Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Baseball Newsstand > Baseball Primer Newsblog > Discussion
Baseball Primer Newsblog
— The Best News Links from the Baseball Newsstand

Wednesday, August 08, 2012

THT: DuPaul: What is WAR good for?

Uhh…to make Edwin Encarnación and his 3.5 WAR into a starr?

This past weekend, I had the honor and pleasure to present a paper at the 2012 Saber Seminar, a charity baseball conference in Boston that raises money to benefit the Jimmy Fund. I’ve cut down that paper a good deal and made some modifications to it based on the feedback I received at the conference. Essentially, I converted the paper into a form that would work well as an article at The Hardball Times.

...WAR has been gaining acceptance, but some of the internet’s best sabermetric minds are distancing themselves from the statistic more than ever, especially as a single-season metric.

The goal of my paper for the Saber Seminar was to evaluate the ability of WAR to describe performance in a given season, as well as to predict future performances in a subsequent season.

...Single-season WAR does a phenomenal job at doing what it says it does. Single-season WAR should not be used to predict win totals or even WAR in a subsequent season. Single-season WAR also is not supposed to reflect the true talent level of a player, which I think is far and away the largest flaw in the way people interpret the statistic. If WAR did reflect true talent, every player would have the same WAR that perfectly encompassed how much value his talent should bring to his team every single year.

Even in the various definitions of WAR, the words “true talent level” never pop up:

The consensus seems to be that WAR is how much value (WINS!!) a player contributes to his team over the baseline of a player who could replace him. WAR does not reflect the true talent level of a player, but instead it describes how many wins an individual player contributes on the actual field, and in that aspect it works spectacularly well.

Repoz Posted: August 08, 2012 at 08:16 AM | 80 comment(s) Login to Bookmark
  Tags: sabermetrics

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. Edmundo got dem ol' Kozma blues again mama Posted: August 08, 2012 at 08:30 AM (#4202980)
Uh, I'm not terribly SABR-y, but this excerpt told me nothing that isn't blantantly obvious. Am I missing something?
   2. PS is probably going to survive his vacation Posted: August 08, 2012 at 08:39 AM (#4202986)
Most people don't know the blatantly obvious. Then again, these are the same people who try to hammer in a screw and declare the hammer useless.
   3. Slivers of Maranville descends into chaos (SdeB) Posted: August 08, 2012 at 09:01 AM (#4203000)
The consensus seems to be that WAR is how much value (WINS!!) a player contributes to his team over the baseline of a player who could replace him
.

It doesn't even do this, correct? You can add up all the WAR that a team produces, subtract it from actual wins, and you won't get the same baseline number in each case. WAR measures something like theoretical average wins over replacement a player's performance would result in across a large number of iterations, or so I thought.
   4. Steve Parris, Je t'aime Posted: August 08, 2012 at 09:07 AM (#4203007)
Can we please retire the Edwin Starr WAR reference? Please?
   5. villageidiom Posted: August 08, 2012 at 09:15 AM (#4203011)
Can we please retire the Edwin Starr WAR reference? Please?
Good God, y'all.
   6. Pasta-diving Jeter (jmac66) Posted: August 08, 2012 at 09:26 AM (#4203017)
Then again, these are the same people who try to hammer in a screw and declare the hammer useless.

which is ridiculous--it's the screw that's useless
   7. villageidiom Posted: August 08, 2012 at 09:27 AM (#4203018)
Uh, I'm not terribly SABR-y, but this excerpt told me nothing that isn't blantantly obvious. Am I missing something?
Some folks can't talk about what has happened without considering whether it's projectable.

A: Hey, did you see Pedro Ciriaco go 4-for-5 last night? That was awesome!

B: He sucks.

A: But wasn't it great?

B: It's not like we can count on that in the future.

It seems like WAR is the "A" version of that conversation, in numbers. WAR isn't about projection of the future, it's about allocation and/or estimation of the past. It's imperfect, but useful enough for that purpose.

Mostly I see the contention with WAR being akin to the made-up conversation I gave above: two people talking past each other, one focused on the past, one focused on the future. When people are talking past each other, they miss blatantly obvious things.
   8. villageidiom Posted: August 08, 2012 at 09:28 AM (#4203019)
which is ridiculous--it's the screw that's useless
You've just earned the nickname "useless screw".
   9. Pasta-diving Jeter (jmac66) Posted: August 08, 2012 at 10:02 AM (#4203051)
You've just earned the nickname "useless screw".

that's what she said
   10. GuyM Posted: August 08, 2012 at 10:11 AM (#4203056)
The fact that WAR works well at the team level doesn't really answer the interesting question, which is how much confidence should we have in the accuracy of the fielding component? No matter how inaccurate DRS was -- even if it had no value at all -- his equations would come out the same at the team level. My view is that single-season fielding WAR should be viewed as a very crude estimate of a player's actual defensive contribution, and should be given much less weight than the offensive component in evaluating a player's performance. For example, if you had two 8 WAR players in a league, one of whom derived .5 wins from fielding and one of whom had 2.0 fielding wins, I would give the first player my MVP vote (assuming WAR was your voting criteria). It's very likely that some of the 20 fielding runs is an illusion, while we know the offensive contributions were real.

   11. AROM Posted: August 08, 2012 at 11:09 AM (#4203143)
The most unsurprising thing in the article is the finding that WAR correlates to team wins exactly as well as pythagorean percentage. Obviously. WAR (at least BBref version) is simply team runs attributed to individuals. If you couldn't add the individuals back up to get team runs, you're doing something wrong. This only partially applies to Fangraphs WAR, with the pitching part based on FIP instead of actual runs.
   12. Kiko Sakata Posted: August 08, 2012 at 11:15 AM (#4203150)
WAR (at least BBref version) is simply team runs attributed to individuals.


Is this true on the offensive side? I assumed it was something more like expected runs, so that WAR would essentially tie perfectly to something like a component Pythag (what B-Pro calls third-order wins or something like that?).
   13. Dale Sams Posted: August 08, 2012 at 11:15 AM (#4203151)
I think WAR is useful in illuminating theories like, "Don't trade mlb-ready position players for relief pitchers."
   14. Booey Posted: August 08, 2012 at 11:34 AM (#4203179)
What was up with the BBREF WAR re-vamp earlier this spring? I was just starting to really get on board with WAR and then all the totals changed (fairly drastically, in some cases) and it killed a lot of it's credibility for me. It's hard to put too much stock into a stat that fluctuates.
   15. Dale Sams Posted: August 08, 2012 at 12:06 PM (#4203230)
If there were a WAR value for soccer, would the GK usually have the highest? I know, indoors for sure.

edit: Now that I think about it,(outdoors) I'm not sure. Who is really stopping more goal chances, a GK or a defender? The defender I would think.
   16. Sean Forman Posted: August 08, 2012 at 12:12 PM (#4203239)
What was up with the BBREF WAR re-vamp earlier this spring? I was just starting to really get on board with WAR and then all the totals changed (fairly drastically, in some cases) and it killed a lot of it's credibility for me. It's hard to put too much stock into a stat that fluctuates.


