Dr. Strangegloves

August 30, 2011 5:53 PM

Who was the worst defender in the history of baseball? A commenter in a baseball-fever thread compiles a list of the bottom 100 career dWAR figures of all time -- in other words, the 100 players who cost their teams the most wins with the glove. (Joe Posnanski on the WAR metric, for those unfamiliar with it.) The list is an interesting mix of players whose bats allowed them to stay in the game for years despite terrible glovework (Bernie Williams, Manny Ramirez, Dave Winfield) and players who were so bad in the field that they managed to rack up a lot of negative dWAR in shorter careers (Chris Gomez, Dean Palmer.) Toby "Stone Fingers" Harrah is #14 with a -10.9 dWAR. Dick "Dr. Strangeglove" Stuart just misses at -6.1. Some active players have a chance to finish high on the list: Ty Wigginton is only 33 and has already bumbled away enough balls in 2011 to "improve" his ranking from 24th to 15th. Worst of all time? No, it's not the Captain -- Derek Jeter is #2 on the all time list with -13.4 dWAR. Can you guess the "winner"?
posted by escabeche (85 comments total)

If you're interested, the ten all-time (post-1900) leaders in WAR:

1. Brooks Robinson
2. Andruw Jones
3. Roberto Clemente
4. Ozzie Smith
5. Mark Belanger
6. Barry Bonds
7. Carl Yastrzemski
8. Willie Mays
9. Cal Ripken
10. Ivan Rodriguez

Interesting how, with one exception, it just about exactly follows historical perception. That one exception is telling, though - the stats involved are still, by their creator's admission, not super at tracking catchers and first basemen, and while generally good for left-fielders, sometimes gives them wackytown results too. Barry Bonds was a great, great left-fielder pre-steroids, legitimately all-time great, but he makes it onto this list because he has two or three seasons that are just numerically ridiculous.
posted by Simon! at 6:06 PM on August 30, 2011

Leaders in dWAR, that is.
posted by Simon! at 6:06 PM on August 30, 2011

Somehow I don't see ESPN putting an countdown on their website for the day when Jeter passes Mr. Sheffield.
posted by mikedouglas at 6:08 PM on August 30, 2011 [1 favorite]

Interesting how adamant the Posnanski link is about pitching WAR as a work in progress:

Defensive statistics are more advanced and more controversial than ever, and because of that Baseball Reference WAR and Fangraphs WAR can fluctuate pretty wildly (Baseball Reference uses Total Zone as their defensive stat; Fangraphs uses UZR). But the point to me is that WAR is TRYING to figure the defensive contribution. It’s TRYING to get at the whole player, and not just the obvious stuff. It’s TRYING to make an educated guess about a player’s total value. I don’t have to agree with all the conclusions. I don’t have to like the inconsistencies, don’t have to like that someone like Josh Hamilton is ranked as by far the most valuable player in all of baseball (by Fangraphs) and as the fourth most valuable player in the American League (by Baseball Reference). I don’t have to agree with all the conclusions. I can live in shades of gray, I really can. I can decide for myself what stats are worth. I can like WAR without having it determine every single thing I believe or like about baseball.

posted by danb at 6:08 PM on August 30, 2011

As it turns out, I was watching classic Little league World Series games on ESPN Classics this weekend and the #1 player on the list was playing in the game. Dwight Gooden's cousin was quite the younger player.

If you believe this calculation, the Captain should have been traded years ago. What this does not measure is the intangible of being a winner. Jeter simply, like Michael Jordan, knows how to help his team win. His intangibles are terrific.
posted by JohnnyGunn at 6:10 PM on August 30, 2011

If you believe this calculation, the Captain should have been traded years ago. What this does not measure is the intangible of being a winner. Jeter simply, like Michael Jordan, knows how to help his team win. His intangibles are terrific.

Oh my God, now Joe Morgan is MetaFilter's Own!
posted by Horace Rumpole at 6:14 PM on August 30, 2011 [29 favorites]

How WAR is calculated if you are entirely new to baseball and need more than why it is calculated
posted by Blasdelb at 6:19 PM on August 30, 2011

I get it, but I have to say Jeter made what is to my mind the single most brilliant defensive play under extreme duress that I have ever seen, that is the flip at the plate to get out Ozzie Canseco, who was so sure he was going to score he didnt even bother to slide. That was the most heads up baseball defensive play I ever saw. There have been others much more athletic, but for sheer imaginatiion and performing under pressure that play is hard to beat. And I dont even like the Yankees.
posted by jcworth at 6:20 PM on August 30, 2011 [2 favorites]

Yeah, I remember Sheff's last couple seasons in Detroit, where he was all pissy because they just wanted him at DH. He would have had to actively take balls and throw them into the stands in order to be a worse defender.
posted by klangklangston at 6:21 PM on August 30, 2011

If you believe this calculation, the Captain should have been traded years ago. What this does not measure is the intangible of being a winner. Jeter simply, like Michael Jordan, knows how to help his team win. His intangibles are terrific.

