Here’s something I wish I would have seen before posting yesterday’s comment on the Toronto Raptors.
Much has been made of NBA teams employing more and more statistical analysis in their decision-making process. It’s important to remember, though, that such analysis isn’t used by all teams (and even teams where this analysis happens, it’s not clear that the stats trump the other information teams employ). To see this point, consider the following comment from Malcolm Gladwell. (HT: RaptorsHQ)
I remember once having a conversation with a top executive with the Toronto Raptors. I asked her about the stats revolution in basketball and she just kind of shrugged and said, “It’s interesting, and we look at those things, but you have to understand that for our purposes, it’s all [about] character.” The thing that separates players is that some have a work ethic, some don’t; some are coachable, some aren’t; some party all night, some go to bed early. From her standpoint, it’s all those intangibles.
In reading this quote we have to remember… the Raptors have never won more than 47 games in a season, and have only advanced past the first round of the playoffs once. Perhaps (just perhaps) there is more to player analysis than focusing on the “intangibles.”
One last observation… once upon a time I noted that the owner of the Raptors tried to get his general manager in basketball to read The Wages of Wins. Maybe he will have more luck assigning our second book (am I being too snarky? Sorry about that).
– DJ
The WoW Journal Comments Policy
Wages of Wins Journal 
Scott
October 23, 2009
Well, in the Raptors defense, those intangibles are important when scouting rookies and younger players- if two rookies have similar production, I’m taking the one who has a better work ethic, listens to the coach, doesn’t party, etc. But once you get to the vets who have little chance of improving (what you see is what you get) it doesn’t matter quite as much.
Kevin
October 23, 2009
I told a friend who really cares about basketball to check out what you do here. He didn’t like it at first, but we’ve taken stats classes together as undergrads, I knew he basically got it and was able to see how strong the predictions are. When I probed a bit on why he didn’t like it, he revealed the real problem. It was as though I told a lifelong Christian that god was not real. He simply hated that basketball could be reduced to a tidy little model and predicted with such regularity. It killed the fun for him.
I don’t get that myself. I think it’s pretty amazing that such a wild and chaotic looking sport, with wild and chaotic athlete personalities, can be so elegantly described through abstraction. It’s not as if the model is so powerful you KNOW what will happen (though as bad as some GMs are, it’s almost as if they’re trying to make the big picture clearer for prognosticators). There is still dramatic tension between the best teams, and a stochastic enough playoff schema (at least in the later rounds).
But I think what really explains the heat that WoW gets comes from that “no, it can’t be, it can’t be!” moment that people get when they realize that sometimes, we’re not so complicated.
I’m studying theories that attempt to model emotion processes, and eventually, ‘free will.’ You should see the reactions we get…
FanFeedr
October 23, 2009
I think that the larger point that Professor Berri is making is that there is a balance between human judgement and empirical evidence. For those with a statistical background, regression analysis is one approach to predicting the future. It is hard to do for financial markets, and, as far as we know, no one has cracked that nut.
The issue is different in basketball. The variables are fewer, and perhaps more importantly, we can look back over time to figure out if the forward-looking analysis has borne fruit, that is to say, has mapped to expectations.
Prof. Berri has smartly put forth that WPA has been both historically as/more accurate than other measures, and that that accuracy has stayed the same over time.
Certainly you want hard-working players, but there is no better measure of that effort than efficiency.
Michael Jordan infamously stayed out late, gambled and ate junk food. We can safely say that he was efficient and deadly on the court.
That is not to say that there isn’t a correlation, but that the means of measure shouldn’t be eyeballing a player and interviews.
Kenny “Sky” Walker had a wicked vertical game but was woefully incompetent in every other phase of the game. Let’s presume he worked hard and didn’t party (we don’t know this.) He was still not fit to play in the NBA.
We rest our case.
Tball
October 23, 2009
I think a similar issue is playing out with Rondo and the Celtics. I think Ainge is afraid of Rondo’s intangibles and afraid of signing him to a big money, long term deal. They think his hard work is related to this next contract and that he has had self-discipline problems that they’ve been able to control with the carrot of the next contract, but they are afraid of what Rondo becomes once that deal is locked up and they lose that carrot.
