Projecting whether or not a prospect can shoot is one of the hardest things to do in the scouting community. There has never been a real, concrete way to accurately tell. Since the college three-point line is shorter than the NBA line, the college 3pt percentages aren’t reliable. Also, some prospects shoot incredibly high or low from three, but have good/poor free-throw percentages or shoot a low number of threes. So, I tried to determine a way to judge a college 3pt shooter.
I started by using the Nylon Calculus formulas created by Andrew Johnson. Johnson created two different formula to judge shooting back in 2014. The formula are as follows:
Formula 1: NBA 3 Point % = .175 + .128 * Free Throw Percentage + .00449 * 3PTA per 40 + .163 * Three Point Percentage
Formula 2: NBA 3 Point % = .22 + .01571 * 3pt Made per 40 + .1389 * Free Throw Pct
Why those specific variables? Here’s Johnson’s reasoning:
Free Throw Percentage basically conveys information about the player’s shooting stroke without the ‘noise’ that can effect any particular three point attempt from shot clock, pass placements, or opponent contests.
3PTA per 40 is a variable that I think one needs to be a bit careful interpreting. In essence, it conveys something of the confidence both the player and their coach have in their three point stroke. On the other hand we can’t necessarily interpret as a causal variable, Josh Smith and Charles Barkley spent their careers proving that jacking up threes doesn’t necessarily cause your accuracy to rise.
Three Point Percentage, of course, mimics most closely the variable we’re looking for: three point accuracy in the NBA. But the high variation and relatively low number of attempts in pre-NBA shooting make that number unreliable as a predictor.
Step 2 was a slight adaptation of these formulas. I noticed that the projections for Formula 2 tended to be higher than usual when compared to the real percentages for NBA players. So, I changed Formula 2 to the following:
NBA 3pt% = 0.22 + 0.00808 * 3pta per 40 + 0.115 * Free Throw Percentage
This seemed to lead to slightly more accurate projections.
Step 3 was taking a sample of NBA players and comparing their college projections to their actual percentages. I took a sample of 100 players from the 2007-2017 draft classes. On average, Model 1 over-projected the percentages by .167% and Model 2 over-projected the percentage by .006%. Both models got 24 of the 100 players within 1% of their real percentage and 47 of the 100 within 2% of their real percentage.
The takeaway? That number can’t perfectly tell 3pt percentage. There is no perfect number. Just for reference, both models over-projected Michael Kidd-Gilchrist by about 12.5% and under-projected Jayson Tatum by 7.1%. However, sometimes the numbers work. Model 1 predicted Jimmy Butler and Gary Harris perfectly, while Model 1 missed Anthony Davis, Terrence Ross, and Eric Bledsoe by .01%.
Scouts need to use both the film and the numbers to predict shooting, not just one or the other. This shouldn’t come as a big surprise to basketball fans, but the marriage of the numbers and the eye-test is the best way to go about an evaluation.
With all that in mind, here are the projections for some members of the 2018 Draft Class:
|Name||3PT%||3PTA/40||FT%||Model 1 Score||Model 2 Score|
|Gary Trent Jr.||0.402||7.7||0.876||38.7%||39.0%|
|Jaren Jackson Jr.||0.396||5||0.797||36.4%||35.7%|
|Wendell Carter Jr.||0.413||1.9||0.738||34.5%||32.0%|
|Marvin Bagley III||0.397||2.1||0.627||32.9%||31.1%|