Wait... are you serious?
posted by The Michael The at 6:23 PM on August 30, 2011 [3 favorites]

Fuck the heck?
posted by SpiffyRob at 6:28 PM on August 30, 2011 [5 favorites]

"If you believe this calculation, the Captain should have been traded years ago. What this does not measure is the intangible of being a winner. Jeter simply, like Michael Jordan, knows how to help his team win. His intangibles are terrific."

If they had a stat for proclaiming cliched and wrong received baseball knowledge, you'd lose points only for failing to say "clutch" at least once in your comment.

But if you looked at the POS column, the positional WAR which accounts for both offense and defense, you'd see that both Sheff and Cap'n Yankme have positive WAR. And if you looked at Jeter's career stats, you'd see that his terrible dWAR is largely reflected in the huge number of errors that he had in a handful of seasons, regularly in the high teens and sometimes over twenty.

So maybe don't trade them, but realize that they're actually pretty shitty in the field, and wouldn't last in the NL.
posted by klangklangston at 6:29 PM on August 30, 2011 [4 favorites]

I get it, but I have to say Jeter made what is to my mind the single most brilliant defensive play under extreme duress that I have ever seen, that is the flip at the plate to get out Ozzie Canseco, who was so sure he was going to score he didnt even bother to slide.

That was indeed a great play. Jeter is a very smart defender; his problem is that his range is really small, so tons of ground balls get by him into left field for singles.

(It was Jeremy Giambi who was out at home, by the way.)
posted by dfan at 6:32 PM on August 30, 2011 [2 favorites]

Jeter simply, like Michael Jordan, knows how to help his team win.

If you go by what happens on the field, Jeter knows how to help a team with the best talent in the major leagues lose in the first round of the playoffs year after year.

But more to the point, this: if you believe this calculation, the Captain should have been traded years ago. -- is crazy. If you believe the calculation -- and I do -- Jeter has been one of the best shortstops of the era, despite his ineptitude at fielding his position, simply because he is a great hitter and always has been.

But let's not talk about Jeter like we always do, let's talk about other crappy fielders!
posted by escabeche at 6:32 PM on August 30, 2011

Oh, the trolling of Yankee fans that is about ensue on my twitter and facebook feeds . . .
posted by KingEdRa at 6:34 PM on August 30, 2011 [6 favorites]

I'm a little sad to see Eddie Yost on there. The Walkin' Eddies (Yost, Joost, Stanky, and Lake) are a big part of my favorite bizarre phenomenon of 50s baseball, when shortstops started drawing 120 walks a year.
posted by Copronymus at 6:35 PM on August 30, 2011

And if you looked at Jeter's career stats, you'd see that his terrible dWAR is largely reflected in the huge number of errors that he had in a handful of seasons, regularly in the high teens and sometimes over twenty.

Jeter has never made a ton of errors (he's never come that close to the league leader). What kills him on the defensive metrics isn't the errors, it's the singles he doesn't make a play on at all.
posted by dfan at 6:38 PM on August 30, 2011 [1 favorite]

escabeche: "Worst of all time? No, it's not the Captain -- Derek Jeter is #2 on the all time list with -13.4 dWAR. Can you guess the "winner"?"

I'm a Braves fan. I knew.

Re #2, Derek Jeter #2,: Fucking Yankees, Reports Nation
posted by ob1quixote at 6:43 PM on August 30, 2011 [1 favorite]

Why did this guy reverse-engineer "dWAR" rather than just using the original Total Zone data that was used to calculate the defensive component of BB-Ref's WAR numbers?

All defensive statistics should be treated with a fair amount of suspicion, and historical ones even more. What's more, the Total Zone defense numbers used to generate BB-Ref WAR are generally thought far inferior in accuracy to UZR (which is used in Fangraphs' WAR). Basically every number here needs to be taken with a gigantic grain of salt compared to any number that concerns hitting or pitching.
posted by RogerB at 6:46 PM on August 30, 2011

Dave Kingman? I always thought Dave Kingman was the worst. Doesn't even appear on the list. It was that sort of endearing ability to do anything to lose that had the Cubs put him at third base for a time.
posted by dances_with_sneetches at 6:46 PM on August 30, 2011

Oh, correct DFAN, twas Giambi... I knew it was one of the lesser brothers... ; )
Yeah, not athletic by any means, but I will never forget that play (A;s fan, alas)
posted by jcworth at 6:50 PM on August 30, 2011

Horace Rumpole: "Oh my God, now Joe Morgan is MetaFilter's Own!"

