http://www.draftexpress.com/article/Win-Scores-2008-NBA-Draft-Recap-2953

“I’m just not sure if his statistical approach to evaluating prospects is more useful than a doctor looking at a player’s physical evaluation.”

First, not insulted, all in good fun…

Second, I didn’t realize that’s the conversation I jumped into. I guess I don’t see the sense in worrying too much about injury potential with guys in the draft. First of all, injuries are capricious and random. I don’t care what a doctor says, there is no way to predict which guy will fracture his kneecap or break his wrist etc. It can happen to anyone.

Second, the time to worry about it is when you sign them to their second and third contracts, when you are actually taking a significant financial risk. If Blair blows out his knees in his second or third year and he can’t play, so what, it’s a low risk high reward gamble that just didn’t pay off, which is basically what nearly every draft pick is anyway, except that the reward is much much higher in Blair’s case.

Leaving Blair aside. the fact that so many people passed on Lawson is mindboggling given how outstanding his numbers were in college. By any metric you look at, and if you use your eyeballs too, it’s clear that Lawson is going to be much much better than those for instance, Flynn and Holiday, statistically, at every step of their careers. What would have been wrong with taking Lawson at 8 anyway?

Man, I have to learn to give up on the players that get away…

Edit: Here is what I just was on True Hoop on small guards, seemed apropos

“What we see is the biggest offensive advantage for qualified small guards 6-foot-1, while players 6-foot-4 suffer disadvantages. So while players 6-foot-3 definitely struggle at the NBA level, the absolute most you can say is that there is no strong relationship between size and production. We see a similar effect when we look at a variety of measures, including John Hollinger’s Player Efficiency Rating. This speaks well for a lot of players who are often dismissed as being too small to play small guard. This group includes Ben Gordon, Raymond Felton, Jason Terry — all top-notch players who are often considered liabilities because of their size. … I’d think twice before dismissing the “undersized guard.” Because at this level, there’s probably a good reason he’s made it to the NBA.”

You have to use a small amount of common sense when you look at the draft rankings. It’s not a substitute for, but a complement to scouting.

I don’t have them in front of me, but if you take Hollinger’s historical results and compare them to the actual drafts, I remember Hollinger’s being far better. And that’s just one guy, doesn’t mean he’s the end all and be all of statistical analysis.

On Doerr, I think Owen is correct that WoW clearly values players differently than the average NBA decision maker.

“I think Rashidi’s point was that while statistical analysis of draft prospects is, as Doerr admits, a “a very imperfect science”, X-rays, CT Scans, Blood Tests, Past Injuries, and Family History are a much more perfect science.”

Maybe that’s true a small fraction of the time, but how many draft prospects are busts for medical reasons??? On the other hand, how many are busts due to lack of skill? I would guess that the latter out number the former 100 to 1.

There are plenty of examples of players being red flagged for medical reasons (not just draft prospects, but veterans too) and never having a problem. I’m not a doctor, but projecting someone’s health 5 years down the line does not appear to be an exact science by any means.

On the other hand, a lot of the guys who have career ending or career altering injuries simply cannot be predicted. What doctor could have told the Bulls that Jay Williams was dumb enough to jeopardize his career by riding a motorcycle and would then proceed to fall off and tear his knee to shreds? Where doctors telling Phoenix about Q’s back before they made a significant investment in him (which they pawned off on the Knicks, of course)? Did doctors know that Grant Hill’s ankle was shot for half a decade due to an injury he suffered in Detroit?
I’m not ripping NBA doctors or the medical profession, just saying that a doctor can only tell you so much about the average player, while a statistician can tell you a good deal about most players.

“In many ways a player’s medical is a crystal ball. If a guy is 6?6?, weighs 288 lbs, and has no ACLs, it’s probably as sure a bet that a drafter can make that his college stats won’t translate well to a six month, 82 game season. (At least it’s a less risky gamble than drafting Joey Dorsey over Derrick Rose :)”

Well, first of all, we’ll get a great chance to test your theory by watching Blair play. I have no idea how he’ll do, just saying that we’ll see.

Again, the point is not to draft Dorsey over Rose. You do see a general pattern emerge when you look at the rankings. Using stats can increase your odds of making a good pick/decrease your odds of picking a bust, which is all you can really do in the draft. Take a look at the draft steals the Rockets have pulled off in the 2nd round since MIT grad Daryl Morey joined the franchise (which goes back longer than he has been GM).

I’m not saying that just looking at a players stats and creating some kind of model means you will draft better, I’m saying that if you do it right you will.

Didn’t mean to misquote or misstate the statistical work of Doerr. Ted said if Rashidi looked up “the historical results of Hollinger and Doerr’s models for evaluating draft prospects” then he wouldn’t say that doctors’ opinions are more valuable to drafters than statisticians’ opinions. So I googled “Erich + Doerr + Draft” to see if I agreed with Ted or Rashidi, and his 2007 draft analysis was what came up.

“Read enough of him and you will understand that he is very far from thinking he has any kind of a silver bullet or that drafting is anything but a very imperfect science.”

I think Rashidi’s point was that while statistical analysis of draft prospects is, as Doerr admits, a “a very imperfect science”, X-rays, CT Scans, Blood Tests, Past Injuries, and Family History are a much more perfect science. Therefore, it isn’t much wonder why doctors’ opinions are valued highly at the time of the draft, especially if the best prospects a statistician can suggest are guys that are obviously not going to be drafted or last long in the league.

“The numbers are not a crystal ball. Yes, there is a statistical relationship between what we see in college and the pros, but it’s not a perfect relationship. Or to put it another way… the college numbers tell us something, but not everything.”