Who were the drastic changes. The runs to wins calculation mucked things up for about 3-4 days, but after we reworked that into a better system the changes were not that large. We list the biggest changers here.

http://bbref.com/about/war_explained.shtml
   17. Nasty Nate Posted: August 08, 2012 at 12:29 PM (#4203256)
I think WAR is useful in illuminating theories like, "Don't trade mlb-ready position players for relief pitchers."


Delino DeShields for Pedro Martinez!
   18. Benji Gil Gamesh Rises Posted: August 08, 2012 at 12:32 PM (#4203261)
Sean, is that the same time when the change was implemented where you can't add an individual's oWAR and dWAR to equal his total WAR shown?
   19. Slivers of Maranville descends into chaos (SdeB) Posted: August 08, 2012 at 12:39 PM (#4203273)
My view is that single-season fielding WAR should be viewed as a very crude estimate of a player's actual defensive contribution, and should be given much less weight than the offensive component in evaluating a player's performance. For example, if you had two 8 WAR players in a league, one of whom derived .5 wins from fielding and one of whom had 2.0 fielding wins, I would give the first player my MVP vote (assuming WAR was your voting criteria). It's very likely that some of the 20 fielding runs is an illusion, while we know the offensive contributions were real.


If we grant that fielding WAR is more imprecise than batting WAR, is there a reason to assume that fielding WAR consistently overestimates value?

In your example, let's say that the 2.0 fielding WAR for player B may actually be 0.0. Wouldn't it be equally likely to be 4.0? That is, if the error bar is greater, then it will be greater in both directions, no? Note I'm not talking about projectability, which would require you to regress to the mean. I guess what I'm saying is why do you figure that 20 fielding runs is more likely to be an overestimate than an underestimate.
   20. AROM Posted: August 08, 2012 at 12:48 PM (#4203285)
Is this true on the offensive side? I assumed it was something more like expected runs, so that WAR would essentially tie perfectly to something like a component Pythag (what B-Pro calls third-order wins or something like that?).


The way I did it, I used a baseruns formula to generate custom linear weights for every team, and the end result ensured that everything added up to team runs scored. Sean has switched over to Pete Palmer's batting runs formula, so things won't always match up. In most cases the change is not a big deal.

One example of a fairly large change is a team like the 2008 Twins, who scored a bunch more runs than you would expect from their batting stats. I had him at +45 batting that year, now he's at +28.

I'm not saying one method is better or worse than the other. They both have their advantages and disadvantages. I talked with Sean Forman before the switch and told him the change didn't bother me in the least.
   21. Booey Posted: August 08, 2012 at 12:50 PM (#4203288)
Who were the drastic changes


You wouldn't consider a 10-14 WAR difference from the top guys to be a drastic change? That's like 2 MVP caliber seasons. It doesn't really matter with guys like Ruth, Bonds, Cobb, Mantle, etc, who are no brainers either way. The ones that I worried about were guys like McGwire, Sheffield, and Jim Edmonds, who's loss of 5-10 WAR dropped them from solid lower end HOF caliber choices in the 60's to borderline at best candidates in the high 50's.

Edmonds, for example, looked like an easy choice when he was sailing along with almost 68 WAR. He seemed like the kind of guy that the SABR community could get behind and try to make a push to get him into the HOF (since the BBWAA will almost certainly overlook him), same way they did with Blyleven and Santo. Now he's at 57 WAR. Is that even a HOFer? Is it worth a concerted effort to get a borderline guy recognized?
   22. AROM Posted: August 08, 2012 at 12:54 PM (#4203295)
In your example, let's say that the 2.0 fielding WAR for player B may actually be 0.0. Wouldn't it be equally likely to be 4.0? That is, if the error bar is greater, then it will be greater in both directions, no? Note I'm not talking about projectability, which would require you to regress to the mean. I guess what I'm saying is why do you figure that 20 fielding runs is more likely to be an overestimate than an underestimate.


This is where things get Bayesian. We assume that a true talent +40 run fielder is exceptionally rare, if such an animal even exists at all. Average fielders on the other hand, are common. So the chance of an average fielder getting lucky (not just his play, but lucky in the effect of game recording) at showing a +20 season is much greater than the chance of finding the mythical Rabbit Ozzie Belanger and just observing an unlucky season of him.
   23. AROM Posted: August 08, 2012 at 01:00 PM (#4203302)
Looks like Edmonds got a double whammy. He loses about 30 runs on offense, I guess he played on teams that generally outscored their expected runs, and my system was giving him some credit for that. On the defensive side it's about 5 wins. DRS doesn't like his defense as much as TZ did.
   24. AROM Posted: August 08, 2012 at 01:04 PM (#4203306)
Is that even a HOFer? Is it worth a concerted effort to get a borderline guy recognized?


Compare him to Duke Snider. Make sure you include a picture. Photoshop a picture of Edmonds in a Brooklyn Dodger's uniform, turn it gray, and see how many people could actually tell them apart.
   25. McCoy Posted: August 08, 2012 at 01:05 PM (#4203310)
BRef's WAR formula doesn't have a section for measuring the greatness of half-shirts.
   26. Booey Posted: August 08, 2012 at 02:11 PM (#4203412)
Apologies if my critique of WAR offended anybody. I honestly forget sometimes that some of the people who helped create stats like this post here. A genuine kudos to all who put in their time and effort to make things as accurate and up to date as possible. :)

And for anyone who had anything to do with BBREF, that is probably my all time favorite website. Beyond awesome.
   27. fra paolo Posted: August 08, 2012 at 02:13 PM (#4203414)
My view is that single-season fielding WAR should be viewed as a very crude estimate of a player's actual defensive contribution, and should be given much less weight than the offensive component in evaluating a player's performance.

I think this is true, but I don't agree with your wording, exactly. The 2.0 fielding WAR player almost certainly had an excellent season with the glove, and in that sense it isn't really a crude measure at all. One might still vote for the 0.5 WAR player ahead of the better fielder, because the 2.0 guy might actually be only 1.0 WAR better, not 1.5. But, really, I'd prefer to dig a bit deeper, looking at the different teams' defensive contexts in much more detail, to see if I can explain that extra 0.5.