I might need to start a MeTa for this putdown. Nothing is worse than being called Joe Morgan.

Everyone knows that the best color man in all of baseball was Phil Ruzzuto. Second, while I am an admitted Yankee fan, I do not particularly love DJ. When y'all refer to the Yankee captain, the first person that comes to mind is Thurman Munson who we all know was the best catcher of his era, better than Fisk and if he had lived would be in the HOF right now.
posted by JohnnyGunn at 6:53 PM on August 30, 2011 [3 favorites]

He would have had to actively take balls and throw them into the stands in order to be a worse defender.

I wonder if his DWAR includes the times he actually did this in Milwaukee?
posted by drezdn at 6:54 PM on August 30, 2011

What's more, the Total Zone defense numbers used to generate BB-Ref WAR are generally thought far inferior in accuracy to UZR (which is used in Fangraphs' WAR).

I think this is what it is like when I talk about computers in front of my grandmother.
posted by Justinian at 6:58 PM on August 30, 2011 [8 favorites]

sort of a silly stat to calculate - it's cumulative, so the more years you play, the higher your potential score is. But you only get to play 20 years in the majors if you have other skills.

The total WAR for the most of the top 50 being >0 speaks to this. The first guy with a negative WAR is at 10. Oh BTW - former red sox, brought in by St. Theo.
posted by JPD at 6:59 PM on August 30, 2011

This just cements my opinion that, pound for pound, Willie Mays is the best player ever. Hit for power, hit for average and a great glove man. Say Hey.
posted by jonmc at 7:00 PM on August 30, 2011 [5 favorites]

"Jeter has never made a ton of errors (he's never come that close to the league leader). What kills him on the defensive metrics isn't the errors, it's the singles he doesn't make a play on at all."

You're right — I just took a look, and his rtot doesn't really correlate with his errors, though 2000 was a particularly shitty year for both.

I do get a little chuckle out of noting the number of times that his defense cost the Yankees runs while still being named Golden Glove at his position. But it must be his intangible golden glove or something.
posted by klangklangston at 7:05 PM on August 30, 2011

oh the other thing that's screwy about dWAR is it is scaled by position right? So on an absolute basis a centerfielder (cough - Bernie) could be a much better fielder than a corner outfielder (cough - Manny), but actually come up with a lower dWAR. It makes sense if you think about it, but it makes looking at the absolute numbers for comparisons a little weird. It probably tells you B. Williams should have been a corner outfielder, but it doesn't tell you Manny was actually absolutely a better fielder.

Not to mention if the scalings are off the whole thing is screwy
posted by JPD at 7:07 PM on August 30, 2011

Jeter's "The Flip" (video).
posted by starman at 7:11 PM on August 30, 2011 [1 favorite]

I think this is what it is like when I talk about computers in front of my grandmother.

Ha, and point taken. Still, it's pretty easy to see just from the basic description that the way Total Zone is generated is pretty damn quick-and-dirty (things definitely get more complicated once you start looking into how UZR is calculated and why it's better). TZ really only becomes at all useful for very large datasets like players' whole careers, where UZR is also dubious about single seasons or less, but moderately trustworthy for three or more years in aggregate. The big advantage to the way TZ is calculated, compared to otherwise better defense metrics, is that it can be done for historical games just from Retrosheet's play-by-play recaps. So it's nice in that it enables this kind of interesting historical comparison — but you have to realize that what you're comparing here is just not as meaningful data as the hitting or pitching components of (anybody's way of calculating) WAR.
posted by RogerB at 7:13 PM on August 30, 2011

If you go by what happens on the field, Jeter knows how to help a team with the best talent in the major leagues lose in the first round of the playoffs year after year.

Oh please. The Yankees have made the playoffs 16 of the 17 years Jeter's been in the majors. They got knocked out 6 times in the first round - which means that ten times, they didn't. If you're a Red Sox fan and need help with that, that's a .625 record.

Compare that to the number of active players with more rings than Jeter: Zero.
posted by deadmessenger at 7:16 PM on August 30, 2011

deadmessenger: "Compare that to the number of active players with more rings than Jeter: Zero."

Frankie Crosetti for 3rd best player of all time, Don Mattingly for worst (tied).
posted by Copronymus at 7:22 PM on August 30, 2011 [2 favorites]

Mattingly and Jeter in the same comment. That my friend is some impressive trolling. Like a Dogwhistle to yankee fans. Just add Munson for the trifecta. Kudos.
posted by JPD at 7:25 PM on August 30, 2011

Compare that to the number of active players with more rings than Jeter: Zero.


Exactly. A few more and he'll be as good as Bill Dickey, Herb Pennock, Vic Raschi, Red Ruffing, and Frankie Cosetti!
posted by googly at 7:27 PM on August 30, 2011 [1 favorite]

What this does not measure is the intangible of being a winner.