In many ways a player’s medical is a crystal ball. If a guy is 6’6″, weighs 288 lbs, and has no ACLs, it’s probably as sure a bet that a drafter can make that his college stats won’t translate well to a six month, 82 game season. (At least it’s a less risky gamble than drafting Joey Dorsey over Derrick Rose :)

Again, sorry to insult Doerr, or Berri, or you with my post. I’m sure Doerr is a smart guy (like I know you are from reading your posts over the years). I’m just not sure if his statistical approach to evaluating prospects is more useful than a doctor looking at a player’s physical evaluation.

Below I posted what Berri had to say about his research back in 07. And the draft last year produced similar results. The basic point of their research is that Win Score is relatively durable from college to the pros. But so is the belief in scoring. Which is why Renaldo Balkman played 700 minutes last year and Rudy Gay has played 3000. Basically, while WS is a relatively accurate predictor of what a player’s stats will look like, it’s a crappy indicator of what NBA decisionmakers will think. If you want the latter, I suggest you use PER.

Here is Berri

“First of all, there is a statistical relationship between what a player does his last year in college and what he does his first year in the NBA. Specifically, from 1994 to 2005 there is a 0.69 correlation between a player’s Win Score per minute his last year in college and his Win Score per minute his rookie season in the NBA. In other words, if we shift from correlation coefficient (or r) to r-squared, about 47% of what a player does on a per-minute basis his rookie season can be explained by what he did the previous season in college.

To put this in perspective, in The Wages of Wins we report that a baseball player’s OPS last season explains 33% of his OPS this season. This means the correlation coefficient between OPS this year and last year is only 0.57. So college performance in basketball predicts a rookie’s NBA performance better than a veteran baseball player’s performance predicts his future production in Major League Baseball.

There is an apples and oranges issue in comparing basketball and baseball numbers. Still, I think one could argue that basketball numbers tell us more about the future than baseball numbers. In this sense, basketball numbers are better.

That being said, the numbers are not a crystal ball. Yes, there is a statistical relationship between what we see in college and the pros, but it’s not a perfect relationship. Or to put it another way… the college numbers tell us something, but not everything.”

I didn’t say it, but I looked it up anyway. All I found was Erich Doerr’s 2007 pre-draft analysis. It listed 3 guys as “pick booms”: Nick Fazekas (waived by three teams), Stephane Lasme (played a grand total of 4 seconds in the NBA), and Rashad Jones-Jennings (a guy who doesn’t even have a Wikipedia page telling me who he is).

Of all the players that actually went in the 1st round that year, Doerr seemed to have been highest on Yi Jianlian.

I wouldn’t say Doctors are necessarily more useful than statisticians when it comes to projecting the futures of players, but even Dr. Giggles could have given more useful advice on that draft…

(note: my favorite part of the WoW journal entry this came from is in the comments section, where Dave Berri responds to a certain “Owen” with the comment: “I see you fighting the fight over at Knickerblogger. Keep it up. You seem to be more than holding your own.”)

Sweetney.
Frye.
Hill [Tho the jury is still deliberating on him].
2 to the Bulls.
Gallo: Might be a bust. lol Stay healthy and take your shots D-Gallo!

Late to Mid-Rounders:

Ariza.
Chandler.
Nate.
Lee.
Do What TD Do [Jury still out on him as well].

I’m not outraged by the Hill pick, but also not thrilled. If he can be the kind of defender that guys he’s being compared to like Kurt Thomas and Antonio Daniels are/were, I’ll be happy with the pick. I think people are taking for granted that Hill can easily be the kind of defender and overall consistent, quality player than Thomas and Davis have been.

“Because doctors are “slightly” more credible than statisticians.”

??? If you look at the historical results of Hollinger and Doerr’s models for evaluating draft prospects I don’t think you would say that. You can also look at the success of teams that do a good job of incorporating statistical analysis into their decision making, like the Rockets.
There are plenty of examples of guys getting red flagged for medical reasons and never having any problems. On the other hand, doctors have missed plenty of injuries, including Grant Hill’s which first occurred in Detroit before he signed the $100mm plus deal with the Magic. Not saying that was any doctor’s fault, just that it’s not an exact science.

“Also there’s no point in talking about Lawson or Jennings when the Knicks made it clear that Toney Douglas was gonna be their guy at #29.”

What happens if Douglas goes #28???
If Lawson turns out to be a good NBA player and Hill doesn’t (not saying that’s going to be the case), then who cares what position Lawson plays?

“As investors, teams try to get the most bang for their buck. Statisticians may have said: “Mike Sweetney’s college stats say he’ll be a solid starter in the pros”. Doctors may have said: “Mike Sweetney has a family history of obesity which will see him out of the league in three years”. If the doctors are right, then it doesn’t matter what the stat-men say.”

???? I don’t think doctors are very important when it comes to personnel issues, they probably only red flag a small fraction of possible acquisitions.
When it comes to draft prospects stats are in fact used to predict future performance, and with pretty good results (far better than the actual draft order, at least).
How many draft prospects every season have medical issues that doctors can diagnose with any foresight??? Most injuries are accidental and unpredictable. Every draft prospect has some sort of statistical history.
It’s a no brainer to have a doctor check out the players you’re looking at, but it should also be a no brainer to have a statistician. Having those guys doesn’t necessarily tell you what to do

Basically, by not getting into shape, he’s likely throwing away 20-40 million dollars over the course of his career. And even if NEXT summer he does finally cut the cheeseburgers, I doubt the knicks will find a place for him in the rotation. Right now is his chance because the knicks want to move him, so they have an incentive, even if it doesn’t fit with their game plan, to show him off. Next season that won’t be the case.