Baseball for the Thinking Fan

Login | Register | Feedback

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

Monday, October 20, 2008

Final 2008 Offense Plus Defense (OPD) Results

With the ALCS wrapped (Congratulations all you former Durham Bulls!), and three days off until the World Series starts, I thought I’d get the end-of-year results for offense and defense posted.  It will enhance your “Who should be MVP?” debates.  It will contain the best fielders and best hitters, so you can work out your Gold Glove and Silver Slugger pools.  People can even pay off on their bets where “Who is better?” came up.  Most importantly for all the fans of teams who are eliminated, you can look at players about to be free agents and decide who you want to sign.  Sure, he can hit, but can he field?  Is $10 million too much for a player going to provide you one win above average?
Final 2008 OPD Spreadsheet

The runs are rated above average at position.  The offense is XR, park-adjusted, and specific to the number of outs a player has used up.  Baseball-Reference tweaked their league batting pages to provide all the necessary categories in one place to calculate Extrapolated Runs, so this will be much easier than in years past to generate.  Thanks, Sean!  The defense is DRS (Defensive Runs Saved: ZR converted to runs), explained in my previous work.  It is runs, not plays, above average. The units are the same, so I simply add the numbers together.  The decimal places are for consistency’s sake, not meant to represent accuracy.  There are several runs of give in these (and any) numbers, offense or defense.  Also, for defense, actual chances are used instead of estimated, as those are now available.

First I want to talk about the National League MVP.  I think Albert Pujols clearly deserves it.  He’s got a good lead over the next player, and was terrific all season.  I was surprised to see that Hanley Ramirez’ bat at his position was every bit as valuable.  If Hanley continues to improve on defense, he could be cradling the trophy very soon.  Here are the top 15 in the National League:

Year Lg Player LastName Age Team G POS XR+AA DRS OPD
2008 NL Albert Pujols 28 STL 148 1B 61.2 14.5 75.7
2008 NL Chipper Jones 36 ATL 128 3B 52.5 9.4 61.9
2008 NL Hanley Ramirez 24 FLA 153 SS 61.3 -0.6 60.7
2008 NL Chase Utley 29 PHI 159 2B 38.4 12.9 51.2
2008 NL Lance Berkman 32 HOU 159 1B 38.1 12.5 50.6
2008 NL Carlos Beltran 31 NYM 161 CF 34.6 6.7 41.2
2008 NL David Wright 25 NYM 160 3B 42.7 -4.8 37.8
2008 NL Matt Holliday 28 COL 139 LF 28.6 8.2 36.8
2008 NL Brian Giles 37 SDP 147 RF 25.4 11.3 36.8
2008 NL Ryan Ludwick 29 STL 152 RF 36 0.4 36.4
2008 NL Manny Ramirez 36 LAD 53 LF 31.9 2.2 34.1
2008 NL Brian McCann 24 ATL 145 C 33.4 -3 30.4
2008 NL Dan Uggla 28 FLA 146 2B 29.4 -2 27.4
2008 NL Jody Gerut 30 SDP 100 CF 17.5 9.6 27.1
2008 NL Jose Reyes 25 NYM 159 SS 35.3 -8.7 26.6

It is a list of the NL stars, with some Jody Gerut thrown in.  What is interesting is the lack of any Cubs.  And Manny Ramirez showing up in a third of a season.  If you are looking for some of the MVP candidates that you read about in MSM articles, well, Ryan Howard is down there about 60th with 7 runs above average OPD.  He’s seven wins behind Albert.  Should he win, I suspect that will be the greatest traveshamockery of all time.

The AL MVP race is pretty close.  I would give the nod to Minnesota’s Joe Mauer, ekeing out a win over Grady Sizemore.  Both played well, and a premium position.  There was lots of late season talk about Dustin Pedroia, and he did keeps his level of play up and is just about 10 runs behind Mauer.  He is as close to Mauer in fourth as the second place player in teh NL.  So, he could win it, and while it wouldn’t be the best choice, it is not nearly as hideous as some we’ve seen in the past.

The top 15 AL players:

Year Lg Player LastName Age Team G POS XR+AA DRS OPD
2008 AL Joe Mauer 25 MIN 146 C 37.8 7.37 45.16
2008 AL Grady Sizemore 25 CLE 157 CF 35.6 6.78 42.33
2008 AL Alex Rodriguez 32 NYY 138 3B 41.1 -1.48 39.64
2008 AL Dustin Pedroia 24 BOS 157 2B 22.3 12.6 34.94
2008 AL Kevin Youkilis 29 BOS 145 1B 28.8 5.86 34.69
2008 AL Milton Bradley 30 TEX 126 DH 35.9 -1.95 33.95
2008 AL Carlos Quentin 25 CHW 130 LF 30.8 -0.97 29.87
2008 AL Brian Roberts 30 BAL 155 2B 24.6 3.6 28.23
2008 AL Aubrey Huff 31 BAL 154 DH 29 -1.64 27.32
2008 AL Nick Markakis 24 BAL 157 RF 30.5 -3.32 27.15
2008 AL Carlos Pena 30 TBR 139 1B 24.9 1.4 26.32
2008 AL Ian Kinsler 26 TEX 121 2B 25.9 -0.01 25.93
2008 AL Josh Hamilton 27 TEX 156 CF 32.2 -6.35 25.83
2008 AL Mark Teixeira 28 LAA 54 1B 24 1.12 25.08
2008 AL Marco Scutaro 32 TOR 145 SS 6.6 18.42 25.03

I didn’t really realize just how good ARod’s season was.  He’s a good player.  It’s still amazing to see three Oriole players right there.  Mark Teixeira came over for a third of a season and ripped up the AL pretty good.  He and Manny both had a big impact on the pennant races.

