Baseball for the Thinking Fan

Login | Register | Feedback

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

Thursday, August 07, 2008

NL Defensive Stats to date

Sure, everyone can go to BPro everyday and check out Albert Pujols’ VORP.  Or whatever your favorite site is and its Offensive Stat du jour.Well, if you really want to know how good a player is/has been you need to add his defense to that.  Is Manny going to be okay in Chavez Ravine?  Does David Wright deserve a second Gold Glove?  What you need though is a regularly updated place where you can see how a player is doing the moment you are having that argument or writing that blog piece.  Well, friends and neighbors, thanks to several other Primates (SG, AROM, MCoA), I got off my duff and made a more updatable and publishable format.  I also learned how to more efficiently use Excel to I could vlookup stuff and greatly cut down on the time it takes me.  Coupled with learning from MCoA’s use of Google docs, I can post them without making a giant mess of this page. 

So, like the All-Star Break data, I will post the data via a Google docs.  If someone can help me locate a good offensive stat format that includes all the pertinent data for my offensive calculations, I can update that more often too.  the present problem is that B-R’s “good page” doesn’t have GIDPs or something I need.  I use Doug’s Stats, but he has the positions all screwy, and correcting those takes 4-eva.  So, in the meantime, use VORP (which is my favorite anyway), or something easy for you to find.

Here is the link to NL DRS through August 6, 2008.  Here is a summary bit of information about that data, with Defensive Runs Saved (DRS) in parentheses.

First Base: Mark Teixeira (+11) had a good year in his 100 games with the Braves, but has moved to the AL.  Albert Pujols (+9) is nipping at his heels, and will certainly deserve the Gold Glove.  I watched a bit more of Pujols this year defensively, and I think that DRS may underrate him *slightly* because of his beautiful skill to take a grounder at first and get the runner at second base.  Most first basemen will take the runner at first rather than risk the throw, but Albert makes it smoothly, and that has slightly higher run value.  Bringing up the rear is Mike Jacobs (-9) in Florida.  The Marlins defense could be what costs them the division/wild card.

Second Base: Chase Utley has been the best fielding second baseman the last few years in the NL,a nd usually by a wide margin.  He has a slim lead this year over Adam “He Can’t Hit” Kennedy, but Kennedy has many fewer innings.  Brandon Phillips (+8) deserves an honorable mention.  Damion Easley (-8) has only gotten in half of the season, but has played poor enough to be the worst.  The Mets right side is not very good so far.

Third Base: David Wright (-1) is playing about the same as last year when he won the GG.  Chipper Jones (+10) is playing much better than last year, but is missing some innings.  Pedro Feliz (+9), who was terrific last year, is closing on Chipper.  Everyone should know who the worst fielding thrid baseman is - he’s staggeringly bad every year - Jorge Cantu (-11) of the Marlins.

Shortstop: Old Man River just keeps on rolling along. Omar Vizquel (+8) has only played 450 innings, but has played them very well.  Jimmy Rollins (+6) is closer to full-time.  Similarly, at the bottom of the rankings, Jeff Keppinger (-11) is a third or second baseman.

Left Field: Patt Burrell (+5), Juan Pierre (+5) and Fred Lewis (+5) are all neck and neck in left.  Burrell has 11 assists, so I’d give him the nod.  Really, though, I think this is the year where it is good that the outfield doesn’t require a fielder from each position win a Gold Glove.  One thing that is a bit more problematic is the LF park effect in Houston.  Carlos Lee (-20) is getting Manny-ish numbers out there.  The Pirates former LF, Jason Bay (-10), isn’t going to do well in Boston defensively.  You thought Manny was bad?

Center Field: The Padres have some pretty good defense all around, and the top centerfielder has been Jody Gerut (+8).  Now, again, there is an innings issue.  Gerut has 500 to Carlos Beltran’s (+6) 1000, but if Gerut maintains his pace for the next couple of months, he could earn the top spot.  One of the surprise hitters of the year, Nate McLouth (-15), has been absolutely terribly with the glove.  McLouth may be better suited to a corner position.

Right Field: Cub fan favorite, Jeremy Hermida (+9), is leading the way.  Right on his heels is Brian Giles (+8) of the Padres.  Giles has always been a very good fielder, and is aging pretty well.  It took this long for a Colorado Rockie to show up now that the humidor has been in full effect.  Brad Hawpe (-15) needs to move to another position, I think.

I would be remiss if I didn’t bring up Endy Chavez.  He splits time at all three outfield positions, and has saved more runs than any other outfielder with +11.6 DRS.

Pitchers and catchers are in that spreadsheet too, but most pitchers have very few chances, and the league conversion rate is higher than 97%.  Catchers have even fewer, and convert at 95%.  Most players make all the plays, and the players at the bottom miss two or three of their chances. 

The best defensive team thus far has been the Phillies (+34), and they are above average at every position, which is impressive.  The Phils division lead can be summed up in this, as the Mets (-12) and the Marlins (-20) are well behind them.  The Cubs (+1) are about average across the board, and the Cardinals (+21) are playing well.  The Diamondbacks (-21) are a poor-fielding team, and the Dodgers (+20) are keeping it close with the leather.  The worst fielding team is the Pirates (-30).

there you have a rundown of how the NL is doing.  You can sort it any number of ways with the data provided, or just check out how your favorite player is doing.  AL tomorrow.

Chris Dial Posted: August 07, 2008 at 01:18 PM | 97 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. Frisco Cali Posted: August 07, 2008 at 03:07 PM (#2893989)
Has Austin Kearns lost his elite fielding status - to go along with his loss of power?

Will he still be in the majors when he is 31 (he's 28 now)
   2. Cowboy Popup Posted: August 07, 2008 at 03:12 PM (#2893992)
Reyes really is having a Jeteresque season.
   3. Stevens Posted: August 07, 2008 at 03:13 PM (#2893993)
Geovanny Soto is the worst defensive catcher in the league? Wow.
   4. Bring Me the Head of Alfredo Griffin (Vlad) Posted: August 07, 2008 at 03:22 PM (#2894003)
"The Pirates former LF, Jason Bay (-10), isn’t going to do well in Boston defensively. You thought Manny was bad?"