In other words, when comparing a 0.7 fielder to a -0.3 fielder, I don't think any fielding measure I've read about gives me any confidence in saying that there is a significant difference in fielding value between them that is largely attributable to their own efforts.
   28. Booey Posted: August 08, 2012 at 02:14 PM (#4203417)
So what do y'all think a good HOF cutoff for WAR should be (in general,; there are always exceptions, of course)? Most people seem to think around 60ish. Would anyone here not vote for Mac, Edmonds, Sosa, or Sheff based on merit?
   29. Kiko Sakata Posted: August 08, 2012 at 02:29 PM (#4203439)
Would anyone here not vote for Mac, Edmonds, Sosa, or Sheff based on merit?


In a hypothetical "name the best 250 players and call them your Hall of Fame" they probably all make it, but in a hypothetical "Here's a 2013 HOF ballot; follow HOF voting rules" I don't think Sosa makes my ballot. When I looked it up under the old WAR, I think he was 13th on the 2013 ballot in career WAR and Piazza was just behind him. If you don't care about PEDs, he gets squeezed out by McGwire and Palmeiro; if you like WAR or its sabermetric cousins, he gets squeezed out by Trammell, Raines, et al.
   30. madvillain Posted: August 08, 2012 at 02:32 PM (#4203445)
[wrong thread]
   31. Booey Posted: August 08, 2012 at 02:40 PM (#4203462)
In a hypothetical "name the best 250 players and call them your Hall of Fame" they probably all make it, but in a hypothetical "Here's a 2013 HOF ballot; follow HOF voting rules" I don't think Sosa makes my ballot. When I looked it up under the old WAR, I think he was 13th on the 2013 ballot in career WAR and Piazza was just behind him. If you don't care about PEDs, he gets squeezed out by McGwire and Palmeiro; if you like WAR or its sabermetric cousins, he gets squeezed out by Trammell, Raines, et al.


Sosa wouldn't make my 2013 ballot following BBWAA rules either (10 player limit), but only because the ballot is so backlogged with players who should've been elected already (Bagwell, McGwire, Palmeiro, Trammell, Raines, etc). On a more typical ballot, he makes it easily for me.
   32. GuyM Posted: August 08, 2012 at 02:51 PM (#4203484)
The 2.0 fielding WAR player almost certainly had an excellent season with the glove, and in that sense it isn't really a crude measure at all.

I'd say a 2.0 fielding WAR was "almost certainly" an avergae fielder or above, but that's the extent of my confidence. I mean, DRS and UZR sometimes differ by 9-10 runs on a given player, and they use the exact same data! Honestly, if I had to choose between this year's DRS and a career rating I trusted -- assuming the player had at least 3 prior seasons under his belt -- I'd be more comfortable assuming he performed at his career rate than accepting an annual PBP metric.
   33. Voros McCracken of Pinkus Posted: August 08, 2012 at 02:53 PM (#4203490)
So what do y'all think a good HOF cutoff for WAR should be

Career WAR is not close enough to how I would consider voting for the HOF to have a cut off. For example, I'd take Will Clark over Rafael Palmeiro despite having 13 fewer WAR. I'd take Johnny Evers over Lou Whitaker despite 26 less WAR.

Sammy's interesting in that he was both an excellent defender and an outstanding hitter, but never both at the same time...
   34. Never Give an Inge (Dave) Posted: August 08, 2012 at 02:58 PM (#4203500)

The consensus seems to be that WAR is how much value (WINS!!) a player contributes to his team over the baseline of a player who could replace him

There need to be pretty meaningful error bars around any value stat that includes a defensive component. WAR is good but it would really help to have a sense of what the standard error is based on the number of fielding chances.
   35. Booey Posted: August 08, 2012 at 03:04 PM (#4203514)
Career WAR is not close enough to how I would consider voting for the HOF to have a cut off. For example, I'd take Will Clark over Rafael Palmeiro despite having 13 fewer WAR. I'd take Johnny Evers over Lou Whitaker despite 26 less WAR.


I agree on your first sentence. But I'm curious as to why you'd take Clark over Palmeiro (peak, I assume?) and very curious as to Evers over Whitaker.

I also think WAR applies more to the HoM than the HoF, since they're trying to induct people based almost solely on actual value. I think the HoF should elect people based MOSTLY on value, but I have no problem giving extra credit for historic achievements. Sosa and McGwire, for example, probably didn't produce as many wins in their careers as the more durable and consistent Palmeiro did, but I'd elect both of them over Raffy in a heartbeat (for the HOF, maybe not for the HoM). Dudes were averaging over 60 homers a year in their peaks, and overall value aside, that's the type of dominance the HOF is all about, IMO.
   36. AROM Posted: August 08, 2012 at 03:12 PM (#4203525)
Sammy's interesting in that he was both an excellent defender and an outstanding hitter, but never both at the same time...


Kind of like Dwight Evans.
   37. bobm Posted: August 08, 2012 at 03:22 PM (#4203543)
Looks like Edmonds got a double whammy. ... On the defensive side it's about 5 wins. DRS doesn't like his defense as much as TZ did.


How could DRS tell he was staging phony diving catches? :)
   38. Voros McCracken of Pinkus Posted: August 08, 2012 at 03:29 PM (#4203553)
Partially peak, guys who play forever at a middling or slightly above average level are not my idea of a Hall of Famer.

But also that I tend to weight player rank versus the other players in his league rather than a simple deviation from league average. Clark was for multiple consecutive years one of the best players in baseball. Palmeiro never was. Palmeiro had the advantage of doing his damage in leagues where tons of other guys were doing big damage. Evers and Whitaker works the same way. Johnny Evers was much closer to being one of the the best players in his league than Whitaker ever was, even after adjusting for playing in a smaller (and weaker) league. The same can't be said for Alan Trammell.

Probably the best example is to compare Reggie Jackson who has less WAR than Larry Walker despite being far more likely to be among the league leaders in WAR every year.

This isn't just a personal preference, there's a practical aspect to it as well: by making this adjustment you avoid an absolute flood of guys coming in from the 90s and teens thru 30s. You will tend to have a rather uniform number of players coming from each era.
   39. fra paolo Posted: August 08, 2012 at 03:37 PM (#4203570)
I mean, DRS and UZR sometimes differ by 9-10 runs on a given player, and they use the exact same data!

Which is about one win, or roughly as close as I'd like to go with fielding WAR.

If you're saying the 0.5 player ranges from 1.5 to -0.5 fielding WAR, and the 2.0 player ranges from 3.0 to 1.0 fielding WAR, and therefore that the 'real' tally could be 1.5 vs 1.0 in favour of the 0.5 player, then you probably just ought to ignore fielding WAR altogether. It could just as easily be 3.0 vs -0.5. They cancel out.