The team OPD is also included in that spreadsheet.  So open up the spreadsheet, poke around and argue away at who is the best, who your team needs, and who your team needs to jettison.

Chris Dial Posted: October 20, 2008 at 05:04 PM | 86 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. HMS Moses Taylor Posted: October 20, 2008 at 05:55 PM (#2990137)
What is interesting is the lack of any Cubs.

Not really. The story about the Cubs this season was depth and very few black holes (by the end of the year, Fukudome was an offensive black hole, but that's about it; him and Theriot were the only regulars in the negative here*). They don't have any great players, and their best players didn't have their best years. IIRC, DeRosa was the highest ranked Cubs just about any time you updated this so it's not surprising at this time of year. As you're quick to point out, this was easily a career year for him and the Cubs can't count on this again next year.

*Should Derrek Lee be at a negative -10? That has to be a mistake.
   2. Ron Johnson Posted: October 20, 2008 at 06:37 PM (#2990189)
Lee looks right to me. Hitter's park. Below average for the position. 27 GIDP (though I maintain any inclusion of DPs without a context adjustment is a minor method error) Yeah 8/2 SB/CS but that's not worth much.
   3. Chris Dial Posted: October 20, 2008 at 06:39 PM (#2990191)
That seems a *little* low, but in VORP he's at least 65 runs behind Pujols. I'll re-check the data this evening.
   4. Chris Dial Posted: October 20, 2008 at 06:40 PM (#2990193)
Ron, I have some defensive PF ideas I'd like to bounce off you. Are you game?
   5. Biff, highly-regarded young guy Posted: October 20, 2008 at 06:41 PM (#2990195)
It seems like a foregone conclusion to me that the writers are going to pick Pedroia, but I'm willing to be proven wrong.
   6. Frisco Cali Posted: October 20, 2008 at 06:42 PM (#2990197)
Fearless Prediction:

Sell high on Scutaro.
   7. Danny Posted: October 20, 2008 at 06:46 PM (#2990207)
Fearless Prediction:

Scutaro's offense shouldn't be compared solely to other SS since he spent most of his time at other positions.

His ZR really was off the charts good though, especially compared to what he's done before.
   8. Ron Johnson Posted: October 20, 2008 at 06:50 PM (#2990210)
Chris, fire away -- either here or by email.
   9. Scoriano Flitcraft Posted: October 20, 2008 at 06:53 PM (#2990211)
Abreu: -21.99 Yankee corner OF's were malatrocious.
   10. Chris Dial Posted: October 20, 2008 at 06:57 PM (#2990215)
Scutaro's offense shouldn't be compared solely to other SS since he spent most of his time at other positions.
Right, but he's the only one, and he's really not going to fall very much from that if you spread around his offense, I don't think it will drop much.
   11. Chris Dial Posted: October 20, 2008 at 06:57 PM (#2990217)
Ron, I'll send you an email this evening with some data.
   12. Padraic Posted: October 20, 2008 at 07:36 PM (#2990265)
Nice to see the 'overated' Rollins come in as the 3rd best SS in baseball,* despite the injuries and what has generally been described as a down year.

*I'm not counting Scuturo, since he wasn't a full-time SS.

With the time back from injury and a slight boost to his early season numbers when he was clearly still hurt, he would have caught Reyes too.
   13. Chris Dial Posted: October 20, 2008 at 07:41 PM (#2990272)
With the time back from injury and a slight boost to his early season numbers when he was clearly still hurt, he would have caught Reyes too.
Not likely. Reyes destroyed him offensively, but really had a poor defensive season.
   14. They paved Misirlou, put up a parking lot Posted: October 20, 2008 at 07:45 PM (#2990279)
Lee looks right to me. Hitter's park. Below average for the position. 27 GIDP (though I maintain any inclusion of DPs without a context adjustment is a minor method error) Yeah 8/2 SB/CS but that's not worth much.


OK, but 5 runs worse than Fukudome offensively?
   15. Chris Dial Posted: October 20, 2008 at 07:45 PM (#2990280)
'overated' Rollins
He was overrated because people think he belongs(ed) in an MVP discussion, not in a neck-and-neck battle with Mike Fontenot for the 30th best player in the NL.
   16. Chris Dial Posted: October 20, 2008 at 07:45 PM (#2990281)
OK, but 5 runs worse than Fukudome offensively?
Not a strong year for RF, and it was for 1B.
   17. Mister High Standards Posted: October 20, 2008 at 08:09 PM (#2990315)
Dustin seems like a very reasonable MVP candidate, which is a little shocking.
   18. Ron Johnson Posted: October 20, 2008 at 08:14 PM (#2990319)
Another way to look at how strong 1B was in the NL. Ryan Howard's not a heck of a lot better than average offensively.
   19. DL from MN Posted: October 20, 2008 at 08:16 PM (#2990323)
Chipper isn't even appearing on the MSM MVP ballots which is going to be the real travesty.
   20. YLT Posted: October 20, 2008 at 09:06 PM (#2990367)
The team numbers support my idea that the Dodgers should try to sign Teixiera and then trade Loney. Or are we still thinking that Loney is going to start walking and hitting for power?
   21. villageidiom Posted: October 20, 2008 at 09:24 PM (#2990389)
Chris - In the past were you making an adjustment to Manny's defensive numbers to adjust for the screwiness that accompanies Fenway LF stats? If so, did you continue to make those adjustments to Manny once he moved to LA? Likewise, are Bay's numbers reflecting any Fenway LF adjustments?