I'm sorry, but this just doesn't pass the smell test for me.
   5. Russ Posted: August 07, 2008 at 03:22 PM (#2894004)
Is the Pirates' fielding catastrophically bad because the pitching is so horrible or is the pitching horrible because the fielding is so bad? Or are both so horrible that the interaction of the two makes them the defensive equivalent of the Lovely Ladies?
   6. Bring Me the Head of Alfredo Griffin (Vlad) Posted: August 07, 2008 at 03:26 PM (#2894008)
It seems like Pirate LFs always grade out too low on PBP metrics, while Pirate RFs always grade out too high. I wonder whether it's a park thing?

I'm also surprised that Adam LaRoche is a -3 (seems low) and that Bautista is a -4 (seems high - yuck!).
   7. Bring Me the Head of Alfredo Griffin (Vlad) Posted: August 07, 2008 at 03:31 PM (#2894015)
"Or are both so horrible that the interaction of the two makes them the defensive equivalent of the Lovely Ladies?"

I'm voting for this one. A fair number of the pitchers are pretty hittable, so lots of LDs, and the lineup only has one real plus fielder (Jack), who's drowned out by guys who aren't so hot.

It'd look a lot better with McCutchen in CF and Nate in RF next year.
   8. Padraic Posted: August 07, 2008 at 03:35 PM (#2894020)
Pat Burrell - Gold Glove Winner.

I really appreciate the work that goes into these stats (and their availability), but I won't be mentioning this to friends when arguing about the value of defensive statistics. I think his numbers do reflect the fact that he has turned himself into a good fielder in terms of positioning, routes, and accurate throws, but it's just not possible that he is the best defensive LF in the NL.

Burrell was -50 in Dewan's +/- from '05 to '07, so I would say there is a flaw in the data or an abnormally high number of easy chances, rather than a drastic improvement in talent level.
   9. Edmundo got dem ol' Kozma blues again mama Posted: August 07, 2008 at 03:35 PM (#2894021)
The best defensive team thus far has been the Phillies (+34), and they are above average at every position
This simply can not be true. 5 of the 8 positions are no-brainers -- Ruiz/Coste, Utley, Feliz, Victorino and Jenkins/Werth, and while I think that Rollins has slipped some range-wise, it's not inconceivable that he is a +. I can almost believe that Burrell could be a +, since he does everything right IF he can get to a ball. His range is massively bad, of course.
But RYAN HOWARD is above average? The statue-ranged, bad decision-making, big number of errors, Ryan Howard? The Ryan Howard who can not throw out any picked off runner out at 2B? (I haven't seen any, he may have gotten 1 or 2 runners this year). The only + I can see for Howard is that he can leap, believe it or not and has made of couple of really nice line drive stabs. The only way for him to be a + is that he is the absolute master of positioning, transcendent over all other 1Bs. Talk about not passing the smell test!
   10. Famous Original Joe C Posted: August 07, 2008 at 03:36 PM (#2894023)
"The Pirates former LF, Jason Bay (-10), isn’t going to do well in Boston defensively. You thought Manny was bad?"

I'm sorry, but this just doesn't pass the smell test for me.


I'm no expert myself, but I'm not going to say Jason Bay is a lousy defender because one metric shows himt o be lousy for a half season. Not to disparage the work that Chris and others have done, because it's certainly a step in the right direction and I think it's very good, but all of these numbers should be taken with many grains of salt.
   11. Edmundo got dem ol' Kozma blues again mama Posted: August 07, 2008 at 03:37 PM (#2894024)
And Chris, as Padraic said, I also appreciate the work that goes into these stats.
   12. Chris Needham Posted: August 07, 2008 at 03:38 PM (#2894025)
Has Austin Kearns lost his elite fielding status - to go along with his loss of power?

Kearns is right there on the defensive rankings. He's missed a bunch of time, otherwise he'd probably be #2.
   13. BFFB Posted: August 07, 2008 at 03:39 PM (#2894026)
Is Orlando Hudson really that bad?
   14. Padraic Posted: August 07, 2008 at 03:52 PM (#2894037)
The only + I can see for Howard is that he can leap, believe it or not and has made of couple of really nice line drive stabs.

I think defensive stats actually help capture things like this, so maybe Howard's height makes up for the lack of range, especially since so many defensive plays at 1B are made without actually taking a step.

On the other side of the diamond, having watched Abraham Nunez and Pedro Feliz play, I understand now why Nunez (despite reports) always had pretty bad defensive ratings, and Feliz (despite being a converted OF) always had high ratings. It's height! Feliz (6'1") has great hands and uncanny accuracy, but he also ranks so high because just by jumping and falling to either side, he covers way more ground then someone like Nunez (listed at 5'11" but there's no way). Howard likely makes a lot of plays other 1B cannot make because he's 6'4". This isn't defensive talent in the traditional sense, but it should be counted.
   15. Crispix reaches boiling point with lackluster play Posted: August 07, 2008 at 03:58 PM (#2894040)
Aha, the Yanks got Richie Sexson as a defensive replacement for Giambi!
   16. Xander Posted: August 07, 2008 at 04:04 PM (#2894047)
Is this a ZR-based metric?

The greatest trick the sabermetrician ever pulled was convincing the world that ZR-based metrics had any merit.
   17. Crispix reaches boiling point with lackluster play Posted: August 07, 2008 at 04:06 PM (#2894051)
I think it's a bit harsh to imply that there is only one sabermetrician in the world, and that he is the Devil.
   18. Toolsy McClutch Posted: August 07, 2008 at 04:10 PM (#2894058)
STL Pujols, Albert 2B 3.3 0

!
   19. Harris Posted: August 07, 2008 at 04:21 PM (#2894073)
Ryan Howard is apparently so bad at 1B that he didn't even make the list. He is listed at DH however, which is ironically kind of funny.
   20. Edmundo got dem ol' Kozma blues again mama Posted: August 07, 2008 at 04:24 PM (#2894078)
so maybe Howard's height makes up for the lack of range
Interesting point -- I'll have to watch and see what "height" plays Howard makes.
I have noticed that Rollins just can't quite get to an occasional line drive because of his height.
   21. Dan Posted: August 07, 2008 at 04:24 PM (#2894079)
Ryan Howard is apparently so bad at 1B that he didn't even make the list

Try looking again, closer to the top.
   22. Harris Posted: August 07, 2008 at 04:26 PM (#2894080)
Burrell leads LF with a +5 (better than Endy Chavez who so often gets touted as such an awesome OF).