The measure for that is usually called the "score", is it not? The final "score" of the "game"?
posted by mhoye at 7:28 PM on August 30, 2011

If you believe this calculation, the Captain should have been traded years ago.

Nah, all this calculation means is that he's a below-average defender. He still has historically been an excellent offensive player, which more than offsets his defensive liabilities. He's just not as incredibly awesome as one might immediately assume.

What this does not measure is the intangible of being a winner. Jeter simply, like Michael Jordan, knows how to help his team win. His intangibles are terrific.

Michael Jordan actually had pretty great tangibles too.

I'm pretty sure that if Derek Jeter had been on the Pirates the last 17 years, we wouldn't be discussing his amazing intangibles.
posted by dfan at 7:44 PM on August 30, 2011 [8 favorites]

Michael Jordan actually had pretty great tangibles too.

I'm pretty sure that if Derek Jeter had been on the Pirates the last 17 years, we wouldn't be discussing his amazing intangibles.

No, but he's still 4th in WAR amongst active players. Don't pretend like the data says he's a mediocre player.

I mean he's not MJ, but even without the rings he's a very very good baseball player, probably a marginal HoF'er.

He's the bizarro Joe Dumars - so overrated he's become underrated.
posted by JPD at 7:58 PM on August 30, 2011

No, but he's still 4th in WAR amongst active players. Don't pretend like the data says he's a mediocre player.

I didn't! He's a great player.

I mean he's not MJ, but even without the rings he's a very very good baseball player, probably a marginal HoF'er.

I think he's a clear Hall of Famer, not marginal at all.

Going by the baseball-reference WAR stats, if he were an average defensive shortstop he would have the 25th highest career WAR of all time, which is as good a way to measure overall greatness as any.

When you take defense into account, he still has the 80th highest career WAR of all time. That gets into my Hall of Fame for sure.
posted by dfan at 8:04 PM on August 30, 2011

Sabermetrics is to baseball what the Tea Party is to politics.
posted by any major dude at 8:09 PM on August 30, 2011 [1 favorite]

Sorry, to be fair, that comparison should be 25th among position players if you count offense only, and 53th among position players (no pitchers) overall.

By the way, Sheffield, #1 on this list of players with the least career defensive value, is 29th and 89th. The players on this list generally have a lot of total value.
posted by dfan at 8:12 PM on August 30, 2011

I think the biggest indictment of this stat is that Manny wasn't #1. He combined fielding incompetence with lack of effort in a way that was unprecedented. I'm no Jeter fan but at least he gave a shit.
posted by TheShadowKnows at 8:24 PM on August 30, 2011

"Sabermetrics is to baseball what the Tea Party is to politics."

And that comment is to this discussion what Adam Dunn is to the White Sox.
posted by klangklangston at 8:43 PM on August 30, 2011 [13 favorites]

Compare that to the number of active players with more rings than Jeter: Zero.

Compare that to the number of active players with a higher average team payroll over a career than Jeter: Zero.
posted by Sphinx at 8:50 PM on August 30, 2011 [2 favorites]

I don't see any discussion, I see a lot of statheads using spreadsheets to yet again attempt to denigrate one of the greatest players all time. The fact that you will never admit is that the ability between an average major leaguer and the best is marginal ( a couple of hits a week, an extra step in the field) so that makes unquantifiable values like heart, and chemistry play a much bigger role in a team's success than any individual stat.

statistics should be like a lampost is for a drunk - use them for support, not illumination. - Vin Scully
posted by any major dude at 10:02 PM on August 30, 2011

I see a lot of statheads using spreadsheets to yet again attempt to denigrate one of the greatest players all time.

Hey, I like Gary Sheffield too, but don't you think that's a little strong?
posted by escabeche at 10:13 PM on August 30, 2011 [1 favorite]

How does Bill Buckner not make this list? His error cost the Red Sox a world series. HE IS MY HERO.
posted by three blind mice at 10:25 PM on August 30, 2011

"I don't see any discussion, I see a lot of statheads using spreadsheets to yet again attempt to denigrate one of the greatest players all time."

And I for one am sick of the rigorous, empirical approach the Tea Party has taken with politics.

"The fact that you will never admit is that the ability between an average major leaguer and the best is marginal ( a couple of hits a week, an extra step in the field)"

Well, this year, the difference between Derek Jeter and league average is pretty marginal, you're right. About .3 of a win. So I'm totally willing to admit that (I'm even willing to admit that Jeter is pretty good at baseball — not as good as A-Rod, though).

"so that makes unquantifiable values like heart, and chemistry play a much bigger role in a team's success than any individual stat."

