Baseball for the Thinking Fan

Login | Register | Feedback

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

Monday, January 04, 2010

2009 AL OPD (Offense Plus Defense)

The MVP voting went along the lines most people thought it should, and so it won’t be a surprise in who led the American League in OPD in 2009.  People did want to claim that Joe Mauer didn’t spread his performance out over as many games as other candidates, like Derek Jeter or Mark Teixeira, but the amount he produced then increases the chances they win the games that he does play.  That’s one of the aspects of analyzing data like this.  Mauer was a catcher, which makes him very valuable in how well he hits the ball.  It really allows the Twins to put better performers at other positions and win more games.

Mauer (68.3 XR+AA, 0.8 DRS, 69.1 OPD) outpaced the rest of the AL by 20 runs.  That’s huge.  Typically, you’ll have a handful of hitters close enough to not quibble about the MVP, or at least there is an argument for another player.  When a hitter is two full wins better, those arguments just don’t wash.  In the analysis provided here, offense and defense are included, so that moves some of the better hitters down, which isn’t always apparent from looking at numbers generally.

Last year, there was considerable discussion of how the calculation used for Extrapolated Runs (XR), that I use for calculating offensive contribution, was possibly off.  The offensive portion is XR+AA, which is eXtrapolated Runs (XR), adjusted for park and outs (+), above average at position (AA).  You can see how Defensive Runs Saved (DRS) is calculated here.  The offensive statistics come from Baseball Reference and the defensive raw data comes from CNNSI.com.  My park factors are based on a the last three seasons’ average, averaged with the last season.  Please recall that all these are “Runs Against Average”, and the typical Replacement Level (RL) is in the 25-run range.  So, if you want to compare these numbers to any other site’s Runs or Wins above RL, add 25 runs or so. 

Discussions with Tom Tango and other posters at his site (terpnats, I believe) led me to tweak the calculations a bit (doubles were too low, for example).  Colin Wyers generously provided me with recent era values for the events in XR, and so I am using those this year (and going forward).  I want to thank Tom and his readers and a special thanks to Colin, who helped with this, and generally is a big help in any research I am doing. 

The second best hitter in the AL, as measured as XR+AA, was Derek Jeter.  Jeter had a wonderful season, and many thought of him as MVP with his clutchiness and captainhood.  With traditional defensive stats, someone might say he should be MVP, but Jeter was only average defensively, and a little below (effectively average), and the two players that hit most similar to him snuck past with better defense.

The Rays had three players in the top ten, and a fourth in the top fifteen.  That’s a good number of good players.  It turns out two were likely having career seasons, and won’t repeat, but still, the Rays had a good chance to do something.  Unfortunately they were in the Yankees division, who had five players in the top 25.  Leading the way for the Rays was Ben Zobrist (40.6, 6.0, 46.6), finishing second in OPD.  Zobrist found a new power stroke in 2008, and hit a big handful of home runs in short order in a callup.  In 2009, he played all over the field, primarily at second base, and played solid defense wherever he was penciled in.  His +6 DRS on defense raised his overall performance past Jeter.

In third was Teixeira (36.5, 6.5, 43), who also had +6.5 DRS, to sneak past Jeter.  Jeter (44.7, -3, 41.7) was fourth in the league.  Alex Rodriguez, Johnny Damon and Jorge Posada can be found in the top 25 as well.  Looking at overall OPD, the Yankees were roughly twice as good as the next team - the Angels.  It was a powerful team, but almost all of it was offense.  They managed to stay near average defensively, and had solid pitching.

Jason Bartlett (40, 1.5, 41.5) was just trailing behind Jeter, but had fewer PAs.  Evan Longoria (22.9, 15.2, 38.1), the third Ray in the top ten, was seventh overall.  Carl Crawford (22.7, 7.8, 30.5) was fifteenth in the league.

In the sixth spot in OPD is the Indians’ right fielder Shin-Soo Choo (33.7, 4.8, 38.5).  That type of quality says something to me about teh Indians ability to identify good outfielders.  It isn’t clear they know how to keep them.