Burrell's late inning defensive replacements:
So Taguchi: -0.9
Eric Bruntlett: 0.1
Jayson Werth: -0.3
Greg Dobbs: -0.3

Stupid Chollie.
   23. Crispix reaches boiling point with lackluster play Posted: August 07, 2008 at 04:26 PM (#2894081)
Howard's defense was good by Zone Rating three weeks ago. This may support TempleUSox's claim.
   24. Edmundo got dem ol' Kozma blues again mama Posted: August 07, 2008 at 04:27 PM (#2894082)
Ryan Howard is apparently so bad at 1B that he didn't even make the list.
You didn't start high enough in the list -- Howard is 4th at 1B!
   25. Harris Posted: August 07, 2008 at 04:28 PM (#2894084)
Try looking again, closer to the top.
Just saw it. That indicates that this metric is crap. Sorry.
I went to the negatives for first base and didn't see him so I put a search field in and he popped up as DH. Didn't consider the possibility that he was in the top 5. Because he isn't.
   26. Chris Dial Posted: August 07, 2008 at 05:15 PM (#2894131)
Geovanny Soto is the worst defensive catcher in the league? Wow.
Well, he's only got 9 chances (or thereabouts). That's why the disclaimer on catchers and pitchers. One miss, and it's curtains for you!
   27. Chris Dial Posted: August 07, 2008 at 05:16 PM (#2894134)
Howard's defense was good by Zone Rating three weeks ago. This may support TempleUSox's claim.
And now? I don't think this supports Templesox at all.
   28. Chris Dial Posted: August 07, 2008 at 05:18 PM (#2894139)
The greatest trick the sabermetrician ever pulled was convincing the world that ZR-based metrics had any merit.
Thank you. Do you have any particular reason for thinking it isn't worthwhile?
   29. Chris Dial Posted: August 07, 2008 at 05:19 PM (#2894146)
Burrell leads LF with a +5 (better than Endy Chavez who so often gets touted as such an awesome OF).
Burrell is +5.4 in 850 innings. Endy is +4 in *155* innings. Endy would be about +20 in Burrell's innings.
   30. Padraic Posted: August 07, 2008 at 05:20 PM (#2894149)
Just saw it. That indicates that this metric is crap. Sorry.

Yes, your inability to read a spreadsheet indicates that a metric is flawed. I have problems with some of these ratings - and particularly the confidence with which they are presented - but there's no need to be a jerk.
   31. Chris Dial Posted: August 07, 2008 at 05:20 PM (#2894150)
This simply can not be true.
The reason it is true is that it is the sum of all the players at the position, not just the starters.
   32. Chris Dial Posted: August 07, 2008 at 05:22 PM (#2894152)
and particularly the confidence with which they are presented
Pardon? There's lots of background on this data within "Dialed In", laying out issues with the data and accuracy. Re-posting those things 100% of the time within the same area is, well, IMO, not needed.
   33. Chris Dial Posted: August 07, 2008 at 05:23 PM (#2894154)
but there's no need to be a jerk.
he knows me.
   34. Edmundo got dem ol' Kozma blues again mama Posted: August 07, 2008 at 05:23 PM (#2894155)
Chris, I've never heard anyone state that they think Howard is any better than a "bad" fielder. Even the Phillies announcers will state that "he is not one of the better 1Bmen" or some such stuff. Do you really think height (and it's not that Howard is 7'2" or anything) could really make that much difference at first? Any kind of positive rating is so counter-intuitive.
   35. Chris Dial Posted: August 07, 2008 at 05:31 PM (#2894172)
Do you really think height (and it's not that Howard is 7'2" or anything) could really make that much difference at first? Any kind of positive rating is so counter-intuitive.
No, I think that 1B is a tough enough position that the difference between Howard (+4) and, say, Carlos Delgado (-7) is about 12 plys (in Howard's innings). 12 GBs where it didn't deflect off his glove, or the pitcher properly covered, or one play a week where he got an easier play.

At a half season of innings, you can easily get these anomalies. Howard just needs a bad week to be -3. It's a little volatile when you don't have enough innings played. MGL tried to throw a few million disclaimers in when he posted his mid-season scores, but people fixate on it too much.

This is also a *relative* ranking, so as other players play better, Howard will drop some. But he could also have a good season with the glove, just like a player can have a good season with the bat.
   36. Gaelan Posted: August 07, 2008 at 05:33 PM (#2894176)
I'm sorry, but this just doesn't pass the smell test for me.


Bay was horrible last year too. I think these measures have a lot of merit but you have to look at them over the long haul. The one thing we have to remember, however, is that while the numbers may go up and down, generally speaking, the play of the player does not. Injuries aside players do not have good and bad seasons and Pat Burrell is still a horrible defender.
   37. _ Posted: August 07, 2008 at 05:35 PM (#2894181)
Are SB/CS part of the catcher numbers?
   38. Chris Dial Posted: August 07, 2008 at 05:53 PM (#2894208)
Are SB/CS part of the catcher numbers?
No, these are not real catcher ratings. Those I do seperately. I guess I shouldn't have left them in. 36 of the 49 catchers have perfect ZRs, with Brian McCann leading the way at 26 for 26. Soto is a mere 6 of 9. I don't know why he doesn't take more balls though. That's very low for catchers and that many innings. Kendall is 33 of 36.
   39. _ Posted: August 07, 2008 at 06:25 PM (#2894260)
Soto is a mere 6 of 9. I don't know why he doesn't take more balls though.