Based on what I've read about fielding metrics over the years, in the 2.0 vs 0.5 comparison, the 2.0 player more than likely had a better season with the glove. How much better can't be as precise as 1.5, though.
   40. Booey Posted: August 08, 2012 at 03:50 PM (#4203583)
This isn't just a personal preference, there's a practical aspect to it as well: by making this adjustment you avoid an absolute flood of guys coming in from the 90s and teens thru 30s. You will tend to have a rather uniform number of players coming from each era.


But how much accounting do you do for league size and strength? There are twice as many teams and players as there was 60, 70, 100 years ago. Shouldn't there be twice as many HOFers? If Palmeiro was consistently in the top 20 players in his league almost every year, wouldn't that be equivalent to being in the top 10 every year in say, 1920? A HOF that has the same number of players from the 1910's as it does from the 1990's seems to be horribly and unfairly skewed against the modern decade, IMO.

Example: Honus Wagner and Barry Bonds were both easily the NL's best players during the bulk of their careers, but who were the next best position players during their careers?

NL 1900-1917ish - Wagner...and probably no one else who would crack most people's top 100 list.

NL 1990-2007 - Bonds...but also the likes of Pujols, Chipper, Bagwell, Biggio, Piazza, Larkin, Sosa, Gwynn, Sheffield, Walker, Thome, Griffey, Vlad, Helton, Edmonds, Rolen, Kent, etc (not saying all these players are top 100, but ya know what I mean...)
   41. GuyM Posted: August 08, 2012 at 03:56 PM (#4203591)
If you're saying the 0.5 player ranges from 1.5 to -0.5 fielding WAR, and the 2.0 player ranges from 3.0 to 1.0 fielding WAR

I wouldn't say that, for the reasons AROM outlined above: it's very unlikely the second player really was +30 runs. I would say the +2 player probably had a better fielding season, but I really have no idea how much better. If this is all the data I have, I would therefore rely heavily on the offensive WAR to choose between them. However, that may not be all you have. If the +2 fielder has a history of great fielding, then it's more likely to be real; if it's, say, Jeter, then you're going to discount it very heavily.
   42. fra paolo Posted: August 08, 2012 at 04:10 PM (#4203606)
If the +2 fielder has a history of great fielding, then it's more likely to be real

This to me is the problem with the usual scepticism towards fielding metrics. We think of them as individualized, but they are more like RBI than hits. I would look at the team and league context before I'd look at career. It just might be the case that for one season a player's context helped him unusually. And such players still deserve credit for being there and getting the job done.
   43. GuyM Posted: August 08, 2012 at 04:18 PM (#4203623)
They are not like RBI at all. Plays made in the field are almost always individual accomplishments (except GDP). "Context" neither helps nor hurts the fielder. That's the whole point of the fielding metrics -- to try to hold context (i.e. opportunities) equal
   44. AROM Posted: August 08, 2012 at 04:20 PM (#4203628)
When you're trying to rank 75,000 seasons throughout major league history, it helps to have a system, recognize the limitations, and live with the consequences. When you're trying to rank award candidates, you can break away from the system quite a bit and trust your eyes. There are probably 10-15 guys who are worth considering every year.

If the numbers say Mike Trout is +15 run, and you see that he's one of the fastest players around and saw him make that leaping catch against the O's, then that number seems right. If the same system tells you Mark Trumbo is +15 runs in the field, then you can probably safely discount that. In an award situation, I'd probably have fielding stats count for maybe 25%, with the other 75% being what I see and what I hear/read from knowledgeable observers of the game.
   45. fra paolo Posted: August 08, 2012 at 05:28 PM (#4203699)
Plays made in the field are almost always individual accomplishments (except GDP).

Yes, let's forget about the pitcher (and his pitch-by-pitch 'dialogue' with the umpire) and the positioning.

Fielding metrics have been chasing the chimaera of individual accomplishments for too long, and it is just impossible to achieve the same level of confidence that we think we have with batting numbers. But I think we both agree with my second clause.

We should be thinking differently about what we can accomplish with fielding statistics. But that's not the same thing as discounting them quite so heavily as people seem to do.
   46. GuyM Posted: August 08, 2012 at 05:45 PM (#4203709)
Yes, let's forget about the pitcher (and his pitch-by-pitch 'dialogue' with the umpire) and the positioning

On the contrary, the point is to understand exactly what the pitcher accomplished. The easier the resulting BIP was to field, the less credit you want to give to the fielder. Again, that's the goal of all fielding metrics.

I'm not sure why you're averse to considering career data. Surely we are much more likely to figure out what a fielder accomplished this year if we take into consideration how good a fielder he is in general (i.e. what he did in prior years). In fact, with the data we have now it may be impossible to make good single-season estimates of fielding without taking account of career data.
   47. PreservedFish Posted: August 08, 2012 at 06:14 PM (#4203735)
I'm not sure why you're averse to considering career data. Surely we are much more likely to figure out what a fielder accomplished this year if we take into consideration how good a fielder he is in general (i.e. what he did in prior years).


I think this approach could be very useful, but I also recognize that it's kind of a Frankenstein's monster, an unholy combination of value and ability, that the math dorks around here might be scared of.
   48. Walt Davis Posted: August 08, 2012 at 06:53 PM (#4203765)
Probably the best example is to compare Reggie Jackson who has less WAR than Larry Walker despite being far more likely to be among the league leaders in WAR every year.

How much of that is simply in-season durability? For relatively easy comparison, let's take their age 23-32 seasons which is pretty much Reggie's WAR prime. (Technically he is a bit better 22-31 but Walker didn't play at 22 so I'd rather control for age.) Reggie had 54 WAR and 36 WAA and averaged 5.4 WAR per 600 PA.

From 23-32, Walker had 46 WAR but in only 5201 PA which is 5.3 WAR per 600 PA.

Walker produced 1 WAR per 111 PA during this time. From 1990-1999 (essentially excluding 1994), NL seasons with 600+ PA and a better WAR/PA rate produced only 44 seasons (about 5 per). The only names which repeat more than once are Sandberg, Piazze, Bonds and Bagwell. A durable Larry Walker would have competed to be among the league leaders every year. Over that 1990-1999 period (min 2500 PA) only 7 players, both leagues, produced WAR at a better rate than Walker -- Griffey, Bonds, Bagwell, Larkin, Lofton, Piazza and AROD.

And Reggie's performance in this regard was not exactly awe-inspiring. He had 8 top 10 finishes, 4 of them top 5 and none at the top. Using those same criteria for the 69-78 AL, you get 43 seasons (fewer teams but no strike). The multiple repeat names there are Reggie, Grich, Carew and Bando -- good players but that's not quite as scary a list of competitors as Walker's although the main difference is no Bonds for Reggie. Over those 10 years, 6 players produced at a better rate than Reggie -- Morgan, Bench, Carew, Grich, Schmidt and Fisk.