Great stuff, BTW.
   22. dcsmyth1 Posted: October 20, 2008 at 09:31 PM (#2990400)
Back to D Lee. I watched most Cubs games, and didn't realize that the following players all contributed more:

R Cedeno, D Ward, KOYIE HILL (10 G), F Pie, E Patterson (13 G), C MCGEEHEE (9 G), H Blanco, M HOFFPAUIR, M Murton (19 G)

Maybe it's just me, but mightn't there be a problem with the baseline Dial is using?
   23. Eric J can SABER all he wants to Posted: October 20, 2008 at 09:42 PM (#2990416)
22: If memory serves, the baseline Dial is using is average. So someone who is mildly below average in 150 games could easily come out worse than Casey McGehee, who isn't a major league player and plays like it, in 9 games. If Lee plays at, say, 95% of average in 17 times as many games as McGehee plays, McGehee would have to play at 15% of average to match him. If it's 93%, McGehee would have to be below 0, which is difficult for even the worst of players. And 93% of average isn't necessarily a gaping hole.
   24. Yeaarrgghhhh Posted: October 20, 2008 at 09:45 PM (#2990418)
Another way to look at how strong 1B was in the NL. Ryan Howard's not a heck of a lot better than average offensively.

Is the baseline based on 2008 and only the player's league, or is it broader (e.g., a 3 year average of the position for both leagues)? If the former, does that mean that Hank Greenberg would be average or below average according to this system because he played in the same league as Foxx and Gehrig?
   25. dcsmyth1 Posted: October 20, 2008 at 09:56 PM (#2990428)
Eric 23, I understand exactly how the avg baseline works. Having always advocated repl level for over 20 years, I was actually starting to reconsider towards .500, based on the difference between the concepts of 'franchise value' and 'in game' value. But seeing results like the McGeehee vs D Lee comparison is enough to rid me of any such notion...
   26. Ron Johnson Posted: October 20, 2008 at 10:01 PM (#2990434)
Duffy, it's you.

Seriously, the fact that he's using average as a baseline is no big deal. Works pretty well for full-timers and comparing the regular's rating to the team numbers will tell you what you need to know about the backups.
   27. Eric J can SABER all he wants to Posted: October 20, 2008 at 10:01 PM (#2990435)
Duffy, I'm a replacement level advocate myself (in more ways than one, probably), so I'm not going to disagree.
   28. bigboy1234 Posted: October 20, 2008 at 10:06 PM (#2990439)
Chris, is there a reason LAA doesn't have a DH total in the "AL Team OPD" section? Thanks for the great work as always.
   29. Walt Davis Posted: October 20, 2008 at 10:06 PM (#2990441)
Punto more valuable than Thome, Ordonez, Ortiz, Ichiro and future MVP :-) BJ Upton.

I love it!

Maybe it's just me, but mightn't there be a problem with the baseline Dial is using?

As #23 said in maybe not the clearest fashion, the baseline is average so playing time "hurts" you. In 9 games you can do little damage compared to the average player. Measure against replacement and Lee would come out much better.

Say you're 1 run below average per 15 games. That would make you 10 runs below per 150. That would also be about 10 runs above replacement per 150.

Now if you play in 150 games, you're 10 below average or 10 above replacement. If you play in 15 games, you're one run below average or 1 run above replacement. If you're in fact replacement level but play in only 15 games, you're still just 2 runs below average and "look better" than the full-time player who's a win above replacement.

At the team level, I think measuring vs. average makes sense. At the individual player level, if you limited it to "starters" or you were measuring/projecting for several seasons, average makes sense. For example, for HoF discussions, I think one should measure against average. But if you're measuring for a single season and especially if you're mixing part-timers and full-timers, measuring against replacement makes a lot more sense.

Then there's the ridiculously high baseline for average NL 1B in 2008. Adam LaRoche had a 120 OPS+ this year and came in below average offensively. This isn't so much a measure of how good LaRoche/Lee are compared to the average ML 1B (in terms of value or talent), it's how much value they produced vs. the average NL 1B in 2008. So don't get too worked up.
   30. Yeaarrgghhhh Posted: October 20, 2008 at 10:14 PM (#2990447)
22: If memory serves, the baseline Dial is using is average. So someone who is mildly below average in 150 games could easily come out worse than Casey McGehee, who isn't a major league player and plays like it, in 9 games. If Lee plays at, say, 95% of average in 17 times as many games as McGehee plays, McGehee would have to play at 15% of average to match him. If it's 93%, McGehee would have to be below 0, which is difficult for even the worst of players. And 93% of average isn't necessarily a gaping hole.

IIRC, last year JR House in 2-3 games had more value to the Orioles than Nick Markakis in 160. Maybe I just don't understand what this system is measuring, but this seems like a fundamental flaw with the system.
   31. Yeaarrgghhhh Posted: October 20, 2008 at 10:17 PM (#2990450)
At the team level, I think measuring vs. average makes sense. At the individual player level, if you limited it to "starters" or you were measuring/projecting for several seasons, average makes sense. For example, for HoF discussions, I think one should measure against average. But if you're measuring for a single season and especially if you're mixing part-timers and full-timers, measuring against replacement makes a lot more sense.

Ah...should have refreshed before I posted.

Then there's the ridiculously high baseline for average NL 1B in 2008. Adam LaRoche had a 120 OPS+ this year and came in below average offensively. This isn't so much a measure of how good LaRoche/Lee are compared to the average ML 1B (in terms of value or talent), it's how much value they produced vs. the average NL 1B in 2008. So don't get too worked up.

