanks to Chris and everyone else.

Chris, how was Rich Aurilia's defense this year at his various positions?
   121. Foghorn Leghorn Posted: October 31, 2006 at 08:59 PM (#2229110)
Your metric (or any other) will have other sources of uncertainty that are far larger than 2%.


Amen.
   122. Los Angeles Waterloo of Black Hawk Posted: October 31, 2006 at 09:43 PM (#2229138)
Great work, JohnW.

Is that Retrosheet data available easily on the website, or is there some kind of database you had to download and parse?
   123. Mike Emeigh Posted: October 31, 2006 at 09:47 PM (#2229145)
Is that Retrosheet data available easily on the website, or is there some kind of database you had to download and parse?


The hit location info needs to be parsed out of the event files using BEVENT or CWEVENT, and put into a DB for analysis, but it's a relatively straightforward process.

-- MWE
   124. SG Posted: October 31, 2006 at 09:54 PM (#2229151)
Chris, how was Rich Aurilia's defense this year at his various positions?

Player          Pos     INN     RSpt
Aurilia
Rich    1B    329.2    -2
Aurilia
Rich    2B    41        1
Aurilia
Rich    3B    356       5
Aurilia
Rich    SS    198.2     0 
   125. North Side Chicago Expatriate Giants Fan Posted: October 31, 2006 at 09:55 PM (#2229153)
Thanks SG.
   126. Foghorn Leghorn Posted: October 31, 2006 at 10:19 PM (#2229175)
Again, I want to endorse SG's and Kyle's response and data. They have done some outstanding work.
   127. John Walsh Posted: October 31, 2006 at 10:45 PM (#2229200)
BlackHawk,

Thanks!

There is some work involved in getting a working database setup going with the retrosheet pbp data,
but once you're over the hump, you can get info out pretty fast.
   128. Alex meets the threshold for granular review Posted: October 31, 2006 at 10:48 PM (#2229204)
Dial - what happened to "The Misunderstood?" I was all revved up for that. :)
   129. AROM Posted: November 01, 2006 at 02:54 AM (#2229324)
I'll second that for SG and Kyle. I could not have done the park factors without SG's data.

And here it is:

Park Factors

I think I'll start a thread for this one as well.
   130. Los Angeles Waterloo of Black Hawk Posted: November 01, 2006 at 02:54 AM (#2229325)
I guess I'll have to teach myself how to use a database one of these days. Or wait until someone else takes the time to publish all this wonderful data.
   131. Darren Posted: November 01, 2006 at 04:53 AM (#2229369)
Rally,

Why are your park factors expressed as /400 chances? Is that supposed to be approximately one season? Dial's stats say that corner LF gets about 350 chances per 162, or ~325 chances per 150 games. Wouldn't it be better to do LF PF as per 325 chances (150 g) and RF and CF with their corresponding numbers?
   132. Darren Posted: November 01, 2006 at 04:57 AM (#2229372)
Any chance we're going to see SG's database posted somewhere at some point? Or did I already miss it?
   133. Darren Posted: November 01, 2006 at 05:05 AM (#2229375)
Also Rally, the example given your site says "Example: Manny Ramirez has a .750 zone rating for -25 runs per 400 chances." It sounds like you're saying that Manny's actual zone rating was .750, when in fact it was .694. Did you mean to say that "If Manny's zone rating was .750....?"

Thanks.
   134. Chris Dial Posted: November 01, 2006 at 05:18 AM (#2229378)
Rally, didn't I link to that at the beginning of my article? Is that something new?

Yes, you are going to get the databases as son as we get them manageable (or you can check out Kyle's in his Forum link.)
   135. GGC don't think it can get longer than a novella Posted: November 01, 2006 at 05:18 AM (#2229379)
Darren, this isn't SG's stuff, but one of the Kyles came up with a spreadsheet.
   136. Darren Posted: November 01, 2006 at 05:31 AM (#2229385)
Chris,

You linked to his previous post, which wasn't quite the Park Factors, but something like them. His current post now has actual park factors based on matched pairs.

Thanks GGC (and to the appropriate Kyle).
   137. Darren Posted: November 01, 2006 at 05:31 AM (#2229387)
link says the Kyle file has expired. :(
   138. AROM Posted: November 01, 2006 at 05:52 AM (#2229398)
Also Rally, the example given your site says "Example: Manny Ramirez has a .750 zone rating for -25 runs per 400 chances." It sounds like you're saying that Manny's actual zone rating was .750, when in fact it was .694. Did you mean to say that "If Manny's zone rating was .750....?"

The .750 / -25 are just examples off the top of my head, not his actual ZR.

