A quick review – have you heard about this book? Put out by the Baseball Prospectus – stats nuts who apply their science to innovative baseball analysis. Been a best-seller, near as I can tell.
I am about ½ way. So far it is fine – it geeks out big time on the statistical analysis as it tries to tackle interesting questions such as do lineup orders matter, does the catcher really have an effect on their pitching squad, did Jeter deserve a Golden Glove (I didn’t know that he is generally considered a sub-par defensive SS). However, the book fails where books like Moneyball succeed – it focuses on the numbers, which is fine, but does so at the expenses of the game. It almost feels like baseball as a game is being used as a random number generator to feed their statistical analysis. They clearly must have a passion for the game or they wouldn’t devote so much energy to it, but that doesn’t come through in the book. It reads like a collector’s book where the author cares more about the act of collecting than the useful attributes of the things they collect. There is little sense of the characters and stories of the game or the insertion of real-life interpretation – they articulate why bunts and late-inning subs and other standard changes are over-used, but the book doesn’t make you hear the crack of a bat or understand what should really be done when the chips are down in the bottom of the ninth.
Makes me wonder of the sabermetrics approach to management needs a bit more tempering with real-world Jim Leyland-type know-how.
The theories are interesting but as in removes the hunch and clutch and intangibles from the game for cold calculations, I have always found sabermetrics to be less than useful.
The very example you provide, help show they are missing part of the picture. According to Saberheads, Bill James and Boston Fans, Jeter is a sub-par defensive Short Stop. According to those who watch him everyday, other players and sportswriters he is above average and does some things no other shortstop does. The problem is that what he does has no measurement.
On the daily scale he goes back on balls better than any shortstop and does this great leap, grab, spin and throw in one motion that leads to a lot of extra outs.
There is also his history of occasionally making impossible plays like the flat out flying leap into the stands against Boston 2 July’s back and the famous flip to Posada in the Playoffs against Oakland to get Jeremy Giambi at the plate.
None of this gets measured.
Off the Jeter defense of his defensive ability.
The stats never take into account that in a line up, a lead off mans job is different from a clean-up hitter.
Another problem, if one cleanup hitter bats .300+ with 40 homers and 110 RBIs he might not be more valuable than the .260 hitting 30 homer guy with the same 110 RBIs but this second apparently lesser play has had twice as many late inning RBIs that put his team ahead to win the game.
Classic examples are Ortiz, Edgar Martinez, George Brett were/are all extremely clutch players. Carlos Delgado put up similar or better numbers with the Jays but was not a more valuable player.
I thought I would be interested in the book, since I started reading Bill James in the 1980s, but I found it to be disappointing.
In a way, admittedly, it’s the logical conclusion of what James started. The more information one has on every pitch and every play, the better the modeling will be. This is what Project Scorecard started out to be before it disappeared under the squabbling.
But one of the things that James did brilliantly was to take the stats and turn them into meaningful nuggets of information. And produce results that could be applied by baseball people as well as fans to understand who was being productive and what should or shouldn’t be done during a game.
I didn’t see any of that in this book. I kept looking for how the number crunching could be applied by a manager during a game, but it wasn’t there. Knowing that in seven situations out of ten on a 2-1 fast ball in a pitcher’s park with men on second and third and one out it’s better to hit a ground ball off a left-hander with fewer than two seasons of experience is strictly Monday morning quarterbacking, to mix a metaphor.
Most of the less detailed info the book was old stuff to anybody who followed sabermetrics since James, and it was so boringly, confusingly, and obsessively written that I doubt most newcomers would ever be able to plow through it and get anything worthwhile.
I’m glad I got it out of the library and didn’t spend money on it. Project Scorecard is not Project Spreadsheet.
As for clutch players… Well, James has backed off his dogmatic assertion that there’s not such thing as a player who is consistently measurably better in the clutch than another player of equal skills. What he says now is that there is no evidence that can be given to support it, but that statement is not quite the same thing. Mostly, however, anybody who talks about players being good in clutch situations suffer from memory bias. A few big plays create a reputation that numbers consistently fail to back up. However, since it takes so few big plays to be important, it’s hard to measure whether they are statistically meaningful or not.
