A Higher Level of Competition, NCAA or the D-League?
USA Today Sports
The NBA D-League has always fascinated me as a bit of a no-man’s land. Some of the best players in the country play basketball for little money, not much adulation, but just to keep their NBA hopes alive. And some certainly make it.
In fact, last year Glen Rice Jr. became the highest draft pick to come out of the D-League, selected at 35th overall. P.J. Hairston is looking to do him a couple better and is currently ranked as the 26th prospect in the upcoming draft by Draft Express.
In addition, the NBA’s 10-day call up period is in full swing, where NBA teams can sign players, often out of the D-League, for up to two, 10-day contracts to give them a trial. Conversely, mid-season seems to be increasingly the time when teams in the playoff hunt send rookies who are not seeing the floor down to the D-League.
So, it seemed like a good time to do an analysis to compare the competition level between the NCAA and the D-League.
The method I used is made possible by the fact that most players in the D-League played in NCAA basketball. I picked current D-League players that recently played in the NCAA and compared their stats using a player metric called Alternative Win Score (AWS). AWS is one of a number of linear box score metrics that score players based on their positive and negative contributions on the floor.
A study by Neil Paine a couple of years ago found that AWS was one of the better such metrics at predicting winning percentages by teams, particularly for teams with a high degree of turnover in their roster. That indicates that AWS is giving us a fairly context-free measure of a player’s contributions and makes it a particularly good metric to look at prospects for the NBA playing in different environments.
One contextual issue that I did have to adjust for was pace. The D-League plays much faster than the NCAA, and faster than the NBA. Looking at Team Rankings for the NCAA and the NBA site for the D-League, I found that the median college team has 70 possessions per 40 minutes, while the D-League has 84 in same time period.
As a result, counting stats are well overstated for the D-League (not that one should put any weight in the counting stats in any case). But this also inflates the box score metrics, including AWS; any reasonably productive player, would record more rebounds or steals or other positive plays per 40 minutes, given more opportunities. It also accentuates any on-court disaster since the player would have more chances for turnovers, missed shots and fouls. But the players we’re interested in here aren’t, generally speaking, the on-court disasters. They’re the prospects.
In order to make the stats comparable I normalized them to the NBA’s median pace, slightly slower than the D-League.
For reference, dear departed HoopData listed the median AWS score as 4.6 per forty minutes last year, with former D-Leaguer Shelvin Mack as the median player. In that area are guys like Jason Smith, Marcus Morris, Caron Butler and Brandon Bass, who would all generally be considered rotation players at this stage of their careers.
- The median pace-adjusted AWS for the NCAA players before going into the D-League was 7.8 per 40, right around the level of Chris Bosh, Joakim Noah, and Carmelo Anthony last year. What I would term second-tier stars.
- Of the 31 players I looked at 23 had lower production in the D-League than they had the year before in college as measured by AWS.
- The median difference was -1.6 AWS per 40, putting their D-League production at the level of a good starter like J.J. Redick.
- On specific statistics, True Shooting Percentage fell a median of -1.7%, pace-adjusted rebounds per 40 recorded median of -1.65, steals fell 0.42 per 40 and personal fouls went up 0.6.
- It is important to note that all of the players in the analysis are at the age where we would we expect some improvement in their production against the same level of competition. So the fall in production in the D-League probably understates the difference in levels of competition.
- Conversely, there maybe some selection bias with players who over performed their final year in college returning to their longer term mean.
- There was a smaller subset that also had time on the court in the NBA, unsurprisingly there was a fall in production against the stiffer competition in the big leagues, though less than the gap from the NCAA to the D-League.
So are D-League players better than college players? Yes and No. The answer on both sides lies in their age, with D-League players being much closer to their peak than most college players. On the other hand their peak is lower than the one and done lottery picks.
In terms of P.J. Hairston, going through bigger, older more seasoned players in front of smaller crowds with a tougher travel schedule, he’s definitely taking the tougher road to the NBA draft. So far he’s one of the relatively few players who has produced more in his brief start in the D-League than he did last year in college. Thanks largely to his torrid shooting to date Hairston has a AWS of 12.7, which would be Durant-ian numbers, if sustained. Hairston’s other numbers, like rebounding, turnovers and personal fouls have seen the typical degradation jumping to the D-League from college, so the bump may not last. But if does he will certainly have earned his way.