Which College Stats Matter in the NBA: Wing Edition

There is an abundance of information nowadays regarding draft prospects. Teams and fans know all their stats, their size and athletic measurements, and probably what their favorite show was as a child. But what actually matters? Prospects move up and down according to their interviews, how they perform at the combine, and what they are able to do in private workouts. While all those are important, it has always been my opinion that a player's performance in actual games is the most important indicator for NBA success. So I decided to do some research.

I found all the significant wing players (those who have played a certain number of NBA minutes) from the 2012-2014 drafts, and compiled some of their choice college stats. My goal was to see if any of these stats correlated strongly with player success, represented through several advanced stats (Win Shares/48, Value Over Replacement Player, and Real Adjusted Plus Minus). I was not seeking to develop a draft model to predict future success, or even try to pin down an exact prototype of a player who might be successful in the NBA. All I wanted was a little more insight into which college stats really matter when evaluating wing prospects.

Here is the table just for the 2012 draft. All the "basic" stats (and true shooting) are the prospects' college numbers, while the three advanced stats on the right are compilations of their NBA years. I used their actual per game statistics (not adjusted for 40 minutes) because the whole point is to see what actually translates to NBA success. If a player played two years or more in the NCAA, I averaged their numbers in their last two years to get the table results.

Taking together the entire pool of 2012-2014 wing draftees, the most important college statistics turned out to be rebounds, steals, and assists (for VORP and WS/48, anyway). Together, they predicted about 52% of the variance in those advanced stats' data, which isn't too shabby considering how simple the equation was. For both models, rebounds and steals were dominant- in fact, steals accounted for 37% of the variation in VORP! That's a lot for one simple number. Of course, there are issues with VORP itself, but there is little question for me that steals and rebounds are the most important numbers to look at when evaluating a college wing prospect.

But what about numbers that didn't mean anything? True shooting, across the board, was the least valuable college statistic. For each draft class, for the total draft pool, for comparing to Win Shares or real plus minus- it had the lowest correlation each time. It was essentially useless in predicting NBA success. While huge inefficiency in college might be somewhat of a warning sign, I would take any evaluation lauding a prospect's efficiency from the floor as a reason to draft them with a grain of salt.

The most interesting number, to me, was age. Usually, the top draft picks in the lottery are younger players who have the most "upside" due to their youth and potential to grow. And the regression equations said that being younger was somewhat helpful. At the same time, looking at the data, many of the best wing players from the last few drafts have been older. Jae Crowder, Rodney Hood, Khris Middleton.... all were 21 or over. So while it is a possibility that more superstars result from younger prospects, age overall doesn't have much correlation with general NBA accomplishment. Here is a more in depth look at the top players from each statistic and how well they bear out future success:

Top 5 rebounders: Andre Roberson, Kyle Anderson, Jabari Parker, Jae Crowder, Michael Kidd Gilchrist

Top 5 in steals: Crowder, Jordan Adams, Kentavious Caldwell Pope, Victor Oladipo, Roberson

Top 5 in assists: Anderson, CJ McCollum, Will Barton, Khris Middleton, Solomon Hill

Top 5 in points: Doug McDermott, McCollum, John Jenkins, Parker, Joe Harris

Top 5 in true shooting: McDermott, Jenkins, Nik Stauskas, Ben McLemore, Oladipo

Almost all the prospects in the assists, steals, and rebounds column have turned out to be at least NBA rotation players. The iffiest of them is Jordan Adams, who has played in only 32 games in the league-- but had a solid rookie season before he was derailed by injuries. Points and true shooting are mostly a barren wasteland of fringe players outside of McCollum, Parker, and Oladipo. These guys are still young, and can improve, but none of them are good rotation players right now. What about the worst performers in each category?

Bottom 5 in rebounds: Tony Snell, John Jenkins, Nik Stauskas, Joe Harris, PJ Hairston

Bottom 5 in steals: Doug McDermott, Stauskas, Tim Hardaway, Snell, Shabazz Muhammad,

Bottom 5 in assists: Cleanthony Early, Muhammad, TJ Warren, Hairston, Jenkins

Bottom 5 in points: Hairston, Kyle Anderson, Snell, Reggie Bullock, Andre Roberson

Bottom 5 in true shooting: Aaron Gordon, Khris Middleton, Harrison Barnes, Anderson, Hairston

The prospects who struggled to accumulate boards, steals, and assists are uniformly non-factors in the NBA other than Shabazz Muhammad and TJ Warren. On the other hand, some of the most promising NBA players struggled mightily with efficiency in college: Middleton, Gordon, and Barnes. In fact, Middleton is one of the most efficient high-usage wing players in the NBA, capable of piling up both 3 pointers and free throws. On the other hand, players who failed to do much outside of scoring at the college level seem to be unable to translate their games to the pros. Why might that be?

Lost in all the arguments about standing reach or vertical jump is how smart a player is on the court. And this is something that can't translate through interviews. Many intelligent players don't play smart, because basketball IQ is different from regular IQ. Rebounds, steals, and assists all are related to how well one understands the game. Assists are about making the right play at the right time, about finding a way to do things that open up the game for others. Steals might be aided by length and quickness, but they are also dependent on anticipation and reading what the opposing player is going to do even before they do. Rebounds are easier to snag for taller players who can jump higher, but they also require knowing the angles of a bounce, the strength of a shot, and the positioning of other players on the court. Athleticism is necessary to be successful, but smart athleticism is even more important.

This is by no means some conclusive study, nor is it entirely new information. But I do think it is significant. Ignoring players' scoring and shooting efficiency from the floor in college is not wise, but rebounds, steals, and assists ought to be weighted a bit more in evaluations. Beware the wing prospect who doesn't do much but score!