There’s no question, though, that all of baseball uses numbers and ideas developed and popularized by the sabermetricians and that you can tell almost immediately the teams that do so badly and the ones that do so well. Sabermetrics is not some magic bullet that creates World Series winners. At best, it gives teams an edge of a few games a year over lesser strategies. Sometimes those few games make the difference, sometimes not. And the difference is far less with short series, which is why Oakland’s use of sabermetrics saved the franchise by putting consistent 90+ game winning teams on the field, but haven’t yet advanced in the playoffs.
Sabermetrics demonstrably works and even the oldest of fogies in baseball have gotten it to some extent. Now if the fans would…
The line that cracked me up the most so far (I am also about half way through the book) is the one about how casual baseball fans might think that Jeter’s first name is “pastadiving” because of how often announcers mention that a groundball has gone past a diving Jeter.
The big problem with sabermetrics is that is great at analysis of past seasons, but piss poor at explaining, and even worse at predicting.
For instance, sabermetrics can pinpoint that Dwight Gooden never was the same pitcher after 1985, and certainly can point to the numbers to prove it. But it cannot explain why, other than mumbling about falling back to the mean. The fact that he was using drugs, and that his pitching coach messed up his mechanics trying to teach him a change-up is not shown at all in the stats and you cannot deduce this from the numbers alone.
Again, Carlos Beltran’s numbers last year would have been inexplicable solely by numbers, but the fact that he was playing injured most of the season does not show up in the stats.
As for clutch hitting; I’m glad to hear that James is inching away from his position on it. He once denied it existed by saying, essentially, “If you could give me one reason why a player would hit better in clutch situatios, then maybe I’d believe it.” That pretty much discredited sabermetrics for me, since the reason is blindingly obvious: personality. But personality has no place in sabermetrics, and that’s a crucial flaw.
Well you know he had to back off when he went to work for Boston, Ortiz is the Babel Fish of Sabermetrics. His clutch hitting goes beyond any numbers that Bill James could produce. When faced with indisputable proof in your own stadium, it is hard to remain in steadfast denial.
The thing is, a sabermetrics guy would tell you that all these things do show up in the numbers, or they don’t exist – for instance, if Jeter’s so good at getting extra outs at short, all other things being equal, he should register more assists and more putouts than the next guy. Makes sense, right? But, after as much extraneous information as possible is controlled for, it turns out that Jeter doesn’t really make all that many plays. Ergo, he must not be as effective as he’s supposed to be. And that, in a nutshell, is what statistical analysis is all about. It’s definitely frustrating when the numbers go against what you “know” is true, I agree, but maybe that makes it that much more worthwhile. I know there are more rabid sabermetrics geeks than I on this board, so I’ll let them handle the more specific arguments (I kind of grudgingly accept the validity of the whole enterprise, myself), but the important question to ask is, if there’s a phenomenon in place, how is it possible for it not to be measurable? If it exists, it should affect wins and losses, which means it should affect runs scored or runs allowed somehow. If you can take a semi-scientific approach to finding out where and how these phenomena occur, well, why not?
But that’s simply not true. It’s plainly false, as a matter of fact. I have watched thousands and thousands of major league baseball games, and played, umped, coached and watched thousands more games at other levels, and Jeter does NOT go back any better than the average major league shortstop. I think I’m a reasonably well versed baseball fan, and I’ve seen a hell of a lot of major league shortstops, and Jeter just does not look very impressive to me. Maybe it’s because I’m in the Canadian market and we’re not bombarded with the Yankees-worship or something, or because I know a lot about both the physical sport and the stats, but to my eyes is it visually very obvious that Jeter’s an average shortstop. He used to be BAD. He’s better than he used to be, IMHO.
The thing is, he looks very good if you never watch any other shortstops, because of course all MLB shortstops are talented athletes. IF you are told Jeter is a great shortstop and you then watch him intending to validate that belief of course he’ll look great; ALL MLB shortstop are terrific defensive players. They’re the guys who were too talented to play first base or the outfield. But because of that, Jeter is middle of the pack among MLB shortstops, because they can all do amazing things. Jeter is no better than the current Blue Jays shortstop, John McDonald. He’s no better than Miguel Tejada. He’s no better than Michael Young. He’s not nearly as good as Orlando Cabrera. He’s not as good as Rafael Furcal. He’s not as good as Juan Uribe. Those guys all do the same things Jeter does and some do it better. (Defensively. Jeter is a much greater hitter than most shortstops.)