The Angels’ top player finished 8th in OPD and is a free agent - Chone Figgins (19.7, 17.9, 37.6).  They had other top performances from Kendry Morales (19.6, 5.8, 25.4) and Torii Hunter (22.5, 0.4, 22.9).

Next on the list was Marco Scutaro (27.6, 9, 36.6).  Last year he appeared on this list, and there was a bit of a hullabaloo because he wasn’t a full-time SS and his offense was compared to SS, and so on.  Well, he’s back.  Scutaro has really blossomed in Toronto, and he will probably be a solid player again next season.  I don’t know about top ten, but he’ll be valuable.

Rounding out the top ten is the Mariners Ichiro Suzuki (27.4 6.6, 34.0).

Some surprises and some confirmations.  You always have to consider offense, defense and position when you evaluate a player.

On a final note, the Boston Red Sox Kevin Youkilis (29.9, 3.8, 33.6) was eleventh.  The Red Sox defense, even allowing for the LF wall was not good.  The Sox notably have moved in a more defensive direction this off-season.

Here is a Google Doc spreadsheet of all the AL, with team, position, games, PAs, XR+AA, DRS and OPD.

Here’s the 2008 OPD so you can compare.

 

Chris Dial Posted: January 04, 2010 at 03:16 AM | 35 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.

   1. alilisd Posted: January 04, 2010 at 01:06 PM (#3426972)
Ichiro a Top 10 player? Preposterous! ;-)
   2. Foghorn Leghorn Posted: January 04, 2010 at 02:27 PM (#3426982)
Ha! I had said something a little snarky, and deleted it in the writeup.
   3. ekogan Posted: January 04, 2010 at 02:51 PM (#3426991)
The defensive rankings for Boston outfield are crazy - not only for LF but RF as well. And DRS is the most pessimistic of the publicly available fielding metrics about Boston fielders by a significant amount.
Data for last year, UZR from fangraphs, TZ is the Rtot column from Baseball Reference
Player            POS        UZR         DRS          TZ
Jason Bay         LF        
-13.0      -21.51         7.9
Jacoby Ellsbury   CF        
-18.6      -12.23        -9.3
JD Drew           RF         10.5        1.33        21.0 


Data for 2008
Player            POS        UZR         DRS          TZ
Jason Bay         LF         
-8.0      -11.23       -10.7
Jacoby Ellsbury   OF         16.5       
-4.59        21.8
JD Drew           RF          6.3       
-7.69        -4.5
Manny Ramirez     LF          1.5      
-10.77         4.1
Coco Crisp        CF         
-8.6      -14.94        -7.0 


Except for Ellsbury 2009, DRS is the lowest of the three in every row.
The maximum amount of disagreement between two systems is for Ellsbury 2008, where TZ & DRS disagree by 26 runs. 26 runs!!!
If you want to give a mulligan because Jacoby was playing all 3 OF positions during that time, TZ & DRS disagree about JD Drew 2009's value by 20 runs.

It seems that whatever park adjustment DRS does for Boston is too strong.
   4. Foghorn Leghorn Posted: January 04, 2010 at 03:22 PM (#3426999)
It seems that whatever park adjustment DRS does for Boston is too strong
I don't park-adjust. I almost always have that caveat, perhaps I overlooked it. That mostly matters with LF. I would suggest others over-adjust. Also, does that TZ include arm? I am not including any throwing, which is usually a factor with Boston OF.
   5. snapper (history's 42nd greatest monster) Posted: January 04, 2010 at 03:25 PM (#3427001)
Great work! Thanks a lot Chris.
   6. Foghorn Leghorn Posted: January 04, 2010 at 03:37 PM (#3427006)
I have looked at BOS LF many times, and worked on a new "change" to account for the Green Monster. It would actually involve removing all the Wall Balls, and that's going to raise the LF performance much more than previously anticipated. It appears that there are close to 20 wall balls on the Fenway LF, as opposed to 1-5 for opposing OF. That's 16-20 runs defensively, some of which is made up by playing shallower (and getting to "front of zone" balls), but I was surprised to see that the "GM Factor" was as large as 16-20 runs. This is in the early development stages, and would have to be adjusted similarly for other OFs, so the effect may only be 13-18 runs, but it's huge. Typically, I just say "Boston LF doesn't count", but removing wall balls is much more effective than a park factor.