I don’t know what Ainge thinks is a good alternative to signing him long term, but I am afraid to find out.
Italian Stallion
October 23, 2009
Even though I agree with your point, I’m going to have to disagree with your interpretation of the quote.
I think it’s virtually impossible that any professional sports franchise doesn’t use some kind of talent/evaluation method.
Some may use more sophisticated mathematical statistics than others.
Some may use visual appraisals of talent and skill.
Most probably use some combination of things.
What I think she is saying is that she believes that work ethic, coachability, and a clean lifestyle trump small edges in ability that you might be able to measure better with advanced stats.
Does anyone think she wouldn’t want James, Wade, CP3 on the team if they liked to stay out all night long and were always late to practice?
However, given a choice of the 3, if one was super clean cut etc.. that’s the one she’d go with even if the stats suggested he was 3rd best by a small amount.
That’s what I “think” she means.
simon
October 24, 2009
IS//
You’re stretching it. She “shrugged” and said ” for our purposes, it’s all [about] character.” I think it’s much more feasible to intepret her comment as “we look at points, rebounds, FG% and those stuff, but basically we go by the eyeball test and the feeling we get when we see the player upclose in person.”
The article hasn’t mentioned it, but Colangelo was widely documented in the media for his usage of the Caliper test to measure the personality potential of his draftees, and Colangelo commented:
“They said his upside and potential were off the charts, … ‘Out of all the athletes we’ve profiled, we’ve never seen anything like this.’ ”
http://sports.espn.go.com/nba/news/story?page=espnmag/bargnani
So Colangelo believes in a systematic test, but interestingly he hasn’t shown much in the way of any statisitcal acumen judging by his recent draft history or free agent signing. Also the DeRozan articles mostly quote the Raptors brass saying what a nice young fella he is and how coachable he is, etc, not what his numbers looked like in college, another indicator of their priority.
To be fair to Colangelo, it should be noted that he first had tried to sign Ariza, so maybe the Turkoglu signing was just an emergency plan. Then again, Bryan could’ve just signed Marion for less.
simon
October 24, 2009
Oops, forgot to mention Colangelo’s comment was on Bargnani
Baron Von Munchasen
October 24, 2009
Kevin – you wrote “I knew he basically got it and was able to see how strong the predictions are.” – which predictions exactly are you referring to? The regular season predictions of wins do seem to come close to actual wins, but has Wages of Wins shown itself to be better at predicting regular season wins than, say, a Las Vegas bookmaker? I’m harping on this point because I hear a lot about how well the Win Produced model is at prognostication, but when I saw an analysis of last season’s prior predictions of regular season wins, there was no advantage over the odds posted at one Vegas sports book back in October. It even looked at numerous weightings and adjustments of playing time, and different lineups using WoW, but the predictions at best were 16-14 (the other weightings were 15-15). So unless the bookies are using WoW, it doesn’t seem to possess any informational advantage that the “betting experts” have. Possibly, this could imply that the bookies would make better GMs than many the ones currently employed. My question is, if Wins Produced is superior in modeling wins, wouldn’t you then have an advantage against those setting the market lines? OR maybe the bookies use Wins Produced themselves?
todd2
October 24, 2009
I’m going to tailgate on Fan and IS’ comments. I love the prof’s work but feel it’s only one piece of the puzzle. Joakim Noah’s WP48 comes in at more than twice the value of Derrick Rose’s. If you had to choose today, which player would you add to your roster?
The character issue reminds me of a conversation between Pat Riley and Jerry West back in the day that made the press. Riley questioned one of West’s acquisitions because of the character issue and the response was, “We don’t want a team full of milk drinkers!” And Ron Harper was traded away from the Cavs because of the people he associated with and went on to win a handful of rings with the Bulls/Lakers. Go figure.
Phil
October 24, 2009
I’d agree that character may not directly translate to a single season’s wins and losses in a significant way. Even considering example 1a of “low character” guys, the mid-late 00’s Knicks, the drama was only a side-note to that the teams were simply not very good.
The Spurs, on the other hand, are a team that pays a lot of attention to character and has led to big payoffs. It’s not just that players play better if they like each other – even though I personally believe they do – it’s that good players will sign with them for less. Getting players like McDyess for a pittance, or getting Duncan for less than a max contract, is a huge boon.
