I’ve been interested in a single question: Where did Landry Fields come from? Why could no one predict in advance that he would be such an effective professional prospect?
The main reason it is so hard to project college players into the NBA is that it is hard to project effective scoring ability (Pts-FGAs-.5FTAs). In fact, I have found there is only a small correlation between the Effective Scoring per 48 that a player posts in college and the Effective Scoring per 48 that he posts in the NBA.
According to a study I just did, the correlation coefficient between the collegiate Effective Scoring per 48 average of rookies who have played at least 300 minutes as a pro and their professional rookie Effective Scoring per 48 average is only 0.24, meaning only 6% of the player’s ability to post effective points in the NBA in his rookie season can be explained by his ability to post effective points in college.
Moreover, I randomly chose 100 players with at least 3 years of NBA experience and found that the correlation coefficient only rises to 0.33, meaning the collegiate scoring ability only explains 10% of NBA scoring ability.
Generally speaking, you can expect a dropoff in a player’s effective scoring ability from college to the NBA. But the drop-off is by no means uniform, and does not always occur. The lack of predictability works both ways. Look at Magic Johnson. Magic Johnson’s True Shooting Percentage in the NBA was better than his True Shooting Percentage in college. How do you explain it? (Probably a lot of zone defenses in college).
You should be safe, though, choosing a college player with a collegiate TS% above 60%. But then you have the example of Michael Beasley. Beasley (whom I called a “lock” at the time of the draft) posted a collegiate TS% of 61% in his one NCAA season. In the NBA thus far he has posted a below average TS% of 51%.
On the other hand, you have Landry Fields. Through four seasons of collegiate basketball, Landry successfully hid his outstanding ability to transform scoring attempts into points. Actually, that’s not fair. Fields was a very good collegiate effective scorer. But, his collegiate TS% of 56% is in the “scary” range, meaning it is too close to the NBA average of 54%. If you factor in an NBA decline, he would be an ineffective scorer.
Instead however, Fields has actually INCREASED his TS% substantially in the NBA (61%), and he has had a tremendous rookie season partially as a result of that fact. But how has he done it? Better choices? Probably. Easier shots? Not necessarily (you would think most any shot in the NBA is more difficult than the same shot in college). Who knows.
So while Landry went one way, former Kentucky PF DeMarcus Cousins (aka “my boy”) has gone the other way and made me look absolutely foolish. At draft time I could not understand how anyone would pass on him. He had everything you want. High production, good size, high TS%, etc All the signs said “Cannot Miss”.
Then he went out and played… and thought he was Larry Bird. He has thus far scuttled his rookie season by making terrible decisions with regard to shot attempts, favoring the jumper over what should be his forte, the inside shot. As a result he has posted the largest decline between collegiate Effective Scoring per 48 and NBA ES48 among those rookies with 200 or more minutes of action.
Now, Cousins may yet turn out to be a productive pro. But he illustrates how hard it is to detect good scoring prospects. You cannot know how effective they will be at the “next level” when it comes to turning scoring attempts into points.
Another example of this is Evan Turner of the Sixers. If you look at collegiate performance, he and Cousins would have been among the players you would have expected to be high percentage scorers in the pros. So far, not so. Turner’s decline in effective scoring has also been substantial.
Therefore it is obvious that intuition still plays a role in the evaluation of NBA prospects. Nevertheless, in the next post I will evaluate this summer’s crop using their collegiate statistics as my main point of measurement.
I will be hedging my analysis.