I am not buying that Drew was anything close to that good this past season. He only played 1080 innings (77-80%).
   7. snapper (history's 42nd greatest monster) Posted: January 04, 2010 at 03:47 PM (#3427007)
It would actually involve removing all the Wall Balls

Don't you only want to remove "wall balls" that hit higher than a normal wall (~8-10 ft)? Otherwise, you're taking out catchable balls.

Of course, I don't know if this is possible from the data.
   8. Foghorn Leghorn Posted: January 04, 2010 at 03:55 PM (#3427013)
Don't you only want to remove "wall balls" that hit higher than a normal wall (~8-10 ft)? Otherwise, you're taking out catchable balls.
I don't think so. Balls that hit the wall but would be theoretically "catchable" are far too hard to judge, and the LF has to "play it safe" to prevent extra bases. If it his the wall, it's mostly uncatchable. there *will* be a few in there sometimes, but I don't think it matters (5%)
   9. AROM Posted: January 04, 2010 at 04:16 PM (#3427029)
How does position adjustment work for guys who played multiple positions? Does it account for Mauer playing a 28 games at DH?
   10. AROM Posted: January 04, 2010 at 04:18 PM (#3427030)
Fun Joe Mauer fact: He's a career 83% basestealer (34 for 41).
   11. Ivan Grushenko of Hong Kong Posted: January 04, 2010 at 04:25 PM (#3427037)
I didn't see in the link how Catcher defense was calculated. Is it there and I just missed it? Also, does XR consider baserunning other than SB/CS?
   12. Misirlou's been working for the drug squad Posted: January 04, 2010 at 04:27 PM (#3427041)
Fun Yuni fact. A score of -19 places you in the bottom 13 of all AL players. Yuni did that for 2 different teams.
   13. AROM Posted: January 04, 2010 at 04:37 PM (#3427055)
I didn't see in the link how Catcher defense was calculated. Is it there and I just missed it?


I believe it also includes PB and errors.
   14. Foghorn Leghorn Posted: January 04, 2010 at 04:42 PM (#3427062)
How does position adjustment work for guys who played multiple positions? Does it account for Mauer playing a 28 games at DH?
I don't adjust the hitting. I don't think it matters. His defense is adjusted specifically to his innings played.

I looked into this last year WRT to Scutaro (counting him at SS when he played all over). The time at DH for Mauer, and using those PAs with a higher baseline won't move his score but a run or two at most (depending on how he did during that time). I suppose for Catcher => DH that could be a few more runs. Accounting for three runs isn't worth the exercise though. It is substantially more work for a negligible gain, and many hitters get PAs at different positions (I don't adjust for pinch hitting PAs either). Overall, it just doesn't matter (IMO).

Okay, so I did a Q&D;version: Mauer vs Catchers is 68 runs (in 606 PAs). 28 games is 20% of his games played. Using that to estimate his PAs, and converting his XR+AA to 80% Catcher and 20% DH, I get 65.2 runs.

So, three runs in this extreme example (catcher with the lowest per out production, versus DH with the highest). It just doesn't matter, no longer IMO.
   15. Foghorn Leghorn Posted: January 04, 2010 at 04:45 PM (#3427066)
Ivan, it's in the link:
"Catchers are done as an amalgam of caught stealing per inning above average, stolen bases per inning above average, errors per inning above average and passed balls per inning above average, at an average base advancement of 0.31 runs per. I have vacillated between a couple of catcher run value calculations."

It doesn't include WPs, which I think is completely wrong due to the 95% standard that separates WPs and PBs. That's the C+ score, measuring traditional catcher defense. There is catcher ZR, albeit pointless (low is -1; high is 0.4 or so).
   16. Ivan Grushenko of Hong Kong Posted: January 04, 2010 at 04:53 PM (#3427071)
Thanks. You've confirmed my blindness.
   17. Craig in MN Posted: January 04, 2010 at 04:56 PM (#3427075)
I don't adjust the hitting. I don't think it matters. His defense is adjusted specifically to his innings played.