Well, Cubs pitchers do lead the majors in Ks.
   40. Mike Green Posted: August 07, 2008 at 06:26 PM (#2894264)
Gerut has been a mediocre corner outfielder in prior years. I'd be shocked if the metric reflects his actual performance in center-field at age 30. Whether it is due to park effects or sample size, something is going on.
   41. Kyle S at work Posted: August 07, 2008 at 06:33 PM (#2894279)
It seems like Cubs pitchers always lead the majors in Ks, even in years where they don't seem like they should. Is there a Wrigley park effect, or are they just good at always having guys who get lots of strikeouts? Trading for Rich Harden didn't hurt, I know, but before that move I wouldn't have called the Cubbies staff a bunch of strikeout artists.
   42. Edmundo got dem ol' Kozma blues again mama Posted: August 07, 2008 at 06:41 PM (#2894301)
No, I think that 1B is a tough enough position that the difference between Howard (+4) and, say, Carlos Delgado (-7) is about 12 plys (in Howard's innings). 12 GBs where it didn't deflect off his glove, or the pitcher properly covered, or one play a week where he got an easier play.
Thanks. Of course I'm just using my biased, glasses-enhanced eyeballs on an old analog TV. Another observation of mine is that the Phils have been doing a nice job on the pitcher covering first. Kendrick, Moyer, Eaton and Hamels are all good athletes, so maybe Howard's getting some bonus from them.
I think if you picked up botched pickoffs in your ratings, Howard's numbers would drop to the negative. :) He has been Little League bad.

At a half season of innings, you can easily get these anomalies. Howard just needs a bad <strike>week</strike>game to be -3. Fixed.
   43. AJMcCringleberry Posted: August 07, 2008 at 06:46 PM (#2894310)
Yikes. Easley and Castillo have combined for -14 in 950 innings.
   44. SouthSideRyan Posted: August 07, 2008 at 06:52 PM (#2894323)
Cub-centric: doesn't pass the smell test

Mike Fontenot as an above average 2B.

Ronny Cedeno as a below average 2B.

Aramis Ramirez as a below average 3B.

Jim Edmonds as not a terrible CF.

And I have to comment that Vizquel as the best SS seems like you have to be using old #s.
   45. _ Posted: August 07, 2008 at 06:57 PM (#2894338)
are they just good at always having guys who get lots of strikeouts?

They certainly have been since about the time Wood and Prior and Clement and Z were together. The 2003 team led the league by a wide margin. Rich Hill and Ted Lilly are no slouches either; and then there's Marmol.
   46. Padraic Posted: August 07, 2008 at 07:00 PM (#2894350)
Pardon? There's lots of background on this data within "Dialed In", laying out issues with the data and accuracy. Re-posting those things 100% of the time within the same area is, well, IMO, not needed.

I don't necessarily mean including methodology, just saying something like "yeah, Burrell is at +5, but his recent numbers have all been poor, so this might not be accurate" rather than just stating that Burrell "gets the nod" because he's tied at +5. I think these numbers work well as a nice starting point for a discussion, but they are presented as facts, i.e "Bay will be bad in Boston" or Hawpe "needs to move positions." To make claims like these based on one metric is the overconfidence I was talking about. Anyway, it's good work.
   47. The importance of being Ernest Riles Posted: August 07, 2008 at 07:42 PM (#2894448)
I'm sorry, but this just doesn't pass the smell test for me.

but it's just not possible that he is the best defensive LF in the NL.

This simply can not be true.

Is Orlando Hudson really that bad?

That indicates that this metric is crap. Sorry.

Any kind of positive rating is so counter-intuitive.

Cub-centric: doesn't pass the smell test


Fascinating, no?

As with offensive stats, defensive stats require regression to the mean. What Chris is presenting is similar to looking at VORP or OPS or batting average. It tells you what a player HAS done, not what his true talent is. It's a value judgment, not a talent judgment. It seems that a lot of Primates are forgetting this. Just as it's perfectly possible for a bad hitter to look terrible yet compile good offensive stats (due to luck or whatever), it's possible for a bad fielder to look terrible yet compile good defensive stats (due to luck or whatever).

Since there are (generally) fewer defensive opportunities than plate appearances, the regression for a single season worth of defensive stats is (probably) even heavier than for offense, no matter how good the defensive metric is. That's why, even when we get to defensive stats that will track the speed off the ball off the bat, the reaction time of a fielder, etc., we still won't know very much about true talent defense.
   48. Slinger Francisco Barrios (Dr. Memory) Posted: August 07, 2008 at 07:55 PM (#2894469)
But he could also have a good season with the glove, just like a player can have a good season with the bat.

I hear this a lot, but I don't see why this is necessarily true if we can't point to a particular change (e.g., in health, in conditioning) one way or the other. Unless by "good" we understand that to mean in large part "lucky".
   49. cardsfanboy Posted: August 07, 2008 at 08:28 PM (#2894495)
I always find it funny to see how many people don't grasp the concept of these numbers just representing how well they have played to this date, and not a listing of their true talent level. The Burrell listing points to the problem with left field on the whole, nobody good defensively is playing in left, people who are judging players by their eyes are most likely comparing them to other players from their memory, not the average pool that is currently playing the game.


BTW, Congrats to Chris for being mentioned in Popular Science (if there was a thread on this I missed it)
   50. Padraic Posted: August 07, 2008 at 08:31 PM (#2894502)
It tells you what a player HAS done, not what his true talent is.

I agree with that, but I still don't think it can be stated with as much confidence as offensive statistics. For example, no matter what the true talent, we know that Ryan Ludwick has X more number of total bases than Geoff Jenkins. So even if we want to say Ludwick is lucky, his contributions still count.

But with defense, the 'play' that gets counted is an artifact of the zone or whatever is set up, in a way that a hit is not.* We can't really know that Pat Burrell is 25 runs better on defense than Carlos Lee in the same way we know Ludwick has been a better hitter than Jenkins.