But does it make sense to compare LaRoche to the average NL 1b in 2008? Why not compare him to the average ML 1b from 2006-2008?
   32. Ivan Grushenko of Hong Kong Posted: October 20, 2008 at 10:28 PM (#2990462)
I don't care what qualifiers you put on Scutaro's performance. The choice of Crosby over him really hurts. Marco! Scutaro!
   33. Eric J can SABER all he wants to Posted: October 20, 2008 at 10:38 PM (#2990470)
I wouldn't call it a flaw in the system so much as a difference of opinion on what you'd like to see measured. There are multiple baselines for which one could make an argument. If you prefer a different one, well, Dial gives you the spreadsheet; it's not too taxing to add, say, 2 runs for every 15 games and switch to a replacement level baseline. (The multi-year baseline would be a little more involved, but still doable.)
   34. Ron Johnson Posted: October 20, 2008 at 10:47 PM (#2990476)
The baseline's 2008 only. And as I suggested quite some time ago to Keith Woolner it's a minor method error. No biggie to me given that the standard error for a full time player in this kind of structure can't be less than about 8 runs.

I use these kind of lists as a starting point -- list the regulars by position and fine tune for any unusual concentrations. (The team totals are most useful for teams who never truly settled on a regular.)

Works particularly well in the AL when dealing with DH versus am equally good hitter who happens to be a bad defensive player.

Ideally you'd want to establish a baseline at long-term replacement level and use a structure (like win shares) such that any defensive value is a positive -- it's simply a question of how much. But I can pretty much get there from Chris' lists. Sort by position and merge.

And which of Greenberg's good years would Foxx and Gehrig have driven him below average? Five other regulars after all and most of them weren't that good -- the other five starters averaged around .288/.356/.431 for the years in question.
   35. Walt Davis Posted: October 20, 2008 at 11:09 PM (#2990488)
But does it make sense to compare LaRoche to the average NL 1b in 2008? Why not compare him to the average ML 1b from 2006-2008?

Depends what you're trying to measure. Who was the most valuable (or best-performing) 1B in the NL in 2008? Did LaRoche have a good season? Where does LaRoche rank compared to MLB 1B in terms of performance over the last 3 years? How do we project LaRoche going forward?

The first question, which seems the one Chris is interested in answering with this spreadsheet (and expanding it to all players of course), should use just this year. The second one could be argued either way. The latter two you would want multiple years of data.

Chris isn't claiming that Derrek Lee is below average in terms of true talent, he's not claiming he'll be below average next year, he's not even claiming he was below average this year compared to MLB 1B -- he's just saying he was below-average in 2008.

The main thing to remember is that the man is providing reasonably complex, reasonably high quality information for free -- he can set the baseline wherever he wants. :-)

And which of Greenberg's good years would Foxx and Gehrig have driven him below average? Five other regulars after all and most of them weren't that good -- the other five starters averaged around .288/.356/.431 for the years in question.

Well, probably none, but let's remember the difference between mean and median. Pujols' monster year could well be pulling the 1B _average_ above the 1B median.

Wait a second! JT Snow played this year? I take it a "I want to retire a Giant" sort of thing -- 1 game, no PA and, near as I can tell, never played the field. Wow, I didn't even know this counted -- he was named in the starting lineup then, apparently, replaced before the first pitch.

And he came up through the Yankees system?
   36. RedSoxBaller Posted: October 20, 2008 at 11:22 PM (#2990498)
I'm not saying I have a better stat or anyting, but this OPD is seriously flawed. Arod is in no way a top 5 player. He batted what .230 with RISP? He just can't hit when it matters, even if you argue that clutch does not exist. Jody Gerut a top 20 player? Really, you would take him over Ryan Howard, Russel Martin, or Nate Mclouth? And where is Justin Morneau in the AL? Surely you must have forgotten him.
   37. dcsmyth1 Posted: October 20, 2008 at 11:23 PM (#2990499)
#26--"Duffy, it's you!"

HaHa. I guess my wording in #25 may have been a bit pompous. But, if I used my real name, I think many of you might recognize it.

Anyway, someone wrote that it's not a big deal, you can just convert the spreadsheet to repl level by adding 20 runs per 162G (or whatever). True, but why doesn't Dial just do that in the first place? Given the likely use this list will be used for by readers, why not simply make a reasonable stab at it. Or, make different lists of starters and backups, etc.

Any of those things are better than just having McGehee ahead of D Lee, them's the numbers, like it or not.
   38. Yeaarrgghhhh Posted: October 20, 2008 at 11:42 PM (#2990512)
Depends what you're trying to measure. Who was the most valuable (or best-performing) 1B in the NL in 2008? Did LaRoche have a good season? Where does LaRoche rank compared to MLB 1B in terms of performance over the last 3 years? How do we project LaRoche going forward?

But the system appears to be measuring who the best player in each league was for 2008, and players like Howard and Lee seem to suffer when compared to players at other positions because the baseline at their position is so high.
   39. Iwakuma Chameleon (jonathan) Posted: October 21, 2008 at 12:42 AM (#2990542)
Wait a second! JT Snow played this year? I take it a "I want to retire a Giant" sort of thing -- 1 game, no PA and, near as I can tell, never played the field. Wow, I didn't even know this counted -- he was named in the starting lineup then, apparently, replaced before the first pitch.