Does this explain Denard Span's curious defensive rating? He's ranked second defensively in the AL for CF (behind Gomez). Span is a very good defensive corner OF, and an ok CF, but not that nearly good. Does all the excellence in the corners get guesstimated into excellence in CF?
   18. AROM Posted: January 04, 2010 at 05:12 PM (#3427086)
On including wild pitches, the question that matters is: Are they completely dependent on the pitcher, or can some catchers prevent more wild pitches than others? If the second, then they are a part of catcher defense. But whether you leave them in or out is only going to affect the ratings by a few runs at most.
   19. Mike Emeigh Posted: January 04, 2010 at 05:25 PM (#3427101)
Are they completely dependent on the pitcher, or can some catchers prevent more wild pitches than others?


They are sometimes dependent on the official scorer. When I'm scoring a game, I *always* check the box score to see whether a pitch that elcudes the catcher is scored as a WP and PB, even when I'm 100% certain.

I agree that it doesn't matter much.

-- MWE
   20. Foghorn Leghorn Posted: January 04, 2010 at 05:32 PM (#3427110)
Does this explain Denard Span's curious defensive rating? He's ranked second defensively in the AL for CF (behind Gomez). Span is a very good defensive corner OF, and an ok CF, but not that nearly good. Does all the excellence in the corners get guesstimated into excellence in CF?
No. I am a "defensive specialist", so Span's time in CF is compared to CFs (etc), and every player's defense is position specific.

Wait - yes, he could have a high overall rating by dominating a corner OF, and those runs saved are added to his CF total. (Checking):
Yes, Span dominated LF: 6/3/-0.5
   21. plim Posted: January 04, 2010 at 05:35 PM (#3427114)
I don't think so. Balls that hit the wall but would be theoretically "catchable" are far too hard to judge, and the LF has to "play it safe" to prevent extra bases. If it his the wall, it's mostly uncatchable. there *will* be a few in there sometimes, but I don't think it matters (5%)


i think that would be the biggest improvement for outfield defense calculations. you may also want to consider it for other stadiums with tall walls (houston, seattle, etc).
   22. Foghorn Leghorn Posted: January 04, 2010 at 05:36 PM (#3427115)
Are they completely dependent on the pitcher, or can some catchers prevent more wild pitches than others? If the second, then they are a part of catcher defense. But whether you leave them in or out is only going to affect the ratings by a few runs at most.
Well, they are *mostly* the fault of the pitcher. So it would take more work breaking that down, for next to nothing, so I leave them out.
   23. Foghorn Leghorn Posted: January 04, 2010 at 05:39 PM (#3427119)
you may also want to consider it for other stadiums with tall walls (houston, seattle, etc).
I agree, but there's an advantage to Boston - it is so shallow you can take Retrosheet data to ID wall balls. I am not sure that's true for other parks (yet). The other parks will have some factor (balls that hit the wall that would have been HRs elsewhere and not count), but that just isn't as common outside Fenway. Houston is the biggest one, and Baltimore will have some.

Until I move that to video, it will be a while.
   24. Lassus Posted: January 04, 2010 at 05:56 PM (#3427140)
I can't at the moment really be bothered to attempt to understand all these numbers; however, I might eventually, so thanks a lot for the work, Chris, this is all quite fascinating to read.
   25. PreservedFish Posted: January 04, 2010 at 05:59 PM (#3427145)
So Bay is a -20 fielder, or he might be a -5 fielder.
   26. Foghorn Leghorn Posted: January 04, 2010 at 06:09 PM (#3427155)
No, Bay is NOT a -20 fielder. He may be a -5 to -15 fielder. Based on his Pittsburgh work, I'd estimate -8. - actually, due to his age, maybe a -12.
   27. Frisco Cali Posted: January 04, 2010 at 06:44 PM (#3427189)
I'm sure this has been asked and answered before, but when I see Carp rated higher than Konerko, I have to wonder what is being measured.
I suppose this is on the FAQ page...
   28. jacksone (AKA It's OK...) Posted: January 04, 2010 at 06:52 PM (#3427199)
I'm sure this has been asked and answered before, but when I see Carp rated higher than Konerko, I have to wonder what is being measured.
I suppose this is on the FAQ page...