*Well, you may be able to argue that a 'hit' is an artifact too, but an advanced base is not.

So even at the level of description of the contribution - as opposed to prediction or evaluating true talent - defensive metrics still are not as reliable.
   51. Gaelan Posted: August 07, 2008 at 08:40 PM (#2894517)
As with offensive stats, defensive stats require regression to the mean. What Chris is presenting is similar to looking at VORP or OPS or batting average. It tells you what a player HAS done, not what his true talent is.


This is totally wrong. Batting average tells you what a player has done. Defensive stats are estimates and guesses about what a player has done. In addition to the sample size variation that is common with things like batting average there is an unspecified measurement error that is unique. So you are completely mistaken if you think these numbers represent what a player has done in the same sense that batting average is a measurement of what a player has done.
   52. Gaelan Posted: August 07, 2008 at 08:45 PM (#2894524)
I don't necessarily mean including methodology, just saying something like "yeah, Burrell is at +5, but his recent numbers have all been poor, so this might not be accurate" rather than just stating that Burrell "gets the nod" because he's tied at +5. I think these numbers work well as a nice starting point for a discussion, but they are presented as facts, i.e "Bay will be bad in Boston" or Hawpe "needs to move positions." To make claims like these based on one metric is the overconfidence I was talking about.


I completely agree with this. Unfortunately Dial does not and hence says things like players can have good seasons with the glove just like they can have good seasons with the bat. There is no reason to believe this is true and plenty of reasons to think it is not. This is a common source of sabermetric error. Saberists do not recognize that players "true" defensive ability does not vary very much while players "true" offensive ability varies a great deal. This leads people to conclude that Jeter's "good" defensive season is legitimate while Ludwick is just lucky when most likely the exact opposite is true.
   53. The importance of being Ernest Riles Posted: August 07, 2008 at 08:46 PM (#2894526)
there is an unspecified measurement error that is unique.

Okay, I stand corrected - I neglected to mention measurement error. Right on.

I guess what bugs me is that the "doesn't pass the sniff test" comments are usually shots at measurement error, when they could very well be due to sample size as well. Didn't say it correctly, though.
   54. Famous Original Joe C Posted: August 07, 2008 at 08:47 PM (#2894527)
Since there are (generally) fewer defensive opportunities than plate appearances, the regression for a single season worth of defensive stats is (probably) even heavier than for offense

It's almost certainly heavier, and should pribably be *much* heavier. Besides number of opportunities, we can measure offensive contributions with a fairly high degree of accruacy because their outcomes are more independent and discreet - there's some luck built in (e.g. "hitting 'em where they ain't" and park effects and all), but it's mostly just pitcher vs. hitter, a single event with a single outcome - there are so many interactions within the field - player positioning, pitcher handedness, so on and so forth, that can cloud the numbers, no matter how well you try to adjust for them.

I don't see at this point how you can use anything other than several metrics averaged together, or multiple year-regressions of individual metrics, amd even then, that you can draw conclusions much beyond "Excellent/Above Average/Average/Below Average/Lousy". Arguing a guy is a better defender than another because one is +8 and one is -1 looking at 4 months of data on a single metric is pretty silly - no offense. Again, I realize that Chris is just providing these for the fun of it (oh, and thanks, Chris, ofr posting these!), and they are fun to look at, but they can't be used as is.
   55. Harris Posted: August 07, 2008 at 08:49 PM (#2894529)
Yes, your inability to read a spreadsheet indicates that a metric is flawed.


but there's no need to be a jerk.


sorry to have offended you. In all seriousness, looking for howard at the top of a 1B fielding ranking is inefficient. That's why I went to the bottom and worked my way up. Guess I just didn't give it enough time. As the explanations of the metric are being presented, it makes more sense to me, but it tells me it shouldn't be used to indicate a person's fieleding ability. It tells how many chances they've had and what they've done with them, and chances don't appear to include balls they should've fielded but didn't get to.
   56. nick swisher hygiene Posted: August 07, 2008 at 08:51 PM (#2894534)
"The one thing we have to remember, however, is that while the numbers may go up and down, generally speaking, the play of the player does not."
--would somebody knowledgeable tell me how defensive stats differ from offensive stats in this respect?
(or if some dumbass wants to tell me, that'd be ok too I suppose...)
   57. Chris Dial Posted: August 07, 2008 at 08:52 PM (#2894536)
just saying something like "yeah, Burrell is at +5, but his recent numbers have all been poor, so this might not be accurate" rather than just stating that Burrell "gets the nod" because he's tied at +5. I think these numbers work well as a nice starting point for a discussion, but they are presented as facts, i.e "Bay will be bad in Boston" or Hawpe "needs to move positions." To make claims like these based on one metric is the overconfidence I was talking about. Anyway, it's good work.
I think that it's fair to make those comments. they are the beginning of the discussion you'd want to have. Plus I have a pretty good idea about players' history. I mean, I don't think those comments come off as overconfident, but okay.

And thanks - I want it to be as open as we can.

Bay, FWIW, will be bad in Boston.
   58. Cowboy Popup Posted: August 07, 2008 at 08:53 PM (#2894538)
Saberists do not recognize that players "true" defensive ability does not vary very much while players "true" offensive ability varies a great deal.

