Pre-Preseason NCAA Draft Projections

Last week I posted  competition adjustments that assess the relative difficulty of putting up boxscore numbers in different leagues around the world.  Included in that list of competitions are a number of prep all-star games and junior/cadet tournaments.  These are competitive venues that feature a number of incoming NCAA freshman.  Today, I am going through the top freshman NBA prospects (via DraftExpress) and pulling every instance I have for that player in my database.  The goal is to gave a very early projection of how each player looks as an NBA prospect based on the numbers.

Is projecting players this early a waste of time?

The one-and-done era creates a strange environment where most fans of the NBA Draft get their first look at the key players only six months before they are selected.  However, these players do have an extensive track record before they set foot on a college court.  This prior record is mostly the bailiwick of subjective scouting , which probably makes better use of tiny samples in alternative settings than number-crunching.  However, many players have produced enough across different prep and international tournaments to start taking a peak at what they say about those prospects’ NBA potential.

Unfortunately, I do not have any data from a some of the important extended prep venues run by Nike and Under Armour.  Obtaining and organizing these data will be a goal for the future, and may dramatically improve pre-NCAA projections.  Until that time, I am limited to mostly low-minute samples.  Projections based on limit minutes are necessary highly error prone, but I am of the opinion that trying to squeeze what I can out of limited information (while always keeping in mind just how uncertain my results really are) is better than pretending I have nothing to work with.  In fact, looking at the correlation between pre-NCAA projections and NBA production historically, the effort does not look like a complete waste of time.  The average correlation between projections based on a given prep all-star game is in the low teens, while tournament play has a correlation just over 0.2.  These are weak relationships, but combining players’ production across venues I can push that correlation up to just under 0.3.  This is still nothing to crow about, but since that correlation is almost exactly the same as the one between RSCI recruit ranking and NBA production, it probably is not something to be ignored either.

Pre-NCAA projections arguably add value beyond a few extra minutes of sample as well, even once the NCAA season has finished.  Team context shapes player performance in ways that complicate NBA projections.  There are ways to account for context using scouting and data, but triangulating on top prospects using multiple venues makes this process a lot easier.

Projections:

To make these projections easy to interpret, I use percentile scores.  Every performance rates somewhere between 0 and 1, and corresponds to the percent of players selected in previous drafts who produced like an inferior prospect .  That means as a player’s score approaches 1, he look like an increasingly better prospect.  Some handy heuristics to work with:  50th percentile or higher looks like a 1st round pick, higher than 75th percentile looks like a lottery pick, anything at or higher than 0.983 and we are looking at a performance strong enough to warrant a #1 pick.

I will be light on the player description below, but I recommend you read about these guys from someone who has actually watched them play beyond highlight clips for more context.

Note:  These projections come from a model that is trained to measure ultimate NBA potential, rather than expected NCAA impact.  That said, there should be a pretty strong correlation between the two.  It may be rational for college fans to tilt their optimism/pessimism about the coming season slightly in response to these model outputs.

#1 Skal Labissiere (Kentucky), Center

 Skal has been touted as one of the top two picks in the coming draft, and even though I only have one complete NBA game’s worth of minutes from him, that is exactly how he projects.   I am going to join everyone else in saying Skal looks like a top-tier prospect.

#2 Ben Simmons (LSU), Forward

 Ben Simmons’ statistical production also lives up to his billing.  Simmons’ minute-weighted average is in the 78th percentile, but that projection is heavily diluted by performances two years removed.  If we focus on recent play, Simmons looks like a top pick.  Also worth noting, Simmons looks like a clear power-forward with minimal tendency towards the wing based on my position projection model.

#3 Brandon Ingram (Duke), Forward

 Ingram has two cases to look at, and posted very different projections in each.  Ingram’s McDonalds All-American numbers tag him as a #1 pick, while his Hoop Summit performance makes him look like a second rounder.  As is true with everyone on this list, it will be interesting to see how he looks as more data comes in.

#4 Jaylen Brown (California), Forward

 Brown joins Skal and Simmons as the third prospect with both the hype and consistent production of a top pick.

#6 Jamal Murray (Kentucky), Shooting Guard

Jamal Murray blesses us with a relatively large sample of production.  Unfortunately for him, with the exception of his dominant performance at the Nike Hoop Summit, Murray’s track-record does not help his case as a future top-10 pick.  I cannot stress enough that all of these projections are only weak estimates that should be updated quickly in response to better data, but Murray’s mediocre production in extended minutes on the international stage sets my expectations for him much lower than his current hype.

#8 Henry Ellenson (Marquette), Forward

 The FIBA World Cup Junior and Cadet events carry more information than most of the other pre-NCAA competitions.  In recent seasons, guys like Okafor, Parker, Smart, Gordon, and Winslow distinguished themselves on the international stage.  Ellenson put up the best numbers from a prospect standpoint (just edging out Giles and Swanigan) on the 2014 USA U17 team.  Don’t be surprised if Ellenson has a major impact in Milwaukee this winter.

#10 Diamond Stone (Maryland), Center

 Stone is the first guy on this list who really sets off some alarms.  In about 250 minutes of pre-NCAA play, Diamond Stone has failed to distinguish himself as even a first-round prospect.

#11 Cheick Diallo (Kansas), Center

 Diallo produced like an elite prospect in all three of his recent prep all-star appearances.  Especially if Diallo does not get a chance to flash his skills on a big stage this season, he might end up being the 2016 steal-of-the-draft.

#15 Malik Newman (Mississippi State), Guard

 Newman gets the biggest red-flag of all the prospects on this list.  Over what is ultimately a reasonably large sample he does not look like a 1st round prospect.  Hopefully Newman can prove me wrong at Mississippi State this season.

#18 Isaiah Briscoe (Kentucky), Guard

 Draft Express has Briscoe projected as the last freshman off the board.  With the exception of his FIBA Americas performance last season, Briscoe has produced like he should be one of the first.  Unfortunately for Briscoe, that international performance was by far the largest sample I have on him, so the bad outweighs the more recent good until more data comes in.

The others:

The players above are far from the only incoming freshman worth considering, and if this season is anything like previous seasons, by the time the NBA Draft rolls around some of the guys above will have disappointed and other freshmen will have taken their places.  Below I list all of the incoming freshmen and their projections based on aggregated pre-NCAA performance.

A few cases above really stand out to me.  Luke Kennard has played well enough over a decent