65 plate attempts for Carp. 621 for Konerko. It's a sample size issue. Anyone can look good (or terrible) in small doses.
   29. snapper (history's 42nd greatest monster) Posted: January 04, 2010 at 07:06 PM (#3427219)
65 plate attempts for Carp. 621 for Konerko. It's a sample size issue. Anyone can look good (or terrible) in small doses.

This, and the fact that Chris is measuring production above average. If you adjusted to runs above replacement, Konerko would get ~20 R's added for playing nearly a full season (621 PAs). Carp would only get 2 or 3 R's for his 65 PAs so Konerko would end up ~+25 RAR and Carp ~+7.

Being an average MLBer has a ton of value that doesn't show up directly in RAA.
   30. AROM Posted: January 04, 2010 at 07:29 PM (#3427251)
Well, they are *mostly* the fault of the pitcher. So it would take more work breaking that down, for next to nothing, so I leave them out.


Fault doesn't really matter. Say every wild pitch is the fault of the pitcher, and you've got a staff that throws 50 wild pitches on a season. Then Yadier Molina comes around, and somehow gets in front of enough of those terrible pitches to save 20 of them from being wild pitches. That's value.

The homerun is always the fault of a pitcher. But if a centerfielder like Torii Hunter can bring two of those back per season, he's providing value.

It should only be ignored if the catcher has nothing to do with it. Say a catcher has a very low percentage of wild pitches for a few years. Then he gets signed by a team that threw a lot of wild pitches. If he brings down the number of WP, he's shown a skill. If those pitchers keep throwing the same # of WP with the new catcher, then there is no skill, it's all on the pitchers and random noise as far as the catchers are concerned.

I think it's a skill, you don't, but I don't think either of us has the data. If I'm wrong, and there's no more evidence of a catcher's ability on WP than there is for catcher ERA being meaningful, I'll gladly revise my catcher ratings.
   31. Foghorn Leghorn Posted: January 04, 2010 at 07:36 PM (#3427260)
I think it's a skill, you don't,
No, blocking pitches in the dirt *is* a skill. Opportunity is varied terrifically (far more so than anything else we analyze), and the catcher influences that number significantly. A home run isn't "all the pitcher's fault." So, instead of slicing paper-thin percentages, we make slightly broader brushes. WPs are the pitchers fault. PBs are the catchers (and they aren't all the catcher's fault either). Let's not digress into that.
   32. Frisco Cali Posted: January 04, 2010 at 09:22 PM (#3427415)
Thanks Snapper. I knew I was missing something obvious.
   33. Harold can be a fun sponge Posted: January 05, 2010 at 01:06 AM (#3427659)
Thanks for doing the work on this, Chris. I liked your writeup.

One thing I've noticed from the comment thread is that there are a few areas where OPD doesn't try to get something exactly right, because it's not worth the effort/complexity for a minor gain (for example, adjusting DRS for park, or counting PAs by position). I agree with Dial about the tradeoff, but I do want to point out that most of the top ten is within ten runs of each other. I don't think it's worth adjusting OPD to gain a run here or there in accuracy, but we should remember that uncertainty when using it to rank players or whatever.
   34. Chris Dial Posted: January 05, 2010 at 03:06 AM (#3427766)
Harold,
I don't disagree that when players are within 5 runs or so, and maybe ten, we look a little more closely. However, I'd suggest hat the difference in the players isn't likely to be definitive anyway, as the error in the measures we have (offense or defense) are accurate to a run or two. There's enough difference in PFs use, specific run environment, situational events to say that, when players are within 5 runs (or so), they are "equal".
   35. Foghorn Leghorn Posted: January 05, 2010 at 10:05 PM (#3428708)
Austin Kearns signed with the Indians. He's an absolutely terrific defensive outfielder. So the Tribe can ID them.

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Chicago Joe
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 0.3955 seconds
68 querie(s) executed