And thanks Darren for answering Chris's question.
   139. Darren Posted: November 01, 2006 at 05:54 AM (#2229399)
Rally, what number did your system have for Manny in 06?
   140. Tom Cervo, backup catcher Posted: November 01, 2006 at 06:32 AM (#2229411)
I'll be re-posting my spreadsheets soon. I've been making some changes and will incorporate Ralley's park factors, too.
   141. Mike Green Posted: November 03, 2006 at 11:58 PM (#2231253)
With the awarding of the "other" Gold Glove to Orlando Hudson, I thought that I should take a closer look. His assist/inning total remains at the top of the league, and his DP rate does also. It is true that he now gets to catch oodles of ground balls courtesy of Brandon Webb instead of Roy Halladay, but I wonder about the counting of opportunities, positioning and so on. It might be an idea to publish 3 year statistics.

Hudson's FRAA remains about the same in Arizona as it always was in Toronto, and I confess to some subjective doubt that Valentin or Biggio were better defensively than Hudson this past year. This may have something to do with the fact that "age 37" and "age 40" are not theoretical concepts to me...
   142. Chris Dial Posted: November 04, 2006 at 06:29 PM (#2231560)
Mike,
I hope to/plan to publish three year marks later.

Valentin was always a strong fielder. His rating isn't surprising to me at all. Biggio's is.
   143. BobbyMac Posted: November 05, 2006 at 03:35 AM (#2231747)
I have a question - Rafael Furcal has now led the majors in RF for 2 straight years, with two separate teams. Does anyone know:

a) Is this significant in any way? (Clearly ZR doesn't like him, and TFB Awards in the BJ Handbook put him at 5th), or

b) Has anyone else ever accomplished this feat?

In 2005, Izturis and Robles were average and below-average on RF.
In 2006, Renteria was very much below average on RF.
   144. bads85 Posted: November 05, 2006 at 10:29 AM (#2231841)
How small? Do you have a chart? Are they the same as ZR? Smaller? Larger?


Dewan changed it this year to "about five feet" in the outfield. Infield vectors are the same, and according to the description in Fielding Bible, they appear to be similar to STATS' Zones. However, the explanation is rather lacking.

The major difference (other than the explanation of your system is clear) between your system and Dewan's is that your system treats all outs as equal while Dewan assigns a value based on the the percentage of outs in that vector -- the same for outs not made. As for Jones, it appears Dewan's system must credit him for more "difficulut outs" than other CFers.
   145. Chris Dial Posted: November 06, 2006 at 02:59 AM (#2232065)
bads,
agreed. And that's granular data, and I don't know if Dewan uses an actual play value or an average play value (the actual play value is the wrong thing to use).

I appreciate your input on the zones. It's possible other descriptions are more cocksure than they should be.
   146. bads85 Posted: November 06, 2006 at 09:48 PM (#2232438)
Chris,

Dewan uses an average play value.

BTW, any thoughts on why your system ranks Pujols with 0 RSpt. while Dewan's ranks him +19 Plus/Minus (best in the majors)?
   147. AROM Posted: November 06, 2006 at 09:55 PM (#2232443)
Is that 2006 for Pujols? Are Dewan's numbers out yet? And if so, where?
   148. AROM Posted: November 06, 2006 at 10:00 PM (#2232449)
One explanation, assuming we're looking at 2006 data, is that Pujols has more balls hit to the edges of his responsibilty than the average first baseman. So he only fields an average amount, but they are tougher plays, so he gets more credit.

Also could be the zones just don't match up. I can't tell how different they are since Dewan has not (to my knowledge) published his zones.
   149. philly Posted: November 06, 2006 at 10:10 PM (#2232460)
Is that 2006 for Pujols? Are Dewan's numbers out yet? And if so, where?

There are 2006 Top 10 leaderboards for each position in the new Bill James Handbook.
   150. bads85 Posted: November 06, 2006 at 10:13 PM (#2232467)
>>>Is that 2006 for Pujols? Are Dewan's numbers out yet? And if so, where?<<<

Yes, those are Pujols' 2006 numbers. They are from the 2007 Bill James Handbook, which shipped last week. Unfortunately, Dewan's system is only presents in Leaderboard fashion in the Handbook. Here is a quick look at how Dial and Dewan's rankings matched up to the GGs:

http://journals.aol.com/bads85/ManyGoFewUnderstand/entries/2006/11/06/nl-gold-gloves-versus-pbp-data-metrics/1292
   151. AROM Posted: November 06, 2006 at 10:47 PM (#2232493)
They are from the 2007 Bill James Handbook, which shipped last week.

I've ordered mine but have not received it yet. Doesn't make me too happy.
   152. Chris Dial Posted: November 07, 2006 at 12:28 AM (#2232572)
That's some very nice work, bads. I apprecite the work you have done.
   153. Los Angeles Waterloo of Black Hawk Posted: November 09, 2006 at 02:17 AM (#2233657)
Unfortunately, Dewan's system is only presents in Leaderboard fashion in the Handbook.

Ugh. I hope they plan distributing that data in some fashion.
Page 2 of 2 pages  < 1 2

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
robneyer
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 0.4059 seconds
70 querie(s) executed