For an extreme example, consider that the Pacers are still recovering from having “low character” guys like Stephen Jackson, Artest and Tinsley as their base. It’s not just the losing seasons and high payroll (although that’s a big part), it’s that NBA basketball is ultimately an entertainment business that markets itself towards families, and part of that is making its players role models.
Did attendance drop after the brawl? I’m not sure. Though having starters suspended certainly didn’t help, nor did having a fan base clammoring for their departure while simultaneously seeing their trade value fall.
I think character matters. It certainly isn’t going to turn Bargnani into a shot-blocking, rebounding machine. And the fact that say, Shaq is not a hard worker doesn’t change the fact that he’s been exceptional for much of his career.
But it certainly matters.
ilikeflowers
October 24, 2009
todd2, that’s a really bad example. Derrick Rose was a rookie and will likely improve. Noah was no longer a rookie but he will still also likely improve. Using WOW there’s an argument for choosing either one since we don’t know how much either will ultimately improve. If I were using WOW I would choose Rose since he is likely to improve more than Noah (having spent less time in the league) and it’s easier to find a 0.200 neighborhood front court player than a 0.200 neighborhood back court player. Additionally, I know that I can get a high-productivity low-scorer like Noah on the cheap at some later date.
todd2
October 24, 2009
Flowers, I wasn’t advocating one over the other (and agree with your argument, btw.) All I did was cite numbers and ask a question. The post weighs the merits of statistical evaluations vs. character/intangibles. I simply tossed out an example to kick around…
Statement
October 24, 2009
You know what’s funny,
Colangelo graduated with an honour’s degree in economics.
Statement
October 24, 2009
Just to add on to my prior comment,
He got a Bachelor of SCIENCE in business management and APPLIED ECONOMICS.
Now I might just be a lowly M.A. student, but I’m sure applied economics has to have some applied econometrics in there as well, no?
Was this guy sleeping through all of his classes or what?
reservoirgod
October 24, 2009
I’ve just created an ESPN NBA.COM Fantasy League based on Win Score. You can check out the scoring settings for the league at http://games.espn.go.com/fba/tools/leaguesettings?leagueId=133748. I’m looking for other members to join the league before the season starts. If you’re interested, send your email address to reservoirgod at hotmail dot com.
simon
October 24, 2009
Statement//
We all know Cornell is a horrible school. ;)
But in seriousness I’ve seen many excellent engineering/economics students who refused to apply their scientific/rational thinking to anything outside academic work, so that’s not unique to Colangelo. Furthermore, these days a GM seems like a glorified accountant in some ways.
khandor
October 25, 2009
The best in any field of activity “learn and know” how to use their own judgment when making decisions in specific situations and, if/when necessary, to combine THAT ability with proper use of situational stats to confirm what their eyes and feel are already telling them is the right/best way to proceed.
It’s irrelevant whether this best person is Michael Jordan, Michaelangelo, Bill Gates, Bill Walsh, Warren Spahn or Warren Buffet [etc.].
The key difference between “the best” and “everybody else” is … exactly as Kenny Rogers/The Gambler once proclaimed … “knowing what to throw away and knowing what to keep, cause every hand’s a winner and every hand’s a loser, the best that you can hope for is to die in your sleep.”
The fact is …
1. If all you know is, “How to make the percentage play, each and every time”, you will most assuredly finish empty-handed; while,
2. If all you know is, “How to play by feel, each and every time”, you will most assuredly finish empty-handed, as well.
It’s not one or the other but being able to pick your spots with accuracy.
Cheers
ilikeflowers
October 25, 2009
Very profound khandor, one would never guess that the truth lies between two extremes or that both cases are simply examples of bad models. Do you have any more wisdom to share?
Baron Von Munchasen
October 25, 2009
Ilikeflowers, it seems you are not a big fan of witty aphorisms, adages, or quotes, and so I have a quote for you: “Wisdom cannot be passed on. Wisdom which a wise man tries to pass on to someone always sounds like foolishness … Knowledge can be conveyed, but not wisdom. “
khandor
October 26, 2009
BVM,
Amen, brotha. :-)