People Sayin' Pop Fell Off

John Hollinger agrees with me. It's an insider only item, but let me summarize the key points:

• "You'd be surprised to learn that San Antonio, not Dallas, has the best point differential in the league, at +7.9 points per game. (Actually, Dallas is third at +7.7 -- Phoenix also noses in ahead of them)."
• "San Antonio won a team-record 63 games a year ago with a point differential of +6.8. This year they're more than a point per game better (They need to be, too: Those 63 wins didn't do them any good in the playoffs); they just haven't been as fortunate in close games."
• "And while the Spurs are disappointed because they're "only" third in Defensive Efficiency instead of their usual perch at No. 1 (they've been there five of the past six seasons), this is the best offensive team of the Popovich Era. San Antonio ranks sixth in the league in Offensive Efficiency, just 2.8 points per 100 possessions behind the Mavs."
• "They're doing this while playing their scrubs for much of the game. No Spur is playing more than 35 minutes per game; Tim Duncan leads the team at 34.6. Tony Parker is only playing 33.0 minutes, Manu Ginobili a measly 27.8. No team is playing their starters less, but I'm going to go out on a limb and say that the Spurs' big three will see a lot more action during the playoffs."
• "The Spurs' problem isn't age or a lack of fire or any of the other ideas trotted out in recent weeks. It's that they're 5-10 in games decided by five points or less, while the Mavs are 12-2."

I concur. The story people should be writing isn't "what's wrong with the Spurs" but why is Gregg Popovich giving so unconcerned about the regular season that he's giving his stars so little playing time. That's an actual reportable factual issue, and if we knew more about his thinking then maybe we could say something about whether or not it's a good idea.

Comments

"It's that they're 5-10 in games decided by five points or less, while the Mavs are 12-2."

Hollinger, who I normally think is pretty smart, is waaaay off the mark in his fetish for point differential.

It's precisely San Antonio's lousy record in close games that should be an enormous warning flag that the team will very likely founder in the playoffs.

I've said it before and I'll say it again, W/L is a much more predictive statistical measure than point differential.

Posted by Petey | February 26, 2007 10:43 AM

Last year, it was Phoenix's lousy record in close games that made it clear they were going nowhere in the playoffs.

Posted by Petey | February 26, 2007 10:49 AM

The Spurs aren't going to get home court advantage for having the best point differential.

In a series against either Dallas or Phoenix, the home court is a much bigger edge than 0.2 ppg of differential.

Posted by quietstorm | February 26, 2007 10:51 AM

Agree with Petey about the importance of the lousy record in close games.

actually want to like Hollinger, and I know that I've found previous columns of his to be enlightening. But you'd think that people devoted to the belief that we have sufficient technical tools to understand basketball through statistics, and that such understanding is better than prior understanding...you'd think those people would find it worthwhile to track whether or not (a) stats gurus like Hollinger were willing to make falsifiable claims, or (b) they were, on average, more right about anything about the season than anyone else (like, say, Greg Anthony).

Posted by SomeCallMeTim | February 26, 2007 10:53 AM

A minor nit - why do sports commenters call a difference in scores a differential? That word would seem to more properly apply to discussing the rate of change of the scores. For that matter, why do they call themselves commentators instead of commenters? What does it mean, to commentate?

Posted by dob | February 26, 2007 10:55 AM

"For that matter, why do they call themselves commentators instead of commenters? What does it mean, to commentate?"

And what's the deal with airline food?

Posted by Petey | February 26, 2007 10:56 AM

I'll add my voice to the Petey/SCMT chorus. I've always been annoyed by the stats geeks' unwillingness to acknowledge the intangible psychological factors that explain why some teams and players and coaches consistently rise to the challenge and others don't. For instance, Dean Oliver, who is probably the godfather of statistical analysis in basketball, attributes the Rockets' 1995 championship to "luck." Ridiculous. That team defied the odds because so many of its key players -- Olajuwon, Horry, Elie, Cassell -- had BALLS.

Having said that, regular-season balls are very different from playoff balls, and so I don't read quite as much into the close-game stats as Petey does. We all remember what happened in Game 3 of the Finals last year.

Posted by Andrew | February 26, 2007 11:08 AM

Has anyone tracked, say, the last several NBA champions, and their record in games decided by five points or less?

