<div class="quote_poster">Quote:</div><div class="quote_post">About a week ago, I rented Leon Gast?s follow up to When We Were Kings (great movie, if you haven?t seen it). It was basketball documentary entitled 1 Love. It was a little disappointing because it tried to do too much in too short a time frame. However, one line did really stick out to me. Hubie Brown was addressed some high schoolers at a camp and said, ?To make it in The Show, you have to have either a game or a specialty.? This quote coupled with my never-ending fascination for the draft inspired me to do a little study to see if there was a quantifiable way to determine which college players in this year?s draft had the best chance to succeed in the NBA. So, I decided to compare this group of potential draftees from the NCAA Division I level and compared them to players in the past using Similarity Scores. Then from that, I was able to calculate their success rates. Defining the pool of players posed a small challenge because the three-point was introduced in the college game in the 1986-87 season. Therefore, I chose the 1990 draft as the starting point to keep the number of seasons before 1986-87 to a minimum, while making the pool as large as possible. Then, the 2003 draft was the cut off point because it?s still a little too early to fully evaluate the quality of the 2004 and 2005 drafts. To do that, it takes three years to get a pretty accurate evaluation. The Similarity Scores were calculated based on fourteen categories, which are listed below: 1. Height 2. Weight 3. Age 4. Minutes per Game 5. Total Games Played 6. Three-Point Efficiency 7. Free-Throw Efficiency 8. Two-Point Efficiency 9. Assist Rate 10. Turnover Rate 11. Rebound per 48 Minutes 12. Blocks per 48 Minutes 13. Steals per 48 Minutes 14. Personal Fouls per 48 Minutes The maximum score is 1000 and a significant comparison would fall in the range between 800 and 1000. So, to come up with a success rate, I analyzed every comparable player that was in that range using Minutes per Game and basic production metrics like PER as the primary factors and assigned a point value to them. They are as follows: 1 ? Definite All-Star, Elite Level Player (ex. Kidd, Shaq, Iverson) 0.75 ? Borderline All-Star, Second Tier Player (ex. Abdur-Rahim, Antoine Walker) 0.5 ? Starter, min. 25 MPG 0.25 ? Borderline Starter, Bench Player, 15 MPG ? 25 MPG 0 ? Fringe Player (<15 MPG), Bust So, first I calculated a player?s Minimum Success Rate by dividing the number of non-busts by the total number of players in the range, then multiplied by 100. Then, I averaged all the point values to get a Quality Index, which I used to get the percent chance that a player will end up as a starter or All-Star. One problem I ran into was that some players didn?t have enough comparable players to get a valid success rate. In those cases, I expanded the range until I got a large enough group of comparable players. With the nitty-gritty mathematical process out of the way, I ran the success rates for around 100 draft prospects, which should cover almost every college player who will be picked. One other note, I adjusted the rates to account for draft position, since a first rounder would have a better chance than a second rounder. So without further delay, here are the players in the 2006 Draft Class with the best chance to be starting at the end of their rookie contract. </div> 1. Ronnie Brewer 2. Rudy Gay 3. Shelden Williams 4. Rajon Rondo 5. Shawne Williams 6. LaMarcus Aldridge 7. Kyle Lowry 8. Brandon Roy 9. Marcus Williams 10. Darius Washington 11. Tyrus Thomas Interesting article. I don't think it's really anything to take seriously, though.
<div class="quote_poster">Quoting MrJ:</div><div class="quote_post"> Interesting article. I don't think it's really anything to take seriously, though.</div> Why not? I bet most NBA teams evaluate players in the draft in part by using a similar sort of statistical analysis.
<div class="quote_poster">Quoting durvasa:</div><div class="quote_post">Why not? I bet most NBA teams evaluate players in the draft in part by using a similar sort of statistical analysis.</div> They definitely use statistical analysis to determine their selections. I have to admit, I wasn't a fan of the statistical approach to comparing players. However, the craft has improved over the last couple of years and these stats are very telling. Right now most stats really cater to the offensive efficiency of players, once they compile more data on the defensive aspects of the game, these stats will be complete. The only player I'm surprised ranked so high is Shawne Williams. Physically he's amazing, but mentally he struggles with basketball.
BTW, some neat college stats here. Specifically, for offensive efficiency check out his leaderboard. Possessions used are in parenthesis as well, so combined it gives a good idea of how good they were overall on offense. Amongst players who used at leats 28% of offensive possession, Nick Fazekas is right at the top, with Redick and Morrison behind him. Roy would have been number 1, but he just missed the cut (27.5% possessions used).
I'd be interested to see how the rankings were for players coming out in past drafts. I guess it would be the only way to determine if this really is that accurate.
<div class="quote_poster">Quoting Chutney:</div><div class="quote_post">I'd be interested to see how the rankings were for players coming out in past drafts. I guess it would be the only way to determine if this really is that accurate.</div> I emailed the author, and he told me that he tried it last year and his formula accurately predicted their "success" at a 2/3 clip.