While Jeter's heart alone manages to keep unicorns alive in Central Park, and his chemistry has already cured feline AIDS, the WAR of the Yankees plus the replacement value of 48.6, it predicts that they should have 81.79 wins. I think that Jeter's heart and chemistry might actually be holding them back — that or the pernicious habit of Bud Selig rounding down for the standings.

But please, tell me more about how the unquantifiable stats are really more quantifiable as wins.
posted by klangklangston at 10:57 PM on August 30, 2011 [6 favorites]

Oh yeah, I meant to link to: Yankees team stats.

Maybe Jeter could take lessons on heart from Granderson. Or Nick Swisher, who is often accused of chemistry.
posted by klangklangston at 11:01 PM on August 30, 2011

Brooks Robinson probably fucked your grandma.

Just sayin'. He was the man.
posted by bardic at 11:22 PM on August 30, 2011

Take away good numbers and leave the intangibles, and you'll never have a player worth a damn. Take away intangibles and just leave good numbers, and you have a Hall of Fame career.

Numbers matter. More than anything else. Anyone who says different, tell me exactly how you objectively determine Willie Mays was better than Claudell Washington.
posted by grubi at 11:32 PM on August 30, 2011

Numbers matter. More than anything else. Anyone who says different, tell me exactly how you objectively determine Willie Mays was better than Claudell Washington.

you'd watch them play. Also, you cannot objectively compare players of two different eras so that point is moot. A great player can only objectively be compared to his contemporaries. The steroid era has now made that even more of a necessity.
posted by any major dude at 11:43 PM on August 30, 2011

Way to turn Rounders into a math contest guys. This is what gets me about baseball: it takes ages, nothing much happens, and everyone is more interested in numbers like .409.
posted by marienbad at 1:41 AM on August 31, 2011

You should file some kind of complaint with whoever forced you to read this post, then.
posted by Holy Zarquon's Singing Fish at 3:56 AM on August 31, 2011 [1 favorite]

Nothing is worse than being called Joe Morgan.

Oh, don't be such a Tim McCarver.
posted by Thorzdad at 4:43 AM on August 31, 2011 [2 favorites]

Everyone knows that the best color man in all of baseball was Phil Ruzzuto.

Jerry Trupiano, working with and Joe Castiglione. Not even close - it was like having two awesome uncles explaining the game as it happened. Troop: low and measured voice, almost a growl, Joe: high and intense and sincere, both genuinely liked each other and knew the game. Best broadcaster combo I've ever heard - To enjoy baseball, you watch it on TV. To understand baseball, you listen to it on the radio with guys like Joe and Troop.

Trupiano was replaced by some schmoe with a more "corporate" voice, and Joe still does color work, but he was a better play-by-play man. No one could put as much visceral disgust and bitter disappointment into the phrase "Swing and a popup."
posted by Slap*Happy at 4:54 AM on August 31, 2011

Dave Kingman? I always thought Dave Kingman was the worst.

He was an okay defender. He's legendary as the guy who did the least with the most. His big negative was his strikeouts, he'd swing at anything. The only reason he made it to the bigs was as tall as he was (6'6") and as long as his arms were (seriously), he'd often hit the ball anyway. The reason he stayed in the bigs so long is that I don't think anybody, ever, hit a ball harder than Kong. (at bat starts about 1:08 in) Those arms would come around, slowly, and it would either be Yet Another Strikeout or Hey Hey!

In that same game, he hit two others, including this first pitch fastball clear out of the park and down Kenmore Avenue. Of course, this being the Cubs, they *lost* this game 23-22. It wasn't just Wrigley -- he'd hit like this everywhere, and even when the winds were screaming from the Southwest, that would help balls heading out center and right field, but down the left field line, you were hitting across the wind and the wind coming over the upper deck would curl downwards and turn left field homers into doubles off the wall or outs. Didn't matter to Kingman, he'd happily pull the ball to left, hit it down the longest line in the Major Leagues, and hit the houses across the street.

Kong was the only guy I know of who managed to hit a ball completely out of the park in a domed stadium, when he hit a popup and found a hole in the Metrodome roof. He also hit the longest out in the game, when he nailed a speaker hanging from the ceiling of the Kingdome -- a good 250 feet out and 150 feet above the field. It bounced off the speaker and was caught for the ought.

Stats tell the tale. Career Home Runs: 442(!!). Career Batting Average: .236(?). Walks/Strikeouts, a miserable .335(???), leaving him with a still above average .780 OPS, but someone who hit the ball that hard -- a .478 career slugging percentage (there are only 99 with a slugging percentage over .500) should have had an OPS well over .850. If he'd had a .500 walk/K ration, he'd be in the hall. I think he's the only eligible player in the game to have over 400 HRs and not be in the hall -- and he's not going to be. (Correctly, IMHO.)