The whole point to sabermetrics is not to take the fun out of the game but to try to deal with these sorts of things honestly. Jeter gets raves about his defense because he plays for the Yankees and because he’s really marketable and famous. He simply is not a great defensive player. It’s not just that the statistics say so, although they do say so; he doesn’t LOOK great (as compared to other major league shortstops.)
Yes, he sure did make a great play to throw Jeremy Giambi out at the plate that one time; saying that makes Derek Jeter a great fielder is very much like saying that Ed Sprague was an all-time great home run hitter because he hit that huge homer off Jeff Reardon in the 1992 World Series.
I am not suggesting that the Baseball Prospectus guys are perfect. They have their own biases, to be quite honest, and have shown themselves to be less than professional journalists. But trotting out this old “The stats don’t tell you the whole story, I know things are different” line is just so tired and silly. You’re going to find yourself on the wrong and illogical side of a LOT of arguments.
What does this even mean? Surely leadoff hitters can put up good stats? Stat heads loved Rickey Henderson, and have long argued that Tim Raines should be a Hall of Famer. How do statistics hurt leadoff hitters?
.
Pop Quiz!
So far in 2006, has David Ortiz, in the late innings of close games:
A) Hit better than he does the rest of the time?
B) hit worse than he does the rest of the time?
C) Hit about the same?
The answer is… ready for it? B. Ortiz has not been as good in the late innings of close games as he normally is. Over the course of the season, Ortiz is hitting .285 with 35 dingers, 99 RBI in 389 at bats; he has a .997 OPS. In the late innings of close games he is hitting .242 with 3 homers and 8 RBI in 33 at bats.
Whether you want to believe Ortiz is a clutch hitter or not is almost entirely dependent on what statistic you want to look at. Over the course of his career he has no tendency whatsoever to hit better in “clutch” situations. He hasn’t hit better in the late innings of close games, hasn’t hit better with runners on as opposed to not, so on and so forth. Even his career postseason stats are not substantially superior to his regular season norms.
Incidentally, Edgar Martinez was a career .266 hitter in playoff games. Why would you think he was a great clutch hitter, when his numbers actually got a bit worse in the clutchest games of them all? Well, because of your subjective memory; you remember the 1995 ALDS when he destroyed the Yankees, drove in 7 runs in that one game. You DON’T remember that in the ALCS that year he went 2 for 21 and was a big reason why his team lost. You don’t remember the 3-for-20 against the Yankees in the 2001 ALCS, when he helped a 116-team win choke the pennant away. That’s why subjective opinions must be challenged by the facts.
I do not usually parse posts, but I see no other way to reply.
Actually I have been watching baseball since 1969, with an understanding of what I was watching since around 1972. I only played a little and I was a decent 3rd base coach. Thank you for dismissing my ability to judge a baseball player based on your disagreement with my view. I said Jeter was not Sub-par, I did not say he was great. I did say “According to those who watch him everyday, other players and sportswriters he is above average and does some things no other shortstop does.” I have rarely seen a SS that goes back on the ball close to as well as Jeter. I have seen a lot of very good defensive Short Stops eat up far more ground balls than Jeter. I did not claim otherwise. Did he deserve the Gold Glove last, no, but Gold Gloves are commonly given for the wrong reason to the wrong player. I do believe he is an above average Short Stop.
The impossible play examples were nothing more than the intangibles that do not and cannot be measured. Dismiss them as you wish. Again I did not say Jeter was a great fielder. Re-read my post and see if I said that. I mentioned he does 2 things better than other shortstops. You disagree on the one.
In Relative value, a good lead-off hitter is almost always valued lower than a good 3-5 hitter as in Sabermetrics it appears that OPS is overrated. Ricky Henderson is probably the greatest lead-off hitter in the history of baseball, you might as well have brought up Ty Cobb, the other candidate. It is harder to find a very good lead-off hitter than a very good 3 or 4 hitter. Sabermetrics and Fantasy leagues undervalue the contributions of a lead off hitter in my opinion.