From what I understand, he went in and stood around at first while the pitcher was warming up and then subbed back out before he actually had to play. The purpose being, of course, as you already guessed, to simply retire a Giant.
   40. Mefisto Posted: October 21, 2008 at 01:42 AM (#2990562)
Chris, is there a database you could use for baserunning (non-SB)? Just eyeballing the figures, that could make a difference in the AL leaders.
   41. Chris Dial Posted: October 21, 2008 at 01:44 AM (#2990564)
Sorry work was so hard on me today. Walt and Ron answered everything pretty much for me, but I'll chime in a few unanswered ones.
But the system appears to be measuring who the best player in each league was for 2008, and players like Howard and Lee seem to suffer when compared to players at other positions because the baseline at their position is so high.
IMO< those players were less valuable because their teams' opponents were putting much better players on the field at their position. It's not impressive, nor particularly helpful for the Cubs, for DLee to outhit Cristian Guzman in raw numbers. DLee needs to outperform the other first basemen in the NL. He doesn't "suffer" - he under-performs compared to what other teams got from the position he plays. If the Cubs had gotten Joey Votto's performance instead of DLee's, they would have been a better team. Players are always (IMO) properly compared to how the position they played hits.

The system *is* measuring who the best players were in each league at each position.
   42. Chris Dial Posted: October 21, 2008 at 01:48 AM (#2990567)
why doesn't Dial just do that in the first place? Given the likely use this list will be used for by readers, why not simply make a reasonable stab at it. Or, make different lists of starters and backups, etc.

1. There is no definition of replacement player that is demonstrable. If I use one, someone argues its the wrong one. There isn't a wrong "average".
2. I don't really want to, mostly because I don't believe in it, and it's just a different line.
3. I put the games played on there so you could just look at it and see who the starters were. The spreadsheet can be copied into your personal Excel where you can change it in whatever manner you'd like. It's pretty open source and freely available in a format for you to use.
   43. Chris Dial Posted: October 21, 2008 at 01:53 AM (#2990569)
In the past were you making an adjustment to Manny's defensive numbers to adjust for the screwiness that accompanies Fenway LF stats? If so, did you continue to make those adjustments to Manny once he moved to LA? Likewise, are Bay's numbers reflecting any Fenway LF adjustments?
I have never park-adjusted any defensive stats formulaically. I have tried to always write a caveat for Fenway's LF. I have recently backed into a park factor solution that I will bounce off some trusted statheads, and then write up, and then retro-actively apply to the larger database. But that's not a fast process. Joe Arthur (Google him) has done some great work for a Fenway park factor. My research out of the gate (very ROUGH DRAFT) looks like 0.84 effect FOR LF ONLY. So, the LF for Boston would see his ZR improve by dividing by 0.92. The largest problem with defensive PFs is going to be that the effect of teh Green Monster (and the Crawford Boxes) isn't consistent. It depends on how shittily the Sox pitchers pitch. Hopefully, I can get enough data to smooth those out.
   44. Chris Dial Posted: October 21, 2008 at 01:54 AM (#2990570)
Chris, is there a reason LAA doesn't have a DH total in the "AL Team OPD" section? Thanks for the great work as always.
Yes, there is a reason. Everyone that DHed for them played another position primarily (I didn't double-check this, but pulled the numbers from B-Ref).
   45. Chris Dial Posted: October 21, 2008 at 01:58 AM (#2990572)
Chris, is there a database you could use for baserunning (non-SB)? Just eyeballing the figures, that could make a difference in the AL leaders.
I think one could use John Dewan's baserunning numbers, which are in "runs", so those could simply be added. I certainly think with Sizemore that close to Mauer it could matter. That was a good point I hadn't considered, *but* 3-5 runs is about the max spread.

(And good to see you - I was filing my travel report and thought of you today).
   46. Chris Dial Posted: October 21, 2008 at 02:00 AM (#2990574)
And thanks to Walt, Rona nd Eric J (sorry I left you out the first time!). After 10+ years of doing this it's gotten where others can explain it better than I can.
   47. Eric J can SABER all he wants to Posted: October 21, 2008 at 02:07 AM (#2990579)
I certainly think with Sizemore that close to Mauer it could matter.


I'm not sure it would; Sizemore is a better baserunner than Mauer, but the average CF would be a better baserunner than the average catcher as well. I would guess that it'd be very close to a wash.
   48. fra paolo Posted: October 21, 2008 at 02:36 AM (#2990599)
Looking at the team stats makes me realize just how bad the Tigers' pitching was.
   49. WillYoung Posted: October 21, 2008 at 02:51 AM (#2990601)

I'm not sure it would; Sizemore is a better baserunner than Mauer, but the average CF would be a better baserunner than the average catcher as well. I would guess that it'd be very close to a wash.


And Mauer is a very, very good baserunner. He has inherited the "deceptively quick, and surprisingly effective baserunner" role on the Twins from Corey Koskie who had inherited it from Paul Molitor. You'll rarely see Mauer thrown out on the bases yet he gets exceptional jumps and will go from first-to-third and other things much better than you would expect.
   50. Chris Dial Posted: October 21, 2008 at 03:12 AM (#2990612)
Good points, Eric and Will.
   51. DCW3 Posted: October 21, 2008 at 06:02 AM (#2990667)
It's probably better to say not that Casey McGehee was "more valuable" than Derrek Lee, but that he "hurt his team less." Below-average players have value in that they prevent their teams from being stuck with even worse options, which is the reason for the replacement-level framework. But they are still a negative. If a team gets average performance from every other position and has the 2008 version of Derrek Lee at first base, then they are going to lose more games than they win.
   52. dcsmyth1 Posted: October 21, 2008 at 11:46 AM (#2990694)
It's probably better to say not that Casey McGehee was "more valuable" than Derrek Lee, but that he "hurt his team less/quote]

I understand what you are trying to say, but in the end it's just semantics. The more I think about it, the more I think the best solution is just to have different lists--starters, regular backups, and occasional/fringe players. That way you can stick with the .500 baseline and not look ridiculous, or have to do a dance with wording.
   53. Ron Johnson Posted: October 21, 2008 at 12:00 PM (#2990704)
It's not even accurate to assert Lee hurt the Cubs. Anybody involved in this discussion knows that average players aren't freely available. Milwaukee's 1B situation was closest to league average and I'd think they'd want something for Fielder.