I don't have a problem on ranking Reggie ahead in prime based on that durability but that's counting stats not quality. I don't necessarily mind ranking Reggie ahead because more of Walker's production is wrapped up in less reliable defensive numbers but that's questioning the reliability of the numbers, not really a quality argument. I am not at all convinced that, even in comparison to his contemporaries, that Reggie was a better player than Walker.
   49. fra paolo Posted: August 08, 2012 at 07:42 PM (#4203807)
The easier the resulting BIP was to field, the less credit you want to give to the fielder. Again, that's the goal of all fielding metrics.

Well, yes and no. Where does that leave the pitching-to-contact strategy? Should none of the credit go to the fielder? Should all of it?

I'm not sure why you're averse to considering career data.

Because it's potentially misleading. Why can't a player have a career year with the glove?

As long as we want to dot the decimal point of fielding metrics down to the third of an inning, fielding metric scepticism will make apparent sense. I'm more interested in finding a way out of the corner we've painted ourselves into. That will, I think, require us to put some of that context we have been trying to remove back into our analysis.

Most of my ideas about fielding have derived from thinking about fielding in cricket, of all things. The games are simultaneously different and similar enough to make one realize a few blatantly obvious things that baseball-only approaches seem oblivious to, even when data supporting cricket-derived insights is produced.
   50. PreservedFish Posted: August 08, 2012 at 07:49 PM (#4203816)
Most of my ideas about fielding have derived from thinking about fielding in cricket, of all things. The games are simultaneously different and similar enough to make one realize a few blatantly obvious things that baseball-only approaches seem oblivious to, even when data supporting cricket-derived insights is produced.

Explain.
   51. Walt Davis Posted: August 09, 2012 at 12:27 AM (#4203997)
It's true! I mean, in cricket, you're required to make a play approximately once every 6.5 days* but that one play is worth about 67 runs.

* Wicket keepers excepted -- those guys are pretty awesome.
   52. Sunday silence Posted: August 09, 2012 at 10:14 AM (#4204175)
This is where things get Bayesian. We assume that a true talent +40 run fielder is exceptionally rare, if such an animal even exists at all.


OK here is where you start to lose me. Are you saying that exceptional fielders are LESS likely to be found than exceptional hitters? If that is indeed the assumption, then on what basis do you make that?

If it's just so your numbers fit, or something like that, then that assumption is dubious. They are not playing a different game out there in the field. It's not as if real flesh and blood hitters are hitting the ball, but some kid in his basement has a joystick to control fielders. If that was the case, OK, such an assumption might make sense. But fielders and batters are playing the same game...
   53. Sunday silence Posted: August 09, 2012 at 10:19 AM (#4204181)
The runs to wins calculation mucked things up for about 3-4 days


Is it really necessary to convert runs to wins? Or is it purely ornamental? There's no positional adjustment for say 6 runs created by a CF creates more wins than 6 runs produced by the Catcher, is there? so it's purely decoration?
   54. Sunday silence Posted: August 09, 2012 at 10:22 AM (#4204183)
One example of a fairly large change is a team like the 2008 Twins, who scored a bunch more runs than you would expect from their batting stats. I had him at +45 batting that year, now he's at +28.


this unclear whether you are referring to a player or team...
   55. JJ1986 Posted: August 09, 2012 at 10:24 AM (#4204185)
Are you saying that exceptional fielders are LESS likely to be found than exceptional hitters?


I think people who are exceptionally talented at fielding are just as likely to be found as people who are exceptionally talented at hitting. Even being an exceptional fielder, though, has more limited opportunities. There are only a certain number of plays that an exceptional fielder has available to him that a good or great fielder wouldn't make. While an exceptional hitter can theoretically best a good hitter by a practically infinite amount.
   56. Sunday silence Posted: August 09, 2012 at 10:28 AM (#4204190)
WAR is good but it would really help to have a sense of what the standard error is based on the number of fielding chances.


I am not a math expert, but based on my understanding of the method; I have a feeling that the standard way of computing confidence interval and standard deviations may not be appropriate for the fielding component. Just a hunch, could be wrong...
   57. Sunday silence Posted: August 09, 2012 at 10:37 AM (#4204207)
Fielding metrics have been chasing the chimaera of individual accomplishments for too long, and it is just impossible to achieve the same level of confidence that we think we have with batting numbers


WHy? WHy do most/many of you keep insisting on this? I know it SOUNDS right; but is it really?

I cant understand the reason why the level of confidence should differ, they are playing the same game out there. A guy goes up to the plate 500 times and we assume he's faced a bell curve of pitches, good ones, meatballs, etc. Same goes for fielding. A catches 500 balls, one assumes a bell curve of tough catches, routine catches...

THe counter is that someone e.g. Undru JOnes is consistently calling people off of easy balls. Well 1) if that is so lets see someone document this and 2) this resoning cant apply to players e..g catchers, or 3B who are not in position to do that. So wouldnt there numbers be more accurate? So is there any evidence of this?
   58. JJ1986 Posted: August 09, 2012 at 10:42 AM (#4204215)
WHy? WHy do most/many of you keep insisting on this? I know it SOUNDS right; but is it really?


Fielders don't have as many chances as batters do. Also, a huge number of chances for fielders are plays that 99.9% of major leaguers will make and thus offer no marginal value.
   59. Kiko Sakata Posted: August 09, 2012 at 10:53 AM (#4204238)
Is it really necessary to convert runs to wins?


It's necessary to compare across different run-scoring environments. Individual runs are more valuable the lower the run-scoring environment - e.g., Petco v. Coors, non-DH v. DH league, 1968 v. 2000. I suppose you could do something akin to ballpark adjustments to "normalize" runs, but then those "normalized" runs wouldn't tie to actual runs.
   60. AROM Posted: August 09, 2012 at 11:28 AM (#4204298)
One example of a fairly large change is a team like the 2008 Twins, who scored a bunch more runs than you would expect from their batting stats. I had him at +45 batting that year, now he's at +28.


this unclear whether you are referring to a player or team...


My bad. Those were the batting runs above average for Joe Mauer.

Are you saying that exceptional fielders are LESS likely to be found than exceptional hitters?


I said no such thing. But the magnitude of an exceptional fielder is less than that of a great hitter. It has to be, because much of the responsibility for run prevention falls on the pitcher.
   61. GuyM Posted: August 09, 2012 at 11:35 AM (#4204309)
A guy goes up to the plate 500 times and we assume he's faced a bell curve of pitches, good ones, meatballs, etc. Same goes for fielding. A catches 500 balls, one assumes a bell curve of tough catches, routine catches...