.
Are you saying that from 2003 to 2005 Ortiz was not a clutch player, exhibiting more big hits than any other player in the AL East at least? I would like to see the numbers you are using, just pointing to this season, hardly proves anything. I sincerly hope he has dropped back to normal levels. The Yanks face Ortiz in Boston for 5 games in August. These five games could go a long way towards deciding the season.
In 2001 Edgar was winding down his career and playing on one leg, speaking of facts that hide from the stats. From 1996 to 2000 he hit .600 vs Rivera which does indeed color my perceptions as no one else hit Rivera well except I think Rafael Palmiero. So maybe Edgar was only clutch against the Yanks, I could well be wrong on that one, except of course Seattle fans will tell you how clutch he was and more likely to win the game than A-Rod, Buhner or even Griffey.
For the Yankees we had Paul O’Neill, I am sure you can provide stats that show he was no more clutch than any other player of similar ability, if you believe those stats we will never agree on how to judge baseball anyway.
Ty Cobb is a very strange candidate indeed, since he was not a leadoff hitter. Cobb batted third for most of his career.
We’re not talking about fantasy leagues.
How does sabermetrics “undervalue” leadoff hitters? Can you provide any rationale at all for this claim? Sabermetrics, since its very inception, has been arguing the leadoff hitters are tremendously important, more important than they traditionally have been given credit for; after all, the acceptance of on base percentage has long been one of the big sabermetric pushes. What gives you the idea sabermetrics undervalues leadoff hitters?
As to your claim that leadoff hitters are harder to find than #3 or #4 hitters, I’m not sure that’s true. Teams tend to put their best hitter at #3 or #4 just because they always have, because they believe that is the best place to put them. It’s not that #3 hitters are easier to find - they aren’t easier to find - it’s that a team will find a hitter and put them at #3 if they’re good. A player is not born a “leadoff hitter” or a “cleanup hitter” - he’s a hitter, and then he is put in the role a team feels he’ll suit.
A few years back I remember an announcer commenting about how it seemed most teams had found a good #1 starter but couldn’t seem to find a reliable #5 starter. Well, DUH. Talk about putting the cart before the horse. Everyone seems to have a decent ace because even a bad team can find one good starter, so he becomes the #1, but even a good team has trouble finding five good starters, and the worst one’s always the #5. The ability defines your role, not the other way around.
Regrettably, situational numbers from previous seasons are hard to find, so I would need some time to get this. However, he certainly hasn’t been tearing it up in the postseason. IIRC we had a discussion about this last year and Ortiz’s batting average in some clutch situations was higher in 2005 than his overall batting average, but had not been in 2004. Again, though, we need the numbers. But still, ain’t it funny that he’s not any better in clutch situations this year? Perhaps that’s because - as every single study ever done on this issue has shown - that hitting in the clutch is not a consistent skill.
And just point to 2005 does? You think he’s a great clutch hitter, let’s see the evidence; it’s your claim, you prove it. If you want to see the numbers I’m using, consult mlb.com. Search for Ortiz, look at his stats, and there’s a link called “Splits” that will show you thinks like his numbers in C&L situations, with runners on, leading off an inning, etc. It’s only got 2006, though; for previous years we’ll have to consult an almanac or something.
Yes, I am sure they will tell me that, but it’s simply not true. He wasn’t overly clutch in the playoffs. You can say he was playing on one leg in 2001 but it sure didn’t hurt his regular season numbers; he hit .306 with 23 homers and had the second highest on base percentage in the league. He wasn’t playing on one leg during that 2-for-21 in 1995, or his 3-for-16 against the Orioles in 1997. Those performances all count too, don’t they? If he was such a clutch player, why don’t his playoff numbers reflect that?
Willie Mays in his career batted .247 in playoff games with one home run; does this mean Willie was a choker, or d’ya think maybe it’s just a fluke? Mike Schmidt was the MVP of the 1980 World Series, and then in the 1983 World Series he was 1 for 20; did he just forget how to be clutch?
I mean, step back and take an honest look at what’s already happening here; you’re dodging and weaving in just your first reply, trying to justify why Edgar was a great clutch hitter when I provided facts that he did not hit any better in the most important games of his career than he usually did. That’s what sabermetrics is about; examining the issue honestly, rather than coming to a subjective opinion and then rationalizing it.
Now, I’m not saying you might not be able to find stats that prove I’m wrong. Perhaps you will discover that throughout his regular season career, Martinez was remarkably better in clutch situations. I would find that very convincing. What isn’t convincing, however, is saying “Well, you have shown me facts that say he’s not a great clutch hitter, but I don’t care about that. He was. Neener neener neener.”
So how are you judging Paul O’Neill’s ability as a baseball player? I’m asking that question honestly; I watched Paul O’Neill play baseball for fifteen years and he was a very good player indeed, but as I personally recall he was just as good in non-clutch situations as he was in clutch situations. As a matter of fact, that’s one of the things I remember about O’Neill; his consistency. As near as I can tell you are not, in fact, judging his ability at all; you’re simply making a decision and then refusing to apply reason or facts to anything you’ve decided.
O’Neill played exactly as well in the playoffs as he usually did - in fact, to an almost uncanny extent. His average, OBP and slugging average:
My favorite standard for measuring clutch hitting is future performance. That is, when someone claims that Ortiz has some strange mystical ability to hit well in the clutch (as opposed to being a hell of a hitter who gets a lot of at-bats in clutch situations) I’ll just offer to bet that person (by any reasonable and mutualy agreed-on statistical yardstick) on Ortiz’s performance in the NEXT 20 or 30 “clutch” at bats. If it’s a true ability, then Ortiz will continue to do it. But if it’s just small sample-size, combined with normal expectations, then he won’t do anything special, normally.
I’ve also made this bet for Jeter in the post-season, and have done pretty well there.
There’s something sad about the fact that ESPN has a better stat database than Major League Baseball.
Looking at this situational splits, it’s immediately apparent that you can argue Ortiz is or is not a clutch hitter depending on what numbers you want to draw. His numbers with runners on are essentially identical to his numbers with the bases empty, and his numbers when men are specifically in scoring position are, again, the same. He’s a bit below his norms with a man on third and less than 2 out; close and late, he’s above his norms (we’re not counting 2006 here.) He slumps when leading off an inning (an unsung clutch situation) but has hit well with the bases loaded. Throwing in his 2006 numbers, I just don’t see a lot of convincing evidence the guy turns it on in the clutch.
Besides, if a player can hit better in certain situations, why not hit that well all the time? For all you know, that first inning at bat could be the key moment of the game.
As an interesting aside, the AL team he has the most trouble with? Oakland. The AL team he hits the best? Minnesota. Dunno if that means anything but it’s neat.
RickJay: Is there a white flag smilie? I surrender.
I subscribe to the Yogi Berra & Don Zimmer School of Baseball, so I am placing myself with the Dinosaurs. I cannot argue with Statistics why Paul O’Neill was clutch and if you do not even accept Ortiz as clutch, there is no hope for me.
I certainly don’t mean to trash Derek Jeter. I think he HAS improved his defence (and the numbers back that up) and that is very impressive for a shortstop in his early 30s; it’s very unusual for a player to improve defensively after he’s 30. Jeter is in a position where he could rest on his laurels, but he has apparently chosen to work harder than ever, and you’ve got to respect that. On top of that he’s an excellent hitter and one of the best baserunners I’ve ever seen. He’s headed to the Hall of Fame, make no mistake about it, and he deserves to.
Look, I understand what you’re saying from a statistical point-of-view and there probably isn’t anything to the whole “clutch” thing. But I can honestly say that as a long time Sox fan (25+years) who watches the team play >150 games a season, I have never seen anyone as consistently exciting when a game is on the line as Ortiz.
Also, I’m not sure that “close and late” is a great measure of clutch performance since some games are (at least psychologically) more important than others. Last nights’ situation, for example, had a little more weight behind it since the Red Sox were in danger of falling to 2nd place behind the Yankees (by % points), who already have a slight edge in momentum coming off some trade deadline coups. Like anything, there are certainly qualitative issues that must enter the equation.