Posted by Steve | February 26, 2007 11:38 AM

The thing with the record at close games is that it doesn't hold from season to season. You'll see teams win a lot of close games and the next season, with almost exactly the same lineup, lose a lot of close games. Hollinger (and several other people, too) has actually studied this issue a lot, and nobody has been able to find a team that consistently wins most of its close games, year after year (or overperforms its point differential, it's the same thing). I'm a bit hesitant with regards to the Spurs record for various reasons, but I'm not ready to say that their inability to win 3 close games so far (with that, their record would have been a more normal 8-7) means they are over the hill.

Posted by Carlos | February 26, 2007 11:39 AM

Pop has stated over and over again that he is purposefully resting his players for the playoff push. Homecourt hasn't been all that important to his championship teams. In 2005, they had to go through Phoenix, and in 2003, they had to go through LA. Last year, Dallas went through SA, and Miami went through Dallas, etc.

The Spurs are well aware of the grind that is the NBA playoffs. You have to have enough gas to win four seven-game series. Dallas, a super deep team, didn't have enough last year.

W-L is of course a better indicator of team quality. But there is also usually a strong correlation between point def. and W-L. It isn't there for the spurs for some reason. And dallas has been crazy clutch all year.

Posted by MP | February 26, 2007 11:46 AM

It's precisely San Antonio's lousy record in close games that should be an enormous warning flag that the team will very likely founder in the playoffs.

I've said it before and I'll say it again, W/L is a much more predictive statistical measure than point differential.

Well, this should be easy to figure out. Can Matthew get Dave Berri or someone to run a regression to see if W/L or point differential correlates better with playoff success over, say, the last 20 years? We all know Matthew has the pull to be able to get that, right?

Posted by Al | February 26, 2007 11:51 AM

I think we can guess what Pop is thinking. The (Western Conference) playoffs will be absolutely GRUELING this year and he needs his top players to be fresh. He's also cultivating a deeper bench which he'll need for close 7-game series. Once the Mavs and Suns started running away with wins he probably decided to forego trying to keep up and just keep fresh.

Posted by eriks | February 26, 2007 12:06 PM

'll add my voice to the Petey/SCMT chorus. I've always been annoyed by the stats geeks' unwillingness to acknowledge the intangible psychological factors that explain why some teams and players and coaches consistently rise to the challenge and others don't

That's not quite my complaint. I'm not a huge fan of psychological factors. It just seems to me that a lot of the stat geeks find what can be measured and then turn correlations into some sort of Higher Truth. They rarely seem to make hard predictions that deviate from more standard eyeball analysis and, when they do, they eagerly avail themselves of all necessary caveats.

The fundamental claim seems to be that stat usage can predict certain things better than eyeball analysis. But I remain unclear what those "certain things" are. What extra insight do they think they have?

Posted by SomeCallMeTim | February 26, 2007 12:16 PM

It just seems to me that a lot of the stat geeks find what can be measured and then turn correlations into some sort of Higher Truth.

Exactly. The sport I waste time on is baseball, which is more easily divisible into discrete units of statistical analysis, but hte same problems arise there.

It is evidently true that there is a strong population-level trend against maintaining an unexpectedly strong or weak record in close games in major league baseball. The same appears to be basically the case in professional basketball. This is useful information - we should be rightly skeptical of claims that a certain team is clutch or unclutch or better or worse merely because of their record in close games.

However, that does not at all mean that every team is exactly as good as every other team in close games. It just means that such a skill is difficult to discern through statistical measures. It just means that if we care about the topic, we need ot learn as much as we can at hte individual level - how does hte coach run the last few minutes, who gets the ball, what strategies change, and on and on.

I can tell you that hte Red Sox under Grady Little underperformed hteir expected wins by ~5-7 games a year in 2002 and 2003, and capped that by losing hte most important close game of their recent history up to that point at the end of the 2003 playoffs. Now, becuase Little was fired, because new personnel came in, becuase of a million factors, it's nearly impossible to make a purely statistical case for the peculiar characteristics of the 02-03 Red Sox. However, I was there. The use of the bullpen, the composition of hte bench, the relationships between manager, pitcher nad pitching coach, and a whole lot of other things conspired to make hte 2003 Red Sox into the team that would lose that series more often than a statistical measure would say they should. (Luckily for me, "stat geek" GM Theo Epstein made all the necessary changes the next year, and hit the right run of luck, and, well, yeah.)