The difference (in addition to # of opportunities), is that a hitter may face a couple of hundred opposing pitchers in a season, while a fielder plays behind fewer than 20. And just 5-6 starters will account for a large share of the fielder's innings. Because pitchers allow very different BIP distributions, the possibility that a fielder will have many more or many fewer opportunities than average is much higher than it is for a hitter. It tends to mostly even out over a career, but not over a season.
   62. alilisd Posted: August 09, 2012 at 12:14 PM (#4204337)
I think WAR is useful in illuminating theories like, "Don't trade mlb-ready position players for relief pitchers."


I wish Josh Byrnes understood this!
   63. Tom Nawrocki Posted: August 09, 2012 at 12:39 PM (#4204357)
The article never answers the title question.
   64. CrosbyBird Posted: August 09, 2012 at 12:46 PM (#4204369)
The difference (in addition to # of opportunities), is that a hitter may face a couple of hundred opposing pitchers in a season, while a fielder plays behind fewer than 20. And just 5-6 starters will account for a large share of the fielder's innings. Because pitchers allow very different BIP distributions, the possibility that a fielder will have many more or many fewer opportunities than average is much higher than it is for a hitter. It tends to mostly even out over a career, but not over a season.

Another enormous difference would be the lack of "true outcomes" for fielders. When a hitter strikes out, walks, or hits a homerun (almost every HR), we can evaluate that without any "noise" from the fielders; it's almost entirely pitcher vs. batter. We can't do the same thing for fielding chances because the chances themselves have already been highly influenced by the pitcher vs. batter matchup.

It's not as if our hitting metrics are close to perfect, but all of the factors that we would have to normalize for hitting have to be normalized for fielding as well, so we've got the same problems and more in terms of isolating ability. Combine wider error bars with a smaller sample size and I think you'd have to conclude that fielding metrics are less reliable than hitting metrics. That doesn't mean "unreliable," mind you. I don't mean to trivialize the work others are doing to deal with this very challenging problem.

All baseball events are context-dependent, but I think it's just the nature of fielding that it is harder to isolate than hitting, the same way individual ability in basketball or football is harder to isolate than individual ability in baseball.
   65. GuyM Posted: August 09, 2012 at 02:07 PM (#4204449)
Another enormous difference would be the lack of "true outcomes" for fielders.

What difference? The fielder makes an out or he doesn't. That's just as true an outcome as we have for hitters. (With the exception of throws not handled by the 1B, where there can be some ambiguity regarding which IF is at fault for failing to make the out.)

The difference is in measuring opportunities, not outcomes. We're all willing to assume that hitters all had roughly the same opportunities. We aren't willing (and shouldn't be) to make the same assumption for fielders.
   66. Don Geovany Soto (chris h.) Posted: August 09, 2012 at 02:20 PM (#4204464)
Another enormous difference would be the lack of "true outcomes" for fielders. When a hitter strikes out, walks, or hits a homerun (almost every HR), we can evaluate that without any "noise" from the fielders; it's almost entirely pitcher vs. batter. We can't do the same thing for fielding chances because the chances themselves have already been highly influenced by the pitcher vs. batter matchup.


But that just means we can't measure "true outcomes" for fielders in terms of outs/game (or outs/inning, or what have you). Instead we have to look at outs/opportunities.

If Pitcher A/Batter A (as a unit) gives us mostly pop flies, then we're measuring our fielders' abilities to handle pop flies. If Pitcher B/Batter B (as a unit) gives us a lot of scorching line drives, then we're measuring our fielders' abilities to handle those. And so on.

In other words, we want to objectively measure how often a fielder makes an out per number of chances to do so, but we have to take a number of details into consideration. I'm not smart enough to be specific on that, but I'd guess it's something like speed, angle, velocity, and "landing spot" of a batted ball (assuming a fly/line drive; something similar would be used for grounders). I don't think anyone thinks that Fielder C would automatically save you X runs/game, because if some pitcher goes up there and throws 20 Ks, the fielder won't see X balls in play anywhere near him.

I'm stumbling over how to express this, but I think everyone gets the idea.
   67. jack the seal clubber (on the sidelines of life) Posted: August 09, 2012 at 03:26 PM (#4204563)
WHy? WHy do most/many of you keep insisting on this?


For some reason when I read this I flashed on Nancy Kerrigan after Tonya Harding's thugs broke her knee.

I don't think there is ever going to be a fielding parameter that is as credible as hitting parameters. It just isn't as easy to measure.
   68. The Id of SugarBear Blanks Posted: August 09, 2012 at 03:42 PM (#4204592)
I don't think there is ever going to be a fielding parameter that is as credible as hitting parameters. It just isn't as easy to measure.

And there's an asymmetry in our demands. When a centerfielder catches a can-o-corn that any other competent major league CF would catch, we credit him with, really, nothing. When a hitter hits a double on a BP fastball that 75% of other major leaguers would hit for a double or better, we credit him with a "double."

We give fielders credit only for fielding "tough chances," but give hitters credit for hitting easy pitches.
   69. JJ1986 Posted: August 09, 2012 at 03:44 PM (#4204598)
And there's an asymmetry in our demands. When a centerfielder catches a can-o-corn that any other competent major league CF would catch, we credit him with, really, nothing. When a hitter hits a double on a BP fastball that 75% of other major leaguers would hit for a double or better, we credit him with a "double."

We give fielders credit only for fielding "tough chances," but give hitters credit for hitting easy pitches.


That's not right. Or, it's right only in so much as it applies to the basic hitting runs. The replacement level, which is subtracted from the base runs, is what any scrub would hit and includes that double.
   70. The Id of SugarBear Blanks Posted: August 09, 2012 at 04:26 PM (#4204671)
But the replacement level hitter isn't seeing all BP fastballs -- that's the difference. He's a meh hitter, but he's seeing a representative selection of pitches.

With fielders the only "credit" they're getting (*) is for converting "tough chances" that lesser fielders wouldn't convert. The fielding equivalent of the BP fastball down the middle is getting airbrushed away.

(*) Or certainly should be getting.
   71. JJ1986 Posted: August 09, 2012 at 04:39 PM (#4204685)
But the replacement level hitter isn't seeing all BP fastballs -- that's the difference. He's a meh hitter, but he's seeing a representative selection of pitches.

With fielders the only "credit" they're getting (*) is for converting "tough chances" that lesser fielders wouldn't convert. The fielding equivalent of the BP fastball down the middle is getting airbrushed away.

(*) Or certainly should be getting.