But when he was on, man, was he fun to watch. But he was the ultimate feast or famine. On one day, he'd hit for the cycle, or launch three baseballs into Waukegan. On another day, he'd be wearing a golden sombrero.
posted by eriko at 5:42 AM on August 31, 2011 [2 favorites]

What kills him on the defensive metrics isn't the errors, it's the singles he doesn't make a play on at all.

That's why we call him Derek "Past A Diving" Jeter. Next time Tim McCarver calls a Yankee game, count how many times he says "Past a diving Jeter" when there's a sharp grounder that makes it to the outfield. It will happen a lot.
posted by notmydesk at 6:25 AM on August 31, 2011 [1 favorite]

When they stop keeping score (and determining champions based on that), then I'll be happy to admit numbers aren't important. Until then, there is no such thing as professional baseball without numbers.
posted by grubi at 6:37 AM on August 31, 2011

I'm not a stats geek by any stretch...Is there some sort of adjustment in those dWAR stats that balance for someone who was routinely mediocre year-after-year, versus a player who was actually pretty good for most of their career but stuck-around a few seasons too many and had really bad numbers in their last few years?
posted by Thorzdad at 6:39 AM on August 31, 2011

Is there some sort of adjustment in those dWAR stats that balance for someone who was routinely mediocre year-after-year, versus a player who was actually pretty good for most of their career but stuck-around a few seasons too many and had really bad numbers in their last few years?

Nope, it just sums their career totals. It's just like not being able to tell from a pitcher's career won-loss record whether he was great in the middle but had a couple of terrible seasons at the end (or beginning) or was just good the whole time. (Of course, pitcher W-L records have other issues...)
posted by dfan at 7:02 AM on August 31, 2011

This post and ensuing discussion is the greatest Mefi post in a long, long time. Thank you!

Also, as a Brewers fan, Sheffield can go to hell. He's the only guy in baseball I loathe more than Tony La Russa.
posted by rocketman at 7:10 AM on August 31, 2011 [1 favorite]

Is there some sort of adjustment in those dWAR stats that balance for someone who was routinely mediocre year-after-year, versus a player who was actually pretty good for most of their career but stuck-around a few seasons too many and had really bad numbers in their last few years?

Nope, it just sums their career totals. It's just like not being able to tell from a pitcher's career won-loss record whether he was great in the middle but had a couple of terrible seasons at the end (or beginning) or was just good the whole time. (Of course, pitcher W-L records have other issues...)


This highlights something I think is wrong with a lot of sabremetric analysis (at least the stuff in the public domain) - its just not that sophisticated from a statistical perspective. Its pretty easy to adjust for changes in time like like this, but the Sabre guys aren't doing it yet.

The other thing they seem to address poorly is precision - I.e. there is a lot of false precision in their analysis. Saying someone is either the 29th or 129th best player ever actually has no information in it. But people talk about it like it does.
posted by JPD at 7:19 AM on August 31, 2011

I don't see any discussion, I see a lot of statheads using spreadsheets to yet again attempt to denigrate one of the greatest players all time. The fact that you will never admit is that the ability between an average major leaguer and the best is marginal ( a couple of hits a week, an extra step in the field) so that makes unquantifiable values like heart

See I obviously disagree with AMD on the issue of statistics in sports, but I actually think there is some real value in this sentence. It takes a lot of data for their to be information in that couple of extra hits a week, or an extra step in the field - but I don't ever see that part of the analysis. When is there enough data to be highly confident something is different from the mean?
posted by JPD at 7:21 AM on August 31, 2011

Sheffield is certainly a worthy winner. When he played for the Braves a few years back, watching him and Andruw Jones in the outfield at the same time was an essay in cognitive dissonance.

I'm a little surprised to see Willie McCovey on the list. I don't remember Stretch being a particularly awful first baseman.
posted by steambadger at 7:28 AM on August 31, 2011

I'm a little surprised to see Willie McCovey on the list. I don't remember Stretch being a particularly awful first baseman.


If you read all of the comments there is some discussion that the analysis doesn't work very well for first baseman.
posted by JPD at 7:32 AM on August 31, 2011

If you read all of the comments there is some discussion that the analysis doesn't work very well for first baseman.

If I'm remembering correctly, it also completely gives up on catcher defense and basically just gives them all 0s for everything except maybe throwing out basestealers. WAR is a fun thing to play around with, but the defensive numbers are, at best, a starting point.
posted by Copronymus at 7:56 AM on August 31, 2011

"Take away good numbers and leave the intangibles, and you'll never have a player worth a damn."

*cough* Eckstein *cough*

"This highlights something I think is wrong with a lot of sabremetric analysis (at least the stuff in the public domain) - its just not that sophisticated from a statistical perspective. Its pretty easy to adjust for changes in time like like this, but the Sabre guys aren't doing it yet. "

Actually, a lot of them do. For most predictive SABR stuff I see on blogs, they take the publicly available numbers and run them through adjustments to correct for age or position or era or what have you. The main numbers on Fangraphs and BB-ref are starting points that get messed with a lot. Nate Silver used to do a lot of this before he got called up to the majors.