Posted by DivGuy | February 26, 2007 12:32 PM

Carlos is right. Much like in baseball, "clutchness" in basketball appears to be largely illusory. A team's performance in close games fluctuates pretty randomly, just the way it would if it were all luck.

This confuses me, because when I watch, it certainly appears that some players and teams are "clutch." But rather than conclude that my impressions must be right, I'm reserving judgement. In fact, the very fact that "clutchness" appears to be real makes me doubt that it is. If I can't come up with better logic than, "well, Dallas has this special magic mojo that only takes effect late in a close game, and San Antonio doesn't," then my impression is probably wrong.

Posted by too many steves | February 26, 2007 12:37 PM

And, re: the balls of Horry, Cassell, etc. Perhaps we confuse cause and effect -- we think the 95 Rockets were especially ballsy because they got lucky and won close games. That's what gave those guys the "clutch" reputation.

Robert Horry is the real reason I suspect "clutchness" isn't real. He's the very definition of a player who has a great rep because of a few lucky shots. Watch him in the 4th quarter of a big game sometime -- he just stands at the 3-point-line and fires 'em up. Most of the time he misses, but of course, sometime he makes 'em. Nobody remembers the misses, or the times he fucks the offense by standing at the 3-point line.

Posted by too many steves | February 26, 2007 12:40 PM

Resting your starters in the regular season and giving up home court advantage is a strategy that used to work well for the Lakers, and will probably prove to be effective for the Spurs as they get older. So I'm not surprised that Pop is trying it out.

I think late game performance is important and varies from team to team, but it is very situation specific. Remember the Miami-Dallas series? That had little to do w/ each team's clutchness (except for some strangely effective foul shooting from Shaq), but had a lot to with the league's tendency to reward all star level guards by calling fouls on anyone near them at the end of games in big playoff series. That phenomenon probably peaked in that series. I think winning close games in the regular season is a good sign, but its still mostly luck.

Posted by mpowell | February 26, 2007 12:50 PM

Isn't this "Insider" material? Matt: how can you support this junk? i used to love ESPN but the idea of paying them to read this stuff makes me bilious.

Posted by berger | February 26, 2007 1:12 PM

I think a large part of this effect is that Duncan, Manu and Parker don't play very much in games where the Spurs are blowing out the Sonics or trouncing the Celtics, but they're not being rested in regular series games against the Mavs, Suns, Rockets or Jazz. In other words, they're being "rested" in blowouts, which the Spurs have been on the good side of in a lot of games, just like any coach would do with his stars.

Posted by flickup17 | February 26, 2007 1:15 PM

I just skimmed the comments, so I apologize if someone has already said this, but the big warning flag is that the Spurs can't beat the good Western Conference teams. I don't think they've beat any of the top teams--Phoenix/Dallas/Houston, maybe even Utah--since early in the season.

Posted by Jerry | February 26, 2007 1:18 PM

"And, re: the balls of Horry, Cassell, etc. Perhaps we confuse cause and effect -- we think the 95 Rockets were especially ballsy because they got lucky and won close games. That's what gave those guys the "clutch" reputation."

But the Rockets had a similar run in '94 -- they came back from 2-0 down to Phoenix, and from 3-2 down to New York. Was that luck, too? In '96, they beat a much more talented Lakers team in the first round, and their three wins were by four, six, and eight points. Was that luck?

I'll give you better examples than that. Phil Jackson, until last year, had never lost a series his team was leading. Was that luck? Bill Russell never lost a winner-take-all game in his college, Olympics, and professional career. Was that luck?

Doc Rivers, coach of my beloved Celtics, has, as a player and coach, blown series leads of 3-1, 2-0, and 3-2, and has twice lost three home games in a single series. Is that luck?

I just don't understand why we're willing to discount something that we've all experienced in our own lives, just because it's hard to demonstrate statistically. Don't we all know that some people perform better in high-stress situations than others?

Posted by Andrew | February 26, 2007 1:31 PM

Resting your starters in February doesn't make them a step faster in the 4th Q of a conference finals. Unfortunately for Pop, his team is just old. His starters average over 30 (Dallas' average closer to 27). Their lock-down defender, Bowen, will be nearly 36 when the finals end. I suspect that's going to make the difference, not whether Dirk is averaging 1-2 minutes a game more than Duncan in the regular season.