In both cases, you're trying to count what the actual player is doing above what a replacement level player would do. The replacement fielder would get all of the "easy" chances so they count as 0. The fielder gets credit for everything the replacement isn't getting to.

For hitting, if you were to get more specific, on a BP fastball down the middle, the replacement value might be .8 runs and the hitter you're measuring might double and get .8 runs. He gets no credit, just like the fielder gets no credit for catching a pop up. If he homers, when the replacement value is a double, then he gets positive value. We don't actually measure replacement level hitters like that, but it is the idea behind it, to figure out exactly what a replacement hitter would do given the same situations.
   72. Jittery McFrog Posted: August 09, 2012 at 05:12 PM (#4204706)
But the magnitude of an exceptional fielder is less than that of a great hitter. It has to be, because much of the responsibility for run prevention falls on the pitcher.


You know a lot more about this than I do, so I hesitate to chime in...but: I'm not sure I get this reasoning.

I'm skeptical because every fielder is also a hitter, and we know that teams can, through their roster and positioning choices, exchange offensive production for defensive production. If there are more runs at stake at the plate than in the field, then, if they're being reasonably efficient, teams should tolerate bad hitting much less than bad fielding. A team shouldn't give up 5 runs of offense playing Joe Glove instead of Jim Bat unless Joe Glove saves at least 5 runs with his D.
It seems to me that if we assume these decisions are at some sort of equilibrium then the different run responsibilities of fielding/hitting are going to be baked into the fielding/hitting baselines rather than into the values above the baseline.

In short, I don't see how [ share of fielding < share of hitting ] implies [value of great fielder < value of great hitter ] when the baselines for fielding and hitting value are not selected for independently. Maybe I'm missing a premise or something.
   73. GuyM Posted: August 09, 2012 at 05:24 PM (#4204717)
The missing premise is that the range (variation) in team offense and team defense is the same. So if fielders are providing only a portion of the variation in team defense (pitchers doing the rest), while hitters produce all the offensive variation, then there must be a larger spread in talent among the hitters.

You can also just look at team DER, and see that teams differ in overall defensive effectiveness less than in offensive production.
   74. Jittery McFrog Posted: August 09, 2012 at 07:18 PM (#4204796)
The missing premise is that the range (variation) in team offense and team defense is the same.


Aha, thanks. This makes the conclusion follow, and changes the case from what I thought was an a priori sort of thing to an empirical one. "Because we observe it to be so" is a good reason to believe something, and different from what I thought he was saying.
   75. Walt Davis Posted: August 10, 2012 at 01:00 AM (#4205018)
As noted, a big challenge (the biggest I'd think) in fielding metrics is measuring the opportunities. That is very hard to do reliably, especially in real time. The new FX system might be able to do this much more accurately but right now it must be chock full of measurement error.

But this is also where I think the fielding metrics are being treated somewhat unfairly. They are trying to do something much more difficult than the standard hitting metrics. The fielding metrics are trying to measure the difficulty of each opportunity and adjust for that. That would be the equivalent of adjusting hitting metrics based on the quality of the pitcher and the pitch hit for each PA.

So, as noted, we essentially give a CF no credit for a can of corn ... and if a particular CF happens to see 10% more cans of corn than the typical one, so be it. If a batter does see 10% more BP fastballs or faces crappy relievers or crappy starters more often, that gets measured as hitter value. I know some analysts have tried adjusting for pitcher quality and such and again the FX systems are starting to help us identify good breaking ball hitters, guys who swing a lot outside the zone, etc.

Anyway, an exact hitting equivalent of what the fielding metrics are trying to do would judge hitters on the success of their outcome (relative to average) on every single pitch. He took a pitch a foot outside -- OK everybody but Francoeur does that, so no credit. Hey, he lined that pitch on the outside corner the other way for a single while the typical batter grounds out to the SS.

We tend not to do that for hitters because (a) "opportunity" (PA) is consistently defined and accurately measured and (b) it doesn't seem worth the trouble. But the corresponding "clean" fielding measure is the "chance" but we've decided that straight fielding percentage doesn't really work. We know we have to reward fielders for range which led to RF which is probably the equivalent of going from BA to OBP. But the distribution of BIP apparently is sufficiently varied that we need a finer measure of opportunity.

We'd be in the same boat if, somehow, some batters faced King Felix 20% of the time while others got to face some crappy Rockies pitcher 20% of the time. We'd have no choice but to adjust our hitting measures, especially if there was low year-to-year correlation in how often a batter faced King Felix. Or imagine if, somehow, despite equal playing time some batters got substantially more PA than others either due to talent (cf better range) or something unrelated to their talent (cf playing behind groundball pitchers or no lefty starters).

Essentially, we got lucky when it came to hitting. Given equal playing time, everybody gets a roughly equal number of opportunities (PAs) and it's simple to adjust for that difference. And while a hitter's performance is obviously affected by the mix of pitchers and defenses (and parks) that he faces, they face so many different pitchers (less so defenses) that it comes close enough to evening out over 1-3 seasons.

I think it would be interesting to look at the variation in fielding metrics compared to the variation in hitting performance of (non-platoon preferably) bench players. Reed Johnson has been a bench player* for the last 6 seasons, amassing 1500 PAs, never more than 374 in those seasons. His OPS+ has ranged from 66 to 122. His Rbat has varied from -12 to +9. His WAR has ranged from -.5 to 1.1. It's possible that in some of those "years" he faced substantially tougher pitchers than others. Probably in some years he got to face a lot more lefties than others.

We see fielders in samples of that size going from -12 to +9 runs and we declare WAR broken and wish defensive measures were as reliable as offensive. Maybe we need to chill out a bit.

*Technically I think he was close to full-time to start his age 30 season but he sucked so bad he lost his job and he's been a bench player since.
   76. McCoy Posted: August 10, 2012 at 01:14 AM (#4205026)
Reed Johnson's splits the last 6 years.

L-PA       R-PA     Year  L/R Ratio       VLHP  VRHP      Rbat
85         65       2007     1.31         0.913 0.513     
-12
169        78       2008     2.17         0.848 0.721     
-1
78         46       2009     1.70         0.903 0.628     
-1
109        59       2010     1.85         0.790 0.520     
-5
109        89       2011     1.22         0.797 0.829      9
97         64       2012     1.52         0.861 0.720      3
647        401      Total    1.61 
   77. fra paolo Posted: August 10, 2012 at 11:30 AM (#4205261)
Explain.

Sorry to disappear, but this is equivalent to an essay question that makes a statement and adds 'Discuss'. It is worthy of an article-length explanation, and I'm not ready to do that. But, here is something of an executive summary of the cricket aspects.

It's true! I mean, in cricket, you're required to make a play approximately once every 6.5 days* but that one play is worth about 67 runs.

