Hollinger's Draft Rater, Fun with math Options

[ZoomSlowik]

Some of you may be familiar with John Hollinger, ESPN's math geek when it comes to basketball. I personally love his NBA stuff, and he's tinkering with a prospect rater for the NBA draft. Basically it gives you an idea of who the most productive players are, meaning which guys are theoretically most likely to succeed early. Two major caveats: he relies on the listed heights for players, which are often inaccurate (he speculates that Griffin and Hansbrough are likely a bit generous), and it's a measure of actual production, not potential. Anyways, I'll post some snippets (it's ESPN Insider)

First his top-rated players:

Title of data
Player School Year Score
Michael Beasley Kansas State Freshman 856
Blake Griffin Oklahoma Freshman 725
Kevin Love UCLA Freshman 724
Danny Green North Carolina Junior 649
James Harden Arizona State Freshman 642
Robbie Hummel Purdue Freshman 601
Jerryd Bayless Arizona Freshman 599
Andrew Ogilvy Vanderbilt Freshman 598
Ryan Anderson California Sophomore 586
Dar Tucker DePaul Freshman 583
DeJuan Blair Pittsburgh Freshman 582
Ty Lawson North Carolina Sophomore 566
Tyler Hansbrough North Carolina Junior 558
Matt Howard Butler Freshman 556
Chase Budinger Arizona Sophomore 547
Malik Hairston Oregon Senior 537
Tyler Smith Tennessee Sophomore 528
Roy Hibbert Georgetown Senior 527
Marreese Speights Florida Sophomore 526
Kosta Koufos Ohio State Freshman 525


Some notes he mentioned:

-He only used stats for this season, so it's still a fairly small sample size (roughly half a season). Some guys will fall and rise before the end of the season, especially as the conference season winds down (specifically points out Green and Hairston as a bit flukish).

-Brook Lopez came in at #30, but would have been much higher if he wasn't suspended for the first semester.

-There are a lot of major names that don't appear on the list. Derek Rose is #25, mostly because his assist/TO ratio was low enough to make him a 2-guard in the computer, and Donte Green is #26 because his assist ratio and shooting percentage are major drags. With more seasoning both would likely fly up the charts, and both play for teams that really need them to be scorers.

The ratings were seriously unimpressed with OJ Mayo and DeAndre Jordan. Mayo's brutal AST/TO ratio as well as his poor true shooting percentage suggest he has a long way to go to be an efficient player, though his scoring ability is an asset. Besides a lack of production, Jordan got dinged for a low steal rate (athleticism indicator along with rebounding and blocks) as well as a 1:6 AST/TO ratio. He could REALLY use another year, but would get drafted too high to pass up as things stand. Eric Gordon also finished well off the chart (though better than the last two) because his rebounding rate was weak and his other numbers are decent but not stellar, and his AST/TO ratio is falling rapidly (his wrist injury is probably hurting that). Darell Arthur is another highly rated prospect who got docked for poor rebounding and bad turnover rates, and his stock is slipping as is.

-My own little commentary on some of the "who the hell is that?!" names: Hummel is a versatile big man from Purdue that can bang inside a bit as well as hit the 3. He's likely not an impact player on the next level, maybe a Nocioni type....Andrew Ogilvy is a big Australian center that has a good level of skill, but is not particularly quick and needs to bulk up....Ryan Anderson is a highly skilled offensive big man that rebounds at a good rate, but again, is not really quick or strong enough to be considered a really good prospect....Dar Tucker is a super-athletic wing player that needs a lot of work. We'll hear more about him as a prospect in a year or two....Dejuan Blair is a 6'7" big man at Pitt that is tough and a fairly decent athlete, but will have a hard time being more than a second round pick....Matt Howard is another undersized big man that has scored well and blocked some shots, but isn't a dominant rebounder despite playing in a lesser conference.



Anyways, it's just something I found interesting. If anyone besides me actually reads it I'll post updated rankings when they become available. The formula is still a work in progress, but it does seem to work reasonably well at figuring out which guys you need to avoid, at least when the full data is in at the end of the year.

[ZoomSlowik]

One thing I notice with this system is that it doesn't seem to properly account for size and athleticism for big men. Sure, it told you that Carlos Boozer was much better than where he was drafted, but guys like Curtis Borchardt (#7, ahead of Caron Butler and Tayshaun Prince), Michael Sweetney (#2, ahead of Chris Bosh and Dwayne Wade), Kris Humphries (#8 ahead of Andre Iguodala and Kevin Martin), Channing Frye (#6 ahead of Deron Williams and Danny Granger), Chris Taft (#8, same as previous), Shelden Williams (discussed below) and Paul Davis (#7, still ahead of Aldridge) also managed to sneak into the top-10 in previous drafts ahead of much more productive pros below them. Also, Tyrus Thomas placed considerably higher than Brandon Roy, Rudy Gay, and LaMarcus Aldridge in this rating, and far and away the #1 rated player in the class, with Shelden Williams placing #2. Obviously Tyrus' great athleticism has something to do with that, but it also didn't seem to factor in that he's nowhere near the ideal bulk/strength level for the league yet, making it difficult for him to match his rebounding rate or shooting percentage from college.

That makes me a little leery about Kevin Love that high. He's a phenomenal college player, but I have a very hard time seeing him maintain his absurd rebound rate that plays a major role in his ranking at the next level. Obviously perfection with something like this is impossible, but that seems like a rather consistent issue with this. At best it should be used as a tool combined with actual scouting, as you're always going to have guys that are studs in college that just don't have pro-level abilities and talented players that get drafted before they put it all together.

Also, at least by his system this rates as a REALLY strong class so far. A rating above 500 usually seems to correlate fairly well with a lottery-caliber talent, and there are over 20 of those in this draft despite some of the big time guys being rated poorly. Also, a 600+ rating generally seems to correlate with an elite prospect (21 guys have done that in the last 6 drafts). However, you have to consider that these ratings only incorporate roughly half a season, as some of these guys seem likely to fall off a bit. 5 of those guys have zero chance to even go in the first round this year, and guys like Ogivly, Hansbrough, Anderson, and Hairston are considered bubble first rounders. I would expect that to drop to about 15 guys (with some changing places) by the end of the college season, making it a fairly deep but not stellar class, and likely having few true elite prospects.