You'll get a good sense of where replacement level is by looking at the team OPD. My sense is that replacement level is something close to -30 per Ripkenseanson (IE every inning of every game) -- maybe -35, and yes some teams did worse -- so Lee's "true" value is something close to +18 (a replacement level player could be expected to be around -28 runs in Lee's playing time.) Mcgehee's set against a baseline of roughly zero and finished up at -2. Dead easy to incorporate into Chris' spreadheet. Won't have much impact on what Chris is interested in -- the MVP discussions.

NL RF's an instructive point on the danger of using single year as the baseline. Kearns put up an OPS+ of 66 (but Washington finished close to positional average as a team since Dukes played pretty well) and Francouer gave the Braves 155 games of a 73 OPS+ and there were other open wounds getting significant playing time (Jenkins with 72 starts for instance)
   54. Chris Dial Posted: October 21, 2008 at 12:29 PM (#2990720)
I understand what you are trying to say, but in the end it's just semantics. The more I think about it, the more I think the best solution is just to have different lists--starters, regular backups, and occasional/fringe players. That way you can stick with the .500 baseline and not look ridiculous, or have to do a dance with wording.
Duffy, on a historical basis I used to only do players that played 750-1000 innings. What happens is I get vastly more questions on how this player or that player did. I appreciate you think most of these readers think along the lines of replacement level, but they don't. RL is more ambiguous than you seem to give it credit for. Yes, the people that generally comment like/want RL, but the other 8000 readers "get" average.
   55. Chris Dial Posted: October 21, 2008 at 12:29 PM (#2990721)
Ron,
sorry I didn't get to this last night. My sister was in town with her family as a prospective student. Hopefully I can get to it this evening.
   56. Chris Dial Posted: October 21, 2008 at 12:38 PM (#2990727)
As Ron perfectly notes, the differences between my data and VORP is about 35 runs. Uses MLB, rather than each league, so that accounts for some differences, and then there are the Tango criticisms of VORP.

I *think* I have access allowed to the 2007 data shared in those Google Docs (or in the 2007 articles in my column). That gives you two years. I am working on having the entire pbp era OPD (1987-present) made searchable and accessible.

Then you wouldn't have to worry about much of this - because a database can have extra columns without as much maintenance.
   57. Harris Posted: October 21, 2008 at 01:10 PM (#2990743)
With the time back from injury and a slight boost to his early season numbers when he was clearly still hurt, he would have caught Reyes too.

Not likely. Reyes destroyed him offensively, but really had a poor defensive season.


I think Rollins will take his season over Reyes.
Rollins is prepping for the world series. Reyes is playing golf.

Then again....seeing as how Reyes performs in September...maybe he likes golf more?
   58. villageidiom Posted: October 21, 2008 at 01:25 PM (#2990759)
I have never park-adjusted any defensive stats formulaically. I have tried to always write a caveat for Fenway's LF.
Fair enough.

Player     Team   Gms    DRS

Bay        PIT    106   
-10.1
Bay        BOS     49   
-11.2
Ramirez    LAD     53   
2.2
Ramirez    BOS    100   
-10.8 


That just seems wrong to me. Jason Bay isn't going to be mistaken for Carl Crawford, but his total DRS of -21.3 for 2008 is the third worst in the league, just behind Brad Hawpe (-23.2) and Bobby Abreu (-21.99) and just ahead of Carlos Lee (-21.1). I didn't see much of Hawpe this year, but what I've seen of him the last couple of years hasn't been pretty. I've seen plenty of Abreu this year, and on fly balls he seemed afraid that the gods would smite him for daring to step on the Warning Track That Ruth Built. He was observationally atrocious. Bay was not, at least not to me. (Anyone else?)

FWIW, I'm willing to buy Coco Crisp plummetting to the worst CF, as he was also observationally bad this year. He took bad routes, he got bad jumps, and nearly every risk he took - and he took a lot of them - didn't work. He earned his mark.

And congratulations, Gary Sheffield, for winning the DRS Gold Glove for DH with +1.68.
   59. RedSoxBaller Posted: October 21, 2008 at 01:39 PM (#2990775)
Bobby Areau was absolutely horrific in the field this year, Bay was decent, wouldn't say as bad as yuor system shows. Manny was not a better fielder just because he went to LA. That just proves that this OPD does not adjust for the Green Monster for fielders.
   60. Chris Dial Posted: October 21, 2008 at 01:40 PM (#2990776)
That just seems wrong to me. Jason Bay isn't going to be mistaken for Carl Crawford, but his total DRS of -21.3 for 2008 is the third worst in the league,
When Bay was acquired, I said that he wasn't going to be any good. He's always been a weak fielder IIRC.