Walt is amusingly glib here, but he also picks up on the problem of cricket fielding statistics: apparently the number of opportunities is limited, so there are all manner of sample-size issues. Except what Walt is talking about is largely limited to an even smaller data-set, catches in the slips and at short leg. Most catches at positions like cover or mid-wicket are in the nature of 'pop-ups' or soft line drives. That is to say, the credit really goes to the bowler/pitcher. What is really missing in cricket fielding is the concept of 'stops'. If you don't put a fielder at cover or mid-wicket, a test-standard batsman ought to be taking advantage of this. That shot is going to go for more than a quick single, and might go for a four.

Now, in cricket, the batsman can play through 360 degrees, but our baseball batter has much more limited arc to play in. Also, note that in cricket fielders are regularly moved around to suit the bowling strategy for a given over, which also has other factors taken into consideration, like how old the ball is. So if baseball fielding positions are largely static, and if data shows that this makes sense (which it does), then a lot of baseball fielding plays are equivalent to cricket 'stops', and not cricket 'outs'.

I see from a quick scan that the discussion has moved on, and is now pursuing the distinction between 'easy outs' (little or no credit) and 'difficult outs' (lots of credit). This is exactly what cricket does. A 'stop' that prevents the batters from even running a single is more valuable than a 'stop' that saves a four which is less valuable than a catch which results in an out. Inherent in the game of cricket is a measure of value for all fielding events.

I'm not really interested in taking that approach with baseball in the way it seems to be heading. What I want is a system that can be used for all baseball history to assess the quality of fielders with reasonable confidence. We're very close to that, and what we should be doing now is looking at areas of disagreement in existing systems and trying to work out why they happen, not looking at areas of disagreement and saying that because there is disagreement it's all rubbish.

There's also another angle here, which is in the perception of fielding. We should not be thinking about linear progressions in understanding the value of good or bad fielding, which the discussion here itself has touched upon.
   78. fra paolo Posted: August 10, 2012 at 11:50 AM (#4205289)
The difference is in measuring opportunities, not outcomes. We're all willing to assume that hitters all had roughly the same opportunities. We aren't willing (and shouldn't be) to make the same assumption for fielders.

In other words, we want to objectively measure how often a fielder makes an out per number of chances to do so, but we have to take a number of details into consideration. I'm not smart enough to be specific on that, but I'd guess it's something like speed, angle, velocity, and "landing spot" of a batted ball (assuming a fly/line drive; something similar would be used for grounders). I don't think anyone thinks that Fielder C would automatically save you X runs/game, because if some pitcher goes up there and throws 20 Ks, the fielder won't see X balls in play anywhere near him.

See, knowing what I know about cricket just makes me think these statements are equivalent to embracing a Sisyphean task.

We know that in certain areas of the field a major-league fielder is expected to make the play. That's why they play where they do. So either these count for nothing, because it's the pitcher's job to help the batter hit the ball there, or they count for something, because if the fielder wasn't there (which could happen in cricket) the batter would get on base.

My hypothesis is that we should be doing is discard 'received wisdom' and start from the assumption that there is no reason for the shortstop to play where he does. So full faith and credit for being there.
   79. PreservedFish Posted: August 10, 2012 at 11:58 AM (#4205297)
Thank you fra paolo.
   80. AROM Posted: August 10, 2012 at 01:11 PM (#4205355)
And there's an asymmetry in our demands. When a centerfielder catches a can-o-corn that any other competent major league CF would catch, we credit him with, really, nothing. When a hitter hits a double on a BP fastball that 75% of other major leaguers would hit for a double or better, we credit him with a "double."


I don't think that's any pitch is 75% likely to wind up like that. In home run derby they get nothing but BP fastballs and I don't think the success rate is 75%. Probably 35-50%. And that's when you expect a BP fastball. In a game situation even if you get one you might have been expecting a curve. Certainly some pitches are harder to hit than others, but the range of batting average, based on pitch location in the zone, is probably more like .200 to .350.

The only pitches that come close to the fixed outcome of a popup are the ones so far out of the zone that even Frenchy, Vlad, Yuni, and Ichiro wouldn't swing at them.

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
The Piehole of David Wells
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

NewsblogNY Times: The Upshot: Up Close on Baseball’s Borders
(9 - 2:53pm, Apr 24)
Last: Pasta-diving Jeter (jmac66)

NewsblogCalcaterra: Blogger Murray Chass attacks me for bad reporting, ignores quotes, evidence in doing so
(22 - 2:53pm, Apr 24)
Last: McCoy

NewsblogJosh Lueke Is A Rapist, You Say? Keep Saying It.
(244 - 2:52pm, Apr 24)
Last: base ball chick

NewsblogOMNICHATTER for 4-24-2014
(24 - 2:50pm, Apr 24)
Last: Nasty Nate

NewsblogKeri: Slump City: Why Does the 2014 MLB Season Suddenly Feel Like 1968?
(40 - 2:50pm, Apr 24)
Last: Mess with the Meat, you get the Wad!

NewsblogMichael Pineda ejected from Red Sox game after pine tar discovered on neck
(93 - 2:48pm, Apr 24)
Last: Ron J2

NewsblogOTP April 2014: BurstNET Sued for Not Making Equipment Lease Payments
(2530 - 2:47pm, Apr 24)
Last: Joe Kehoskie

NewsblogDoyel: How was Gerrit Cole not suspended? He basically started the brawl
(41 - 2:46pm, Apr 24)
Last: Pat Rapper's Delight

NewsblogMatt Williams: No problem with Harper's two-strike bunting
(23 - 2:45pm, Apr 24)
Last: The Polish Sausage Racer

NewsblogColiseum Authority accuses Athletics of not paying rent
(20 - 2:42pm, Apr 24)
Last: Jeff R., P***y Mainlander

NewsblogToronto Star: Blue Jays pave way for grass at the Rogers Centre
(11 - 2:42pm, Apr 24)
Last: if nature called, ladodger34 would listen

NewsblogThe Five “Acts” of Ike Davis’s Career, and Why Trading Ike Was a Mistake
(64 - 1:47pm, Apr 24)
Last: Ray (RDP)

Newsblog4 balls, you’re out!
(59 - 1:44pm, Apr 24)
Last: Sunday silence

NewsblogOT: NBA Monthly Thread - April 2014
(506 - 1:23pm, Apr 24)
Last: Jimmy P

NewsblogJonah Keri Extended Interview | Video | Late Night with Seth Meyers | NBC
(12 - 1:05pm, Apr 24)
Last: Greg K

Demarini, Easton and TPX Baseball Bats

 

 

 

 

Page rendered in 1.1346 seconds
52 querie(s) executed