Ability and value are different things, good positioning can have such a great impact on a defenders ability to get to balls in play that you can't simply look at a guy's range and say he's been a valuable defender this year or he hasn't been. If a player is on a team with great positioning for a year, they'll likely turn in a pretty good year, even if they're Pat Burrell or Derek Jeter.
   59. Edmundo got dem ol' Kozma blues again mama Posted: August 07, 2008 at 08:55 PM (#2894539)
I always find it funny to see how many people don't grasp the concept of these numbers just representing how well they have played to this date, and not a listing of their true talent level. The Burrell listing points to the problem with left field on the whole, nobody good defensively is playing in left

Oh, I get it and even conceded I can almost believe that Burrell could be positive. Unless Ryan Howard is fielding like the reincarnation of Keith Hernandez for the 50% of plays that I haven't seen this year and fielded, well, like Ryan Howard the times I'm looking, and with what I've observed with his defensive play and the defensive play of other 1Bers, something quite unusual is going with the numbers. So far there have been three suggestions that push him towards the good side:
1. Height
2. Good pitcher coverage of first
3. Inordinate number of easy plays

Plus that the 2 inept features of Howard are not tracked:
4. Breaks for some balls towards the 2B that he can't make the play. Does Utley get dinged for not making those plays?
5. His utter inability to take a successful pickoff and throw it to 2nd in enough time to get a runner

As Chris pointed out, 12 plays between Howard and Delgado are all that stand between a positive and terrible rating, so maybe #3 is at play here with a couple of extra positive plays helped by #1 and #2. So it is probably reasonable to assume that Howard's +4 will start to diminish for the rest of the year. He helped that along with his 14th error today, and you guessed it, a throwing error.
   60. Slivers of Maranville descends into chaos (SdeB) Posted: August 07, 2008 at 08:58 PM (#2894543)

As with offensive stats, defensive stats require regression to the mean. What Chris is presenting is similar to looking at VORP or OPS or batting average. It tells you what a player HAS done, not what his true talent is. It's a value judgment, not a talent judgment. It seems that a lot of Primates are forgetting this.


The point about the Cubs smell test is that Aramis Ramirez has looked very good this year. Most Cubs fans here agree this may well be the best defensive season of his career so far.
   61. Chris Dial Posted: August 07, 2008 at 09:01 PM (#2894548)
I completely agree with this. Unfortunately Dial does not and hence says things like players can have good seasons with the glove just like they can have good seasons with the bat. There is no reason to believe this is true and plenty of reasons to think it is not.

Why do you say that? Jeter has had some "good" seasons with teh glove, despite me (and others) saying he's just having a good year, rather than saying he is good. Please offer "plenty of reasons to think it is not".

This is a common source of sabermetric error. Saberists do not recognize that players "true" defensive ability does not vary very much while players "true" offensive ability varies a great deal.
Please demonstrate why you think this is true. And really, what I am saying is that within the season, players can have good seasons and bad is precisely because of what matters to the player is his BIP distribution.

I do disagree that if Pat Burrell has caught FBs hit to LF at a higher rate than the average LF that he isn't really having a better year, and that it's "probably wrong". I don't think it is wrong. His BIP distribution may be slightly more favorable, or he happened to see the ball better or got a better jump on some balls. But the fact of the matter is that X balls were hit to his general area, and he caught Y percentage. It is a large area, and the average LF on a *similar* BIP distribution catches 7 fewer balls. I don't think that's wrong. Now it's *possible* that with more precise BIP vectors we could say that those 7 balls would also have been caught, and some other BIPs that Burrell didn't, but this metric is usually consistent with that one, so I don't think there is automatically some change that couldn't be his BIP distribution or his "luck" on line drives.
   62. Padraic Posted: August 07, 2008 at 09:02 PM (#2894549)
This leads people to conclude that Jeter's "good" defensive season is legitimate while Ludwick is just lucky when most likely the exact opposite is true.

Just to follow up, we also have lots of good ways to tell if an offensive season is luck (opponent pitchers, BABIP, park), but I don't see how that's done (yet) with defensive statistics.

sorry to have offended you.

Hey no problem, didn't realize you knew Chris Dial. I actually like this metric a lot - for one it helped me appreciate that Utley really is an excellent defender - so I thought the comment was a cheap shot. But if you know him, feel free to bash DRS in whatever tone you wish.
   63. Chris Dial Posted: August 07, 2008 at 09:16 PM (#2894564)
So, for those that don't think it is utter crap, is weekly (say on Mondays?) often enough to update to keep your blogs/arguemtns flowing?
   64. The Buddy Biancalana Hit Counter Posted: August 07, 2008 at 09:29 PM (#2894575)
So, for those that don't think it is utter crap, is weekly (say on Mondays?) often enough to update to keep your blogs/arguemtns flowing?

I would appreciate that very much.
   65. cardsfanboy Posted: August 07, 2008 at 10:25 PM (#2894634)
I would appreciate it too, but once a month would be ok also. I doubt that the numbers change that much (assuming this is a lot of work for you) if it is easy enough for you, then once a week would be great.
   66. Srul Itza Posted: August 08, 2008 at 12:01 AM (#2894699)
In all seriousness, looking for howard at the top of a 1B fielding ranking is inefficient. That's why I went to the bottom and worked my way up

Suggestion:

Ctrl-F; Howard; Enter

There is no reason to believe this is true and plenty of reasons to think it is not.

I would think the contrary is true. Baseball is both a talent sport and a skill sport. You can learn things about a position as you play it. You can learn better positioning, and gain practice at reading the ball coming off the bat. Some of it is a question of concentration, effort and practice. This will all be offset by the ravages of time.

Of course, some people never learn. But there is no reason to believe this is true of everyone.
   67. Srul Itza Posted: August 08, 2008 at 12:02 AM (#2894700)
Are there really enough chances in one week to make it worth the bother of updating?
   68. Chris Dial Posted: August 08, 2008 at 12:15 AM (#2894711)
Are there really enough chances in one week to make it worth the bother of updating?
Er, sort of. There are about 15-20. Depending how the league does, it could move a run or two. Last year, David Wright was reasonably close to deserving the Gold Glove. The last 17 games, he was abominable. You can dump three or four runs. And also, people just *like* to have the most recent data when they argue/blog. Your point is true, but it also makes it interesting when a player crashes and his team misses the playoffs barely - you can blame someone.
   69. Chris Dial Posted: August 08, 2008 at 12:15 AM (#2894712)
Maybe it is every two weeks.
   70. AROM Posted: August 08, 2008 at 12:20 AM (#2894713)
This is a common source of sabermetric error. Saberists do not recognize that players "true" defensive ability does not vary very much while players "true" offensive ability varies a great deal.


I don'y think this statement is true at all. I guess that makes me a saberist. What group do you belong to so I can also paint those who disagree with a broad brush?