And don't forget: the Spurs inability to close games in the regular season means they will face a much tougher first round oppenent than the Mavs or the Suns. If the season ended today, the Mavs and the Spurs would face teams with losing records. San Antonio would match up against the Lakers, a team they've beaten just once in three tries.

Posted by Rafe | February 26, 2007 1:35 PM

I will give point differential its due:

I think point differential does mean that San Antonio has better playoff prospects than Utah, which has a similar W/L.

I just don't think point differential means they are in the same class as Dallas and Phoenix.

Posted by Petey | February 26, 2007 1:39 PM

Andrew: To answer one of your questions, of course Bill Russell kicking ass in the clutch isn't luck. He was a kick-ass player in all situations. Kobe Bryant is awesome with the game on the line -- but he's always awesome. The question (to me) is, is there some "clutchness" attribute that's independent of a player's ability to kick ass the rest of the time? Are there really guys like Bob Horry, who are pretty average players until the game's on the line, then they're suddenly awesome?

I think the answer is, "we don't know."

Posted by too many steves | February 26, 2007 1:51 PM

But the Rockets had a similar run in '94 -- they came back from 2-0 down to Phoenix, and from 3-2 down to New York. Was that luck, too? In '96, they beat a much more talented Lakers team in the first round, and their three wins were by four, six, and eight points. Was that luck?

Probably. Why would it be so surprising that it should be luck?

I just don't understand why we're willing to discount something that we've all experienced in our own lives, just because it's hard to demonstrate statistically. Don't we all know that some people perform better in high-stress situations than others?

I'm willing to go out on a limb and say that very few of the people posting here are world-class athletes. I'm perfectly willing to believe that there are differences across the population as a whole, but by the time you get to the NBA Finals I think the people who *really* choke in the clutch have been weeded out long ago.

Posted by Josh | February 26, 2007 1:53 PM

This is an entirely empirical question. Only Carlos seems to realize this. How much of winning close games is explained by persistant skill differences vs. ephemeral randomness should be answered by looking at the data. Similarly, whether point differential is a better predictor than W-L is entirely an empircal question to be answered by looking at the data, not by Petey simply repeating himself.

And if it is true that "such a skill is difficult to discern through statistical measures," then that probably means that the skill is SMALL relative to random factors. If the skill is so important, why is it so hard to discern? If you actually have the data, it's quite easy to quantify the level of uncertainty about how much close-game-skill is consistent with the data.

Posted by ed | February 26, 2007 1:55 PM

Of course the Phil Jackson thing wasn't luck, he had two straight runs with the most talented superstars in the game. It wasn't "lucky" that he had them, but it wasn't because he was "clutch" either.

Posted by dbt | February 26, 2007 1:56 PM

This is an entirely empirical question. Only Carlos seems to realize this. How much of winning close games is explained by persistant skill differences vs. ephemeral randomness should be answered by looking at the data.

Great. Now define "persistent skill differences."

. Similarly, whether point differential is a better predictor than W-L is entirely an empircal question to be answered by looking at the data, not by Petey simply repeating himself.

"Predictor" of what?

Posted by SomeCallMeTim | February 26, 2007 2:11 PM

Are there really guys like Bob Horry, who are pretty average players until the game's on the line, then they're suddenly awesome?

My take: I think "clutch" is a bizarre distortion of a more real phenomenon - choking. I think we've all at one point seen a player visibly choke, and know a few players in different sports who regularly fail when the pressure is highest. So, being "clutch" is being able to perform as expected regardless of circumstance.

Posted by brad | February 26, 2007 2:13 PM

i think the flip side of clutch makes this obvious -- there are some players that very clearly appear not to want to take the final shot, who seem to wilt under pressure. if your "star" or "go to" guy is one of those types, you're not going to win many big time playoff series (which almost always come down to some close games). so whether it's not giving in to the pressure, or actually performing better with adrenalin, i think clutchness exists, even if it's only willingness to try to perform (and performance remaining roughly the same as normal) compared to someone who ducks the responsibility.

Posted by dj superflat | February 26, 2007 2:18 PM

You guys have made some good points. I just want to address a couple things:

"Of course the Phil Jackson thing wasn't luck, he had two straight runs with the most talented superstars in the game. It wasn't "lucky" that he had them, but it wasn't because he was "clutch" either."