It takes a lot of data for their to be information in that couple of extra hits a week, or an extra step in the field - but I don't ever see that part of the analysis."

You don't ever see people talking about the margin of error and complaining about small sample sizes? RogerB's upthread making sure that folks know that defensive stats are noisy and need about three years to be predictive.

But luckily, there's a huge amount of data in a given season — most guys have a couple hundred plate appearances, and pitchers throw thousands of pitches. And also handily, the actual wins and losses end up as easy checks for the predictive power of a lot of the more complicated stats.

I wonder how much of the aversion to advanced stats talk in baseball is left over from the fact that the traditional baseball statistics pretty much suck ass; batting average, ERA and pitcher wins are all pretty terrible ways to measure a player's worth.

As for the lists, they're usually not something that anyone takes seriously (except Yankees fans).
posted by klangklangston at 9:06 AM on August 31, 2011

it also completely gives up on catcher defense and basically just gives them all 0s for everything except maybe throwing out basestealers.

That sounds pretty much right to me! Compared to a 3b or outfielder, you have to figure that a catcher's defense doesn't have that big of an effect, and to the extent that it does, it would be via throwing out basestealers (or, indirectly, by inducing runners on 1b not to steal, an effect I'm not sure dWAR measures.) Of course this does leave out the ability to call a good game, but I'm not sure I've ever seen anyone try to quantify this and I'm not sure it can be done.
posted by escabeche at 9:12 AM on August 31, 2011

you have to figure that a catcher's defense doesn't have that big of an effect,

I don't know. Catcher's errors tend to directly cost you a run more often than any other. Usually, the only time a catcher is involved in a play is when the ball is very close to the plate, and the usual time for that is not a bunt but a potential play at the plate, and a catcher's error here will result in at least one run. Bunts, btw, are more often handled by the pitcher or third baseman, since laying down a bunt right in front of the plate is both hard and a great way to be thrown out. The ideal bunt has both a corner infielder, the pitcher and the catcher trying to get it...because they're all the same "distance" (really, the same time lag) away from the ball.

There are occasional throwing errors on steals, but usually, it's either a fielder error or the ball hitting the baserunner and going somewhere else. Catchers than can't reliably throw to 2nd just don't make it to the bigs as catchers. They are, admittedly, often *spectacular* errors when they do happen. Another class is the "throw for the force, even though the guy is coming in" throwing error, that one is both spectacular and instantly costs you a run.

Finally, there are passed balls, the cost of those is a big "depends" -- with a deep backstop, they're worse-- but the fact that they're guarding the plate makes the cost of a catcher's error pay out instantly.

Throwing out base stealers is easier to track. First, measure the success/fail percentage for that catcher, then measure chances per inning. A very good catcher in this regard won't have many people trying to steal on him, a very bad catcher will have both a worse success/fail percentage and a lot more people trying to steal. An inconsistent catcher will show a disparity, probably a worse success ratio, but not as many chances (are you going to try? He might be on today!)

Of course this does leave out the ability to call a good game

Very hard to track. You'd think it would be easy -- compare pitching records with different catchers, but given that most teams don't have many catchers, and many catchers will always catch certain pitchers, you just don't get enough data. You might be able to rate entire careers (never mind the noise factor) but it will still be hard to get enough pitchers throwing to enough catchers to make it a useful measure, and I'm pretty sure that it'll be impossible to rate any but the most veteran catchers -- and at that point, there must be *some* reason they've been in the majors that long, right? If they really can hit and really can't catch, that's what Left and Third are for, right? (And DH for those who play fake baseball.)
posted by eriko at 9:39 AM on August 31, 2011

For my money the worst fielder I've ever seen is the Mets' own Daniel Murphy, who often has a kind of deer-in-the-headlights look to him at any of the positions he's failed to master. There have been times I've actually feared for his safety out there.
posted by AJaffe at 9:42 AM on August 31, 2011

You don't ever see people talking about the margin of error and complaining about small sample sizes?

sure, but I'd rather see a more thorough discussion of the topic. I mean at some point someone ran the regressions to show the relationships as well as the t-stats et al , but now we just get the numbers presented to us w/o the context of testing the hypothesis. I personally don't know when there is information in the data and when there isn't. I mean there might not be much information in saying a guy is .5 std deviations from the mean vs a guy who is .75 std deviations from the mean, but in OPS terms .810 is a lower number than .840 - I don't know if there is information in that? how many PA does there need to be for there to be information in that?