Players can have good/ poor defensive seasons for the same reason as hitters: their relative health. When measures of a player's range vary it could be they are having a good/bad defensive season, or it could be ball distribution/measurement error. Errors are a little different though: Look at Chone Figgins. Last year he made 13 errors in 99 games. This year only 3 in 70 games. I don't know if his true ability has changed much, but that's not measurement error, that's simply a player having a much better season than the one before.
   71. Exploring Leftist Conservatism since 2008 (ark..) Posted: August 08, 2008 at 03:22 AM (#2894822)
For most positions it seems that about twenty runs separate the best from the worst. Is it unreasonable to then assert that a top defensive team will win something like 16-18 games more than a bad defensive team, all other things being equal?
   72. Gaelan Posted: August 08, 2008 at 03:38 AM (#2894826)
Players can have good/ poor defensive seasons for the same reason as hitters: their relative health. When measures of a player's range vary it could be they are having a good/bad defensive season, or it could be ball distribution/measurement error. Errors are a little different though: Look at Chone Figgins. Last year he made 13 errors in 99 games. This year only 3 in 70 games. I don't know if his true ability has changed much, but that's not measurement error, that's simply a player having a much better season than the one before.


I was this close to adding caveat concerning health and errors. Consider me guilty as charged. That being said I think hitters have good/bad seasons for reasons other than health. Hitting a baseball is a very complicated skill. Similar to golf if just one thing gets out of whack, either mentally or physically, then performance is going to fall apart. While the conventional wisdom (around here) is that slumps are mostly luck I think it is a player losing their swing for a little while.
   73. Ivan Grushenko of Hong Kong Posted: August 08, 2008 at 03:41 AM (#2894830)
Chris, I think it's awesome that you do this.
   74. Slinger Francisco Barrios (Dr. Memory) Posted: August 08, 2008 at 01:08 PM (#2894938)
And really, what I am saying is that within the season, players can have good seasons and bad is precisely because of what matters to the player is his BIP distribution.

Then instead of "good" and "bad" you should maybe say "lucky" and "unlucky" in more circumstances. It's when one says "good" and "bad" all the time that one comes off as perhaps assuming more knowledge than one actually has.
   75. Slinger Francisco Barrios (Dr. Memory) Posted: August 08, 2008 at 01:10 PM (#2894939)
For most positions it seems that about twenty runs separate the best from the worst. Is it unreasonable to then assert that a top defensive team will win something like 16-18 games more than a bad defensive team, all other things being equal?

I wouldn't think that any team would have the lion's share of good or bad defensive players; they will be a bit more unevenly distributed.
   76. Swedish Chef Posted: August 08, 2008 at 01:21 PM (#2894949)
For most positions it seems that about twenty runs separate the best from the worst. Is it unreasonable to then assert that a top defensive team will win something like 16-18 games more than a bad defensive team, all other things being equal?

That would be the equivalent of comparing a team with 9 Barry Bonds and one with 9 José Vidros. Such concentrations of talent/ineptness is never seen.

To go with the real numbers, the spread between Philadelpia (+34) and Marlins (-20), is much bigger than what separates them in the standings.
   77. Chris Dial Posted: August 09, 2008 at 12:58 AM (#2896104)
It's when one says "good" and "bad" all the time that one comes off as perhaps assuming more knowledge than one actually has.
Well, one shouldn't be assuming anything.

A season that is better than average is a good season. A season that is worse than average is a bad season. Saying a season is good or bad doesn't imply the underlying reasons. I don't think.
   78. Exploring Leftist Conservatism since 2008 (ark..) Posted: August 09, 2008 at 03:38 AM (#2896322)
That would be the equivalent of comparing a team with 9 Barry Bonds and one with 9 José Vidros. Such concentrations of talent/ineptness is never seen.


True, but Whitey's Runnin' Redbirds isn't too far from what I'm thinking of, and a lot of teams were notrious for carrying mostly stone gloves. It's the tradeoffs that interest me: the value of a super glove in terms of batting runs. Who's the fielding equivalent of a solidly above average player like Carlos Lee, and how much less do you have to pay Superglove than Carlos, and so on...
   79. SkyKing162 Posted: August 09, 2008 at 07:41 PM (#2896548)
Chris, where have you outlined your methodology? Are you using STATS ZR, BIS ZR, both or something else?

Justin updates his player values every two weeks, and his fielding ratings are an average of STATS and BIS ZRs, using your conversion methods.

http://spreadsheets.google.com/pub?key=pfk_WuYpfdux2FC_hs6ROEQ
   80. SkyKing162 Posted: August 09, 2008 at 07:51 PM (#2896552)
I actually know where to find your methodology, I'm really just asking about what source data you're using.
   81. Chris Dial Posted: August 09, 2008 at 10:31 PM (#2896658)
I am using stats ZR and chances from cnnsi. And which Justin? The reds?
   82. SkyKing162 Posted: August 10, 2008 at 05:32 PM (#2897375)
Any reason not to also use and post the BIS zone ratings from THT?

Yes, Justin from On the Reds. http://jinaz-reds.blogspot.com
   83. Chris Dial Posted: August 10, 2008 at 07:24 PM (#2897563)
Any reason not to also use and post the BIS zone ratings from THT?
Because they are posted there? Since DA/DR went away in 1995, I was the next person (published) to take ZR and convert it to runs. I use the same formula I developed in 1997. In 1999 (I think) STATS started doing it similarly with their UZR but to plays, not runs. MGL started doing it in about 2001 or 2002 after seeing my stuff at Fanhome (Tango and Patriot and he worked it over about then - maybe 1999-2000), and he could immediately improve on it because he could afford the raw data.

There's a reason RZR from BIS didn't come out until 2005-6. Plus, I'm none too sure how they define everything.

Basically, I've been doing it the same way for 11 years, and nothing has notably improved upon it with as open source. My objective is really that you don't need me at all. You can simply copy ZR and my methodology from the "Dr. Strangeglove" article, and you can produce numbers as good as anyone. You just can't get that anywhere. I think taking the two and averaging them is a terrific idea, but that's not me. JinAZ has that niche.
   84. SkyKing162 Posted: August 11, 2008 at 12:29 AM (#2897735)
Thanks, Chris. Didn't know if you had some reason to prefer just the STATS data or some other preference like that.

I wish I'd found your calculations earlier, they and your methods are way underrated.
   85. Chris Dial Posted: August 11, 2008 at 02:01 AM (#2897905)
I wish I'd found your calculations earlier, they and your methods are way underrated.
That's very kind. I posted my methodology - heck, I worked it out on the internet, specifically so it could be *our* ratings - where "our" is the public. AROM, JinAZ, SGinATL, and everyone else that uses it does me proud. THe idea is that no one should "own" the data. I developed a good working method (not necessarily the best or anything), but an open one. And truth be told, I owe lots of my method development to Ron Johnson and Dale Stephenson. I try to credit Dale as much as possible (he was mentioned in the latest Popular Science article about this method) as he and I exchanged lots of run value discussions, and Ron helped with the basic framework in asb-nymets in 1997-8.

I'm just glad people find the data I post useful, but I'm even more pleased when someone else takes it and makes their own spreadsheets - like hte guys mentioned.
   86.  Hey Gurl Posted: August 11, 2008 at 04:21 AM (#2898032)
Is it at all possible to break down this sort of thing into a Play-by-play sort of structure, to see which games the +'s and the -'s come from? For example, could you say "he made 4 plays out of 5 chances on June 5th" or whatever?

It would take a HELL of a lot of work, of course, but it would be interesting to look at this, and maybe cross-reference with MLB.tv to try and "explain" why Ryan Howard or, in previous years, Derek Jeter rates the way he does.

Thank you, as always, for the numbers, Chris.
   87. Chris Dial Posted: August 11, 2008 at 02:36 PM (#2898170)
For example, could you say "he made 4 plays out of 5 chances on June 5th" or whatever?

It would take a HELL of a lot of work, of course, but it would be interesting to look at this, and maybe cross-reference with MLB.tv to try and "explain" why Ryan Howard or, in previous years, Derek Jeter rates the way he does
Yes. the reason I haven't is that it would take a helluva lot of work. I just don't have the database skill sfor it, but may by next year. I think Joe Arthur or some others may.
   88. SkyKing162 Posted: August 11, 2008 at 11:25 PM (#2899041)
Is it at all possible to break down this sort of thing into a Play-by-play sort of structure, to see which games the +'s and the -'s come from? For example, could you say "he made 4 plays out of 5 chances on June 5th" or whatever?


I don't think it would be worth it, unless you had a system as advanced as UZR or PMR. On a day to day basis, the +/- numbers don't tell you much, because the difficulty of the balls hit into the zone could vary widely (all routine or all on the edge). Zone rating really only has value in the long run becaues the distribution of balls evens out.
   89. Chris Dial Posted: August 12, 2008 at 12:43 AM (#2899104)
I don't think it would be worth it, unless you had a system as advanced as UZR or PMR. On a day to day basis, the +/- numbers don't tell you much, because the difficulty of the balls hit into the zone could vary widely (all routine or all on the edge). Zone rating really only has value in the long run becaues the distribution of balls evens out.
Right, but you could do more Home/Away and park effects. Comparing players as they changed teams, etc.

I used to have a pitching staff multiplier, and maybe still should. Chicken/Egg stuff.
   90.  Hey Gurl Posted: August 12, 2008 at 01:05 AM (#2899133)
Right, but that's the point! Looking at it on a play-by-play basis, you can see if a player had more "easy chances" or if he keeps getting +'s due to positioning or whatever.

I have the "database skills," but not the database... :)
   91. Chris Dial Posted: August 12, 2008 at 01:33 AM (#2899174)
Right, but that's the point! Looking at it on a play-by-play basis, you can see if a player had more "easy chances" or if he keeps getting +'s due to positioning or whatever.

I have the "database skills," but not the database... :)
Well, I think you might be able to get that mining the raw data from MLB.com. I might can help you with that.

It's kind of the same as pulling pitch f/x data. Have you looked into that?
   92. Scoriano Flitcraft Posted: August 12, 2008 at 06:40 PM (#2899732)
I don't watch the Mets often, but when I do, Endy Chavez is a delight to watch. He's a master on defense.
   93. Barca Posted: August 13, 2008 at 10:16 AM (#2900874)
"David Wright (-1) is playing about the same as last year when he won the GG."

So he is a shoe-in for the award again this year.
   94. BobbyMac Posted: August 13, 2008 at 11:57 PM (#2902017)
Gerut has been a mediocre corner outfielder in prior years. I'd be shocked if the metric reflects his actual performance in center-field at age 30. Whether it is due to park effects or sample size, something is going on.


Gerut played most of his prior years with knee problems that he seems to have managed to put behind him, at least functionally. I do concur that best-range-in-league performance is unlikely to continue, though.
   95. Crispix reaches boiling point with lackluster play Posted: August 14, 2008 at 12:32 AM (#2902135)
"David Wright (-1) is playing about the same as last year when he won the GG."

So he is a shoe-in for the award again this year.


O Why must Pedro Feliz perpetually get the shaft?
   96. Chris Dial Posted: August 14, 2008 at 05:59 PM (#2903120)
O Why must Pedro Feliz perpetually get the shaft?
He's not so DOMINATE! This year. I'm disappointed in him somewhat.
   97. Edmundo got dem ol' Kozma blues again mama Posted: August 14, 2008 at 06:05 PM (#2903130)
He's not so DOMINATE! This year. I'm disappointed in him somewhat.
After the Dobbs/Nunez/Helms three-headed monster last year, we here in Philly have been falling all over ourselves about how great he is. I do like his throws across the diamond -- from the moment of release, every throw looks like it is going to hit the 1B mitt right where Howard puts it.

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
greenback calls it soccer
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Syndicate

Page rendered in 1.0970 seconds
68 querie(s) executed