This is missing my point; extremely talented teams frequently blow leads or lose to inferior teams. Jackson's and Russell's teams virtually never did. If they had at least equal talent, or if it was close, they won. Period.

"I'm perfectly willing to believe that there are differences across the population as a whole, but by the time you get to the NBA Finals I think the people who *really* choke in the clutch have been weeded out long ago."

Ever looked at Chris Webber's eyes during a big game? He looks terrified. How about the Blazers in Game 7 against the Lakers in 2000? They missed, what was it, 15 straight shots? That could be random variance, of course...

See, we could argue about this forever, because it's hard (perhaps impossible) to "prove" what I'm arguing. I doubt the sort of statistical correlation that Steve requested above actually exists. But just because something doesn't apply globally, doesn't mean it doesn't apply locally.

Posted by Andrew | February 26, 2007 2:20 PM

SCMT: a better predictor of future wins (in regular season, playoff or both). For example, I would bet that the Spurs win more games from here on out than the Jazz, based in part on the Spurs' superior point differential.

Posted by too many steves | February 26, 2007 2:38 PM

I think the "is clutch real?" discussion is separate from the larger "are there psych measures that stat-analysis can't measure?" question.

Posted by MJ | February 26, 2007 2:39 PM

I want to respond to ed's comment: saying that this skill is difficult to tease out through statistical means is not the same as saying it makes a small difference relative to random factors. Part of the reason that it is difficult to measure statistically is that so few big games are played. The team that wins the NBA finals this year may play 7-8 close, important playoff games. That team may or may not be quite as good next year, so you can't count on a whole lot more data points down the road. Now there is a difference between statistical uncertainty and luck. If this year's Mavs win a bunch of close playoff games were they lucky or did they play/get coached well down the stretch? They may have been lucky, but they may have also been good and statistics will not allow you to discern the difference. Now there have been some examples of guys who always seemed to make it happen: Jackson teams and Russell teams, for example. The question to ask is: would we expect these kind of outliers in the history of the game if it was purely luck? But that's pretty tough to answer.

Posted by mpowell | February 26, 2007 2:41 PM

Funny how someone used Chris Webber as an example of a choker only a day after he hit the game-winning shot in a nationally televised contest.

Posted by Steve | February 26, 2007 2:50 PM

i think the webber shot is actually a good example -- he didn't take he a shot, he put one back. webber never really seemed comfortable being the man (you could see it in his face), it was often bibby that kept sacto in the big games by being willing to take the big shots (generally 3s). i bet webber shines now that he's option number whatever at high stakes moments like last nite.

Posted by dj superflat | February 26, 2007 3:15 PM

SCMT: a better predictor of future wins (in regular season, playoff or both). For example, I would bet that the Spurs win more games from here on out than the Jazz, based in part on the Spurs' superior point differential.

Right. Except that (a) "in part" is an important part of your evaluation, and yet that's often only implicit in stat-types' claims with no indication of what else goes in there, and (b) there's no indication how discriminating the point differential is--the Spurs are ahead of the Suns are ahead of the Mavs, and I predict wins in something like the reverse order (I wouldn't pick between the Suns and the Mavs; I'd put both ahead of the Spurs).

Posted by SomeCallMeTim | February 26, 2007 3:26 PM

Could there be statistical information that might be, however tenuously, correlated to "clutch" or character? Come-from-behind, lost leads, quarter breakdowns especially 4th quarter,etc. I think I would bet playoff money on a team that consistently outperformed the opposition during the 4th quater of games, irrespective of seasonal differential. Legs.

Posted by bob mcmanus | February 26, 2007 3:51 PM

It is always funny when pseudo-intellectuals like MY pretend to care about sports for the sake of street cred.

MY: You're not fooling anyone.

Posted by Karl | February 26, 2007 4:19 PM

Could there be statistical information that might be, however tenuously, correlated to "clutch" or character?

Here's my suggestion: FT shooting percentage in the last 5 minutes, when lead is less than 5 points. (FG% would be interesting too, but there are far more variables). I'd also compare number of shots take in those 5 minutes to a player's average, to see if he is "hiding" from the ball.

Of course, nobody tracks such stats, do they?

Posted by brad | February 26, 2007 5:45 PM

Hollinger is nuts to declare the Spurs #1 though. Even by his own categories he lists on his site, Dallas is better than the Spurs in every one except for strength-of-schedule over the last 1/4 of the season and total margin of victory (which is close between the two of them).