Someone suggested an issue with Pittsburgh OF PF, but I haven't seen it too much in the numbers. A small effect, but not one that screams "PARK!"
   61. Chris Dial Posted: October 21, 2008 at 01:42 PM (#2990777)
I think Rollins will take his season over Reyes.
It's not about whose *team* had a better year. It's about which player is better. If the Phils had had Reyes instead of Rollins, they wouldn't have to rely on another team taking a nosedive in the last three weeks the last two years.
   62. Chris Dial Posted: October 21, 2008 at 01:43 PM (#2990778)
That just proves that this OPD does not adjust for the Green Monster for fielders.
That and the author's statement that says it doesn't.
   63. Chris Dial Posted: October 21, 2008 at 01:45 PM (#2990782)
I think Rollins will take his season over Reyes.
For that matter, I'm sure Kyle Kendrick is happy is team is where it is, but it would surprise me to hear Phils fans claim he's better than Johan Santana. Well, it may not surprise me.
   64. Mike Emeigh Posted: October 21, 2008 at 02:27 PM (#2990816)
The issue with defensive ratings is this: The ratings are based on the assumption that an average defender would post a 0.0 regardless of the specific situation into which he was dropped. We have no way to verify that this is true, however - and in the specific case of Boston LF, it is quite likely that it is NOT true, because of the effect of the Monster. To figure out what an average defensive LF *would* post in Boston, we'd have to subtract out wall balls first - zone stats do include wall balls because they are *in zone* based on depth, but they are not catchable by anyone - and then figure out what the impact of the fielder playing closer to the infield would be on the balls that weren't wall balls - there are some areas of the field that are *in zone* when the fielder can play a normal LF depth that wouldn't be *in zone* in Boston because the fielder can't play a normal LF depth because of the wall.

Ideally, what I'd like to see is an adjustment for "degree of difficulty", based on the actual in-play distribution against a team and a refit of the zone responsibilities based on field characteristics (the latter would be primarily an OF effect, with some small infield effects due to differences in the amount of foul territory if we include popups, which some people do and some people don't). BIP distributions vary enough from team to team so that I think the underlying assumption that an average defender would post a 0.0 regardless of situation is inappropriate. Groundball pitchers do (to some extent) make it easier on their infielders and harder on their outfielders, and the converse is true for flyball pitchers. If Fenway LF is harder to play than a normal LF, we should try to account for that.

-- MWE
   65. Chris Dial Posted: October 21, 2008 at 02:40 PM (#2990828)
Mike,
much of that is true. I have made some efforts to create a "wall-ball" spreadsheet. You and I have disagreements about how much BIP matters. I do think the zone is workable, in that the difference between a zone-based one and point-to-point analysis makes very small differerences. Yes, there are some advances to be made wrt chance difficulties (In an older version of this I made an effort to account for pitching staff stinkitudes). However, I also think that 20 years of data can do a good job of smoothing of the "wall-balls". It most certainly does wrt number of chances a fielder gets - the real versus projected data bears that out. I am reasonably close to creating park factors (which you will get to vet, of course).

As such, I think we can create a PF for Fenway that accounts for "wall-balls", and I suspect the depth of fielder is also accounted for. Yes, Bay will suffer more at first in Fenway (until he learns how and where to play), but so will his team. We're measuring plays made here, not Bay's "true talent".

I think we need an underlying assumption, and "average" is 0.0. If you can proffer a different baseline, I'm happy to hear it.
   66. rfloh Posted: October 21, 2008 at 02:48 PM (#2990839)
Mefisto Posted: October 20, 2008 at 09:42 PM (#2990562)
Chris, is there a database you could use for baserunning (non-SB)?


Asides from Dewan, BPro have baserunning numbers using Dan Fox's methodology.
   67. villageidiom Posted: October 21, 2008 at 05:07 PM (#2991010)
When Bay was acquired, I said that he wasn't going to be any good.
Was that based more on observation of his defense, or on observation of your metric? If the latter, then congratulations, you accurately predicted the behavior of your own metric, which isn't what I'm asking. If the former, then you should have no trouble telling me what it is about his defensive play that produces the answer your metric provides.

I'm eager to learn.
   68. fra paolo Posted: October 21, 2008 at 05:45 PM (#2991078)
Should these ratings sum to zero? Because they don't if you add up the team numbers. Overall they are negative. So does that represent the value of true outcomes?
   69. SuperGrover Posted: October 21, 2008 at 06:20 PM (#2991128)
Chipper isn't even appearing on the MSM MVP ballots which is going to be the real travesty.


He missed nearly a quarter of the season with injury. Given that there are other legit candidates his absence is not that surprising.
   70. Chris Dial Posted: October 21, 2008 at 06:26 PM (#2991139)
Should these ratings sum to zero? Because they don't if you add up the team numbers. Overall they are negative. So does that represent the value of true outcomes?
they don't add up to zero there because of positional assignments. I assure you they add up to zero.
   71. Chris Dial Posted: October 21, 2008 at 06:26 PM (#2991140)
that is the first thing I check when I create the spreadsheets.
   72. Chris Dial Posted: October 21, 2008 at 06:27 PM (#2991142)
He missed nearly a quarter of the season with injury. Given that there are other legit candidates his absence is not that surprising.
They have 10 slots. He certainly should be on ballots.
   73. Chris Dial Posted: October 21, 2008 at 06:29 PM (#2991147)
Was that based more on observation of his defense, or on observation of your metric? If the latter, then congratulations, you accurately predicted the behavior of your own metric, which isn't what I'm asking. If the former, then you should have no trouble telling me what it is about his defensive play that produces the answer your metric provides.

I'm eager to learn.
That reads oddly. The answer is both. If he isn't good at converting flyballs into outs in Pittsburgh, he isn't likely to be any good at it in Boston. From observation, I think he's a slow reactor, and he's kind of slow in the OF.
   74. Chris Dial Posted: October 21, 2008 at 06:29 PM (#2991149)
Oh, and I think he plays too deep, which people who aren't good fielders tend to do so they won't get burned deep. they then give up more balls in front of them than they should.
   75. EnderCN Posted: October 21, 2008 at 07:23 PM (#2991218)
I'm a bit confused by JJ Hardy defensively in general. According to thehardballtimes he led qualified NL SS in OOZ plays. He didn't make a ton of errors and he is generally considered by fans to be an average to above average SS. Looking at Justin's numbers over at http://jinaz-reds.blogspot.com/ he comes out above average and I've seen other defensive metrics list him as above average.