You can say "Wel the data says this, but the margin of error is x" but people always assume the data is more precise than it is. That's basic human nature.
posted by JPD at 9:43 AM on August 31, 2011

in OPS terms .810 is a lower number than .840 - I don't know if there is information in that? how many PA does there need to be for there to be information in that?

For OPS, it takes about 500 PA (roughly two-thirds of a season for a starting player) to iron out enough noise that it's a reliable indicator of ability. When Samples Become Reliable; How Do You Measure a Player in a Year?
posted by RogerB at 9:53 AM on August 31, 2011

"Take away good numbers and leave the intangibles, and you'll never have a player worth a damn."

*cough* Eckstein *cough*

You're proving my point. Eckstein isn't a player worth a damn.
posted by grubi at 10:04 AM on August 31, 2011

For OPS, it takes about 500 PA (roughly two-thirds of a season for a starting player) to iron out enough noise that it's a reliable indicator of ability.

that's not really the answer - how many PA allows you to say an .840 OPS is different from an .810 OPS. The number of PA need for that should vary depending on where you are in the distribution.
posted by JPD at 10:05 AM on August 31, 2011

"You're proving my point. Eckstein isn't a player worth a damn."

But he has a four gallon heart and sells Grit to earn inflatable rafts.
posted by klangklangston at 10:11 AM on August 31, 2011 [1 favorite]

He did lead the league in hustle five years straight.
posted by grubi at 10:15 AM on August 31, 2011 [2 favorites]

that's not really the answer - how many PA allows you to say an .840 OPS is different from an .810 OPS

I guess I have no idea what you mean by "is different from." If you're using the 810 and 840 as a record of events that occurred in past baseball games, then any difference between them is meaningful and fair to compare. If you're using them as a predictive indicator of two players' likely hitting skill, it works fairly reliably provided you have 500 PA or more of data, as the articles I linked say in more detail.
posted by RogerB at 10:16 AM on August 31, 2011

I guess I have no idea what you mean by "is different from." If you're using the 810 and 840 as a record of events that occurred in past baseball games, then any difference between them is meaningful and fair to compare. If you're using them as a predictive indicator of two players' likely hitting skill, it works fairly reliably provided you have 500 PA or more of data, as the articles I linked say in more detail.


but no matter what there is some confidence interval surrounding those numbers, and it isn't constant. It takes more PA to show .840 is different from .810 than it does to show 1.000 is different from .750
posted by JPD at 10:31 AM on August 31, 2011

but no matter what there is some confidence interval surrounding those numbers

No — if you're using them retrospectively, as a measure of what happened in past games, then there is no confidence interval. In baseball statistics there is no such thing as observation error.
posted by RogerB at 10:55 AM on August 31, 2011

I misspoke. I didn't mean confidence interval.
posted by JPD at 11:08 AM on August 31, 2011

And I'm speaking of them as predictive of the future value of the player
posted by JPD at 11:09 AM on August 31, 2011

My point basically is "Saying someone with an .810 OPS is a worse hitter on a going forward basis that someone who has an .840 OPS" takes more observations than making the same comment about a .750 and a 1.000 OPS batters.
posted by JPD at 11:19 AM on August 31, 2011

My point basically is "Saying someone with an .810 OPS is a worse hitter on a going forward basis that someone who has an .840 OPS" takes more observations than making the same comment about a .750 and a 1.000 OPS batters.

Well, yeah. So you'd do it based on a larger sample size. If player A has an .810 OPS over the past 500 PAs, and an .850 OPS for his career, and player B has an .810 OPS over the past 500 PAs but a .700 OPS for his career, you'd do wise to bet on player A going forward.

Another factor is age. All other things being equal (which, obviously, is rare, but you've gotta form some baseline), you'd take the next few seasons from a 25-year-old with an .810 over those of a 35-year-old with an .840.

Then you throw in other adjustments for things like body type, position, hitting profile (where do his batted balls land? What's his BABIP? Does he have 'old player's skills'?), and from there you can get a pretty good idea of a given player's projection. Of course, that projection could be wrong, but this is all about probabilities and confidence intervals.

What I've just described, of course, is what all of the better player projection systems do.
posted by breakin' the law at 12:09 PM on August 31, 2011 [1 favorite]

Simon!: " Barry Bonds was a great, great left-fielder pre-steroids, legitimately all-time great"

Too bad he didn't listen to Andy Van Slyke and play a little shallower in the ninth inning of Game 7 of the 1992 NLCS.

Not that I'm bitter.
posted by Chrysostom at 9:02 AM on September 3, 2011

As a side note, Bill James has an excellent little piece on why Verlander deserves the MVP. I'm not sure it deserves its own post, in part because I'm a total Tigers stan, but I thought this thread would be a decent place to put it.
posted by klangklangston at 10:16 AM on September 3, 2011