Dallas has a signifigantly better record than the Spurs, a stronger schedule, a superior record in last 10 games (10-0 to 6-4), and a clearly superior margin of victory over the last 1/4 of the season (+9.29 to +7.20).

Hollinger definitely needs to retune his rankings and bring the weight down a little bit on strength of schedule. Since Dallas has a stronger overall schedule strength, while only being .15 below San Antonio in victory margin, while being +2 better in victory margin over the last 1/4 of the season, but with a schedule strength over that last 1/4 .4 below the Spurs, it's clear that he's not only putting too much weight on the last 1/4 of the season, which was evident earlier in the season in his rankings, he's putting too much weight on schedule strength too.

Posted by Jimm | February 26, 2007 5:57 PM

We seem to have this clutch argument once a year.
For those interested in attempts to quantify clutch performance in 2004-2005 see the series by 82games. Interesting that it seems to predict that Wade would be more clutch than Dirk....

http://www.82games.com/clutchplay3.htm

# Player Team PER dPER Diff +/- Rating
1 Ginobili SAS 49.2 11.3 37.9 +9 39.0
2 Stoudemire PHO 53.7 19.5 34.2 +16 36.2
3 Nash PHO 55.1 23.8 31.3 +31 35.4
4 Wade MIA 29.2 1.4 27.8 +79 33.7
5 Nowitzki DAL 44.7 16.0 28.7 +43 32.3
6 Camby DEN 31.8 3.9 27.9 +22 30.6
7 Griffin MIN 35.6 15.8 19.8 +58 30.2
8 James CLE 38.9 12.4 26.5 +13 27.8
9 Stackhouse DAL 36.1 9.2 26.8 +5 27.6
10 Allen SEA 31.8 10.0 21.7 +58 27.0

Posted by Just Karl | February 26, 2007 7:05 PM

yeah, aside from the incredibly boneheaded foul that gave dallas the opportunity to win one of the closest series ever, manu is totally clutch.

Posted by dj superflat | February 26, 2007 7:10 PM

Could there be statistical information that might be, however tenuously, correlated to "clutch" or character? Come-from-behind, lost leads, quarter breakdowns especially 4th quarter,etc. I think I would bet playoff money on a team that consistently outperformed the opposition during the 4th quater of games, irrespective of seasonal differential. Legs.

Roland Beech of 82games attributes such things to coaching, not clutchness. Stats for last year here (I've posted this before).

Posted by Al | February 26, 2007 8:17 PM

the problem with citing records in close games is that you need to know why the game was close: if games are tight and go back-and-forth down the stretch, that's one sort of close game, but if you have a 15 point lead with 3:40 to go and you bring in your subs and the other team hits 4 3s while your bench players don't do jack, that's another "close" and not nearly as revealing.

(this is also why i pay zero attention to point differential.)

PS. i'm always amused at those who think they can read chris webber's face....

Posted by howard | February 26, 2007 11:55 PM

A team plays as pretty as its coach. Popovich makes D'Antoni look like pre-social conscience Brad Pitt.

Posted by Brangelino | February 27, 2007 4:34 AM

The Rockets were indeed lucky...lucky that Jordan had retired to play baseball! For as much as I couldn't stand the Bulls and was glad to see Jordan give some other teams a chance, it is hard to fathom the Rockets beating a Jordan-led Bulls squad.

Hollinger's ratings are what they are. He has been pretty honest about how they are arrived at. I think they are a bit illusory, but they offer an interesting perspective outside of the 'chalk' analysis that we often get in sports. Perhaps his rankings will take on the same sort of usefulness that RPI does in college basketball. Such rankings are ultimately as inaccurate as the writers and pollsters are about the ultimate quality of the teams but offer a reasonable alternative to the bias inherent in most writers.

Posted by Ricky | February 27, 2007 10:51 AM

The fact that otherwise intelligent people say shit like, "Chris Webber is a choker -- I can see it in his face!" is the best proof I can think of that it's worth pursuing a better stats-based understanding of basketball. I hate to make the tired Moneyball analogies, but it really does sound like the dipshit old-timey baseball scouts who would say, "he's got the good face," meaning, "a ballplayer's face." As Billy Beane told them, "we're drafting baseball players here, not blue jeans models."