Why does your defensive system think so little of him, any idea what is different about them? Is it just that he stinks at in zone plays and your system weights that differently than others?

The same thing happened with Hardy last year if I remember correctly, half the systems thought he stunk and the other half thought he was above average.
   76. Chris Dial Posted: October 21, 2008 at 07:26 PM (#2991224)
Ender,
good questions. When I watch Hardy, he doesn't look good to me. BIS zones are different and evidently smaller, but when I asked John about that he said he wasn't sure that all of the BIS zones were within the ZR zones. I don't know.

The possibility exists that he gets few line drives, but that shouldn't amount to more than about 3 runs if he is REALLY getting zero. So, I don't know.

When he gets OOZ plays in BIS, those could be regular plays in ZR.
   77. Best Dressed Chicken in Town Posted: October 21, 2008 at 07:59 PM (#2991251)
Is the table ever going to be made legible?
   78. Mefisto Posted: October 21, 2008 at 07:59 PM (#2991252)
Asides from Dewan, BPro have baserunning numbers using Dan Fox's methodology.


Thanks for that reference; I hadn't seen those before.

They show Sizemore at +5.1 and Mauer at +4.1, so good call by Eric and Will.
   79. Chris Dial Posted: October 21, 2008 at 08:07 PM (#2991262)
Is the table ever going to be made legible?
can you not access the link? It's just a cut and paste from there. I'll speak to the Powers that Be. I cannot fix it - I have tried.
   80. villageidiom Posted: October 21, 2008 at 08:15 PM (#2991266)
That reads oddly. The answer is both. If he isn't good at converting flyballs into outs in Pittsburgh, he isn't likely to be any good at it in Boston. From observation, I think he's a slow reactor, and he's kind of slow in the OF.

...

Oh, and I think he plays too deep, which people who aren't good fielders tend to do so they won't get burned deep. they then give up more balls in front of them than they should.
Thanks, that was more of what I was looking for.
   81. Dan Posted: October 22, 2008 at 03:21 PM (#2991677)
I have never park-adjusted any defensive stats formulaically. I have tried to always write a caveat for Fenway's LF. I have recently backed into a park factor solution that I will bounce off some trusted statheads, and then write up, and then retro-actively apply to the larger database. But that's not a fast process. Joe Arthur (Google him) has done some great work for a Fenway park factor. My research out of the gate (very ROUGH DRAFT) looks like 0.84 effect FOR LF ONLY. So, the LF for Boston would see his ZR improve by dividing by 0.92. The largest problem with defensive PFs is going to be that the effect of teh Green Monster (and the Crawford Boxes) isn't consistent. It depends on how shittily the Sox pitchers pitch. Hopefully, I can get enough data to smooth those out.

After looking at your spreadsheet quite a bit, and your midseason updates, I have a strong suspicion that the Stats data for Fenway is absolutely junk for all 3 OF spots. If I am to believe DRS, then Boston's total defense was worth -34.51 runs, mostly on the strength of strong negative values for anyone who played the OF for the team at all. Something is going on in regards to Stats in Fenway or the positioning the Sox use or something, By THT's stats, the Red Sox had a .700 DER in 2008, good for 4th in the AL vs a league average of .691 (The Rays led the AL with a .712, for perspective). After park adjusting (mostly because of the LF wall), you're probably looking at 3rd or 2nd for the Red Sox in the AL (I saw park adjusted DER somewhere, but I can't recall where or find it at the moment). To get a .700 DER with the OF defense that DRS claims, the Red Sox would either have to be insanely lucky on balls in play or have pitchers that are amazingly good at inducing weak contact.

I'd be interested to see what the correlation is between DER and DRS by team.
   82. Xander Posted: October 22, 2008 at 03:27 PM (#2991684)
The great Justin Morneau appears to be absent.
   83. DKDC Posted: October 22, 2008 at 04:09 PM (#2991729)
To get a .700 DER with the OF defense that DRS claims, the Red Sox would either have to be insanely lucky on balls in play or have pitchers that are amazingly good at inducing weak contact.


Or they could be preventing singles at the expense of allowing extra base hits.
   84. Chris Dial Posted: October 22, 2008 at 08:31 PM (#2992049)
I'd be interested to see what the correlation is between DER and DRS by team.
My initial work indicates it isn't very strong. BIP distribution is HUGE here.
   85. Ron Johnson Posted: October 23, 2008 at 03:44 PM (#2993023)
Part of the mystery of Boston's .700 der comes from the fact that Boston pitchers were good at inducing popups and the like. These are "automatic" 1-1 in der terms and not considered by ZR. Boston pitchers were also better than average at avoiding line drives (though it's not a huge deal). Line drives are very nearly automatic 0-1 in der.
   86. EnderCN Posted: November 01, 2008 at 03:34 PM (#3000947)
Gonna bring up my Hardy question again since more numbers are out.

The fielding bible awards listed him as the #2 SS defensively.
plus/minus lists him #3 with +19


So obviously there is a fundamental differece between the various systems that seems to choke on JJ Hardy. From watching him he seems to make the routine plays but never seems to make a great play to me at least so I'm not saying your rating is wrong at all. I'm just trying to get a grasp on what the differences are that lead to such a huge difference on a single player. Most of the time the various sytems seem to agree on things but this is the 2nd year in a row that they were polar opposites on Hardy.

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Jolly Old St. Nick Still Gags in October
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 0.8605 seconds
68 querie(s) executed