Posted by too many steves | February 27, 2007 11:03 AM

yeah, presuming to be able to read the emotions on another human's face is just ridiculous. there no basis for such subjectivity, regardless of how many people agree that someone's face betrays a generally recognizable, defined emotional state. it's as silly as thinking that emotions are revealed in people's demeanor and faces as a form of communication. thank god we've got numbers to rely on!

Posted by dj superflat | February 27, 2007 1:07 PM

"The fact that otherwise intelligent people say shit like, "Chris Webber is a choker -- I can see it in his face!" is the best proof I can think of that it's worth pursuing a better stats-based understanding of basketball."

You are seizing on the most superficial part of my argument and implying that it's the central piece. Chris Webber is a choker because his teams, college and professional, have consistently underperformed in the postseason -- usually in dramatic and heartbreaking ways. That he generally looks (to me) scared to death at the end of big games is merely a bit of subjective evidence that his postseason failures might not be a matter of coincidence.

Look, you are absolutely free to believe that Webber's and Rivers' and Jackson's and Russell's records in big games are purely a function of overall ability and luck. But that belief is ultimately as unscientific as my analysis of Webber's eyes.

As for the Moneyball people in baseball, "John McG," an occasional poster here, said it best on Malcolm Gladwell's blog, so I'll just quote him:

"I think what gets me about a lot of these number-crunchers is the lack of humility. The Boston Red Sox win a World Series with a sabermetric savvy GM (and the second highest payroll in the major leagues by a healthy margin), and the statheads put out a book about how they invented a new way of winning. There was no book last year about how the White Sox invented a new way of winning last year, and I suspect there won't be one about the Cardinals this year, even though they won championships with even more meager payrolls."

Posted by Andrew | February 27, 2007 2:21 PM

Now you sound like Joe Morgan, Andrew (or, John McG sounds like him). First, "the statheads" did not "put out a book" -- Michael Lewis put out a book about the statheads. Second, the book was written and published before the Red Sox won the World Series.

Oh, and there was a book about the "old way" of winning, and it was about the Cardinals -- or, more specifically, about what a genious Tony LaRussa is compared to the statheads. It was by Buzz Bissinger, the dude who wrote "Friday Night Lights." I didn't read enough of it to fully judge the merits, but let's just say that, as a writer, Bissinger is Chris Webber and Michael Lewis is Bill Russell.

Posted by too many steves | February 27, 2007 2:32 PM

John McG was referring to a book by a Bill James acolyte (I don't remember the name of the book or the author, but I remember reading a review of it) about how the Red Sox's World Series win was a vindication of sabermetrics.

I don't follow baseball very closely, so I won't try to make John McG's case for him. But I hope we can all agree that both statistics and subjective impressions have strengths and limitations, and it is foolish to rely exclusively on either one.

Posted by Andrew | February 27, 2007 2:53 PM

andrew, do you happen to know why havlicek had to "steal the ball?"

because the sainted bill russell, of whom no one is a greater admirer than i, didn't notice that he was throwing an inbounds pass right into a guide wire on the backboard, resulting in it going out of bounds and letting philly inbound.

which is to say, i think maybe chris webber should be cut some slack on the NCAA final. other than that, there is only one season that chris webber was part of a team that had the ability to win the title in the nba, and the refs took it from his team, not his "fear."

webber is a very good, very savvy player, not a great one, but it's not because he "looks scared." wrt dj superflat, for example, i can't always read the emotions on my parents' faces, or my wife's, or my son's, or my best friend's; why should any of us think that we can read the emotions on chris webber's face?

as for statistical analysis, there is plenty of basis for saying that baseball analysis "works;" it's part of why the oakland athletics keep performing so much better than the kansas city royals, even though both have comparable payrolls.

there is as yet no basis for saying that basketball stat analysis "works," and i'm not sure there ever will be: that's up to valiant researchers to assess....

Posted by howard | February 27, 2007 3:06 PM

In the last 5 years, treating east and west seperately, just seeing who made it to the finals, not who won them:

West:
Last 5 finals appearances were by a #1, two #2s, a #3 and a #4 in WL record.
Last 5 finals appearances were by a #1, two #2s, a #3 and a #5 in PD.

East:
Last 5 finals appearances were by a #1, three #2s, and a #3 in WL record.
Last 5 finals appearances were by three #1s, and two #2s in PD.

Not very informative one way or another, though slightly leaning to PD as more important.

Posted by Njorl | February 27, 2007 3:46 PM

re "why should any of us think that we can read the emotions on chris webber's face?"

ummmmm, isn't that what humans try to do in their interactions just about every day of their lives? not to say they always do it accurately, but are you really going to claim people can't generally tell when someone else is upset or sad or afraid or nervous?

and when the look just happens to coincide with not just missing the big shots, but shying from taking them, why wouldn't people put 2 and 2 together?

put another way, how dare anyone talk about michael's will to win? all we need is the numbers, right, because we can't read determination on mike's face, etc. let's throw out all the heuristics that serve humans pretty well for some numbers that may sometimes correlate with what they're trying to establish (leaving aside that it may just be correlation of course, rather than having found the magic ingredient that explains everything).

i realize i'm being somewhat of a crank, sorry. but one last thing re webber -- why is it so crazy to think that one might be significantly affected by having one of your most formative experiences be failure via boneheaded mistake on enormous national stage like webber, leading one to be somewhat worried when the stakes are high again? webber may mature out of it, but it just doesn't seem ridiculous to think he's thinking, not again, when getting jobbed by refs against LA or whatever. webber really isn't worth this analysis, but i find the meta point interesting.

Posted by dj superflat | February 27, 2007 4:11 PM

dj superflat, if you can't read the faces accurately of the people who are nearest and dearest to you, then you can't read chris webber's face accurately either.

when it comes to michael jordan's will to win, for instance, i don't rely on reading his face, i rely on the testimony of his teammates, who are actually in a position to know rather than guess about facial expressions.

you want to say i don't think that webber is a great player partly because i think he shys away from taking the big shot, fine, what do i care? i didn't watch enough sacramento king basketball to know whether that's true or not, but it's at least something that you can get a reasonable sense of by watching a big enough sample of games and observing whether webber tended to pass up open shots late in close games.

but when you say (that is, when one says) i can see the fear in chris webber's face: no, you can't. you're guessing that you can, but that's all it is.

Posted by howard | February 27, 2007 6:19 PM

sorry, but you could just as easily claim people are guessing whenever they discuss the internal states of others. that's a fairly meaningless point, doesn't square by what we generally mean by guessing, etc. if you're gonna tell me that you don't routinely presume to be able to fairly accurately judge the mental states of relative strangers based on non-verbal cues, then you'd at least be consistent. but you'd also be lying. it's possible that what most folk think they recognize as fear/despair on webber's face is something different. but the odds that the consensus is wrong are slim (and this was frequently discussed in the sports world at various times, so folk aren't really just making this up out of thin air, even though webber's just an easy example).

put another way, i can assume that i understand your words based on my prior interactions with humans and the english language, but there's a chance your actually using words differently from most other folk and your comment actually parses the meaning of britney's latest meltdown. but i doubt it. (it's like the deconstructo fools claiming language doesn't really work; ummmm, then why are you using language to communicate with me.)

this was fun, thx.

Posted by dj superflat | February 27, 2007 6:42 PM

Here's the thing: MJ's "will to win" doesn't mean a damn thing if he doesn't actually make the shots. I can have a mean look on my face, and really really want to win, but I still won't play like MJ. Maybe Chris Webber really is a choker. But the evidence should be that he doesn't make the big plays, not that he "looks scared." Tim Duncan usually looks bored, or he looks like a whiny bitch. That doesn't mean he doesn't have the "will to win."

Posted by too many steves | February 27, 2007 6:51 PM

the demeanor is relevant because someone asked whether we can separate this out from luck. that is, maybe no one's a choker, it's just luck of the draw that jordan's big shots all tended to fall while those of others did not. for what it's worth, i find this highly plausible (that is, we've set up a system with enough iterations that someone should just keep winning as a matter of luck (people cite horry here) rather than skill (it's somewhat like being surprised by someone winning x games in a row to win the tourney, it's a forced result), though it's obviously hard to separate luck from skill (the example i like is warren b., his son said his dad sometimes makes trades because of his mood, whatever, they just mostly tend to pan out, which is entirely possible in a market with millions of participants all making trades all the time (someone's gotta be right every time))).

Posted by dj superflat | February 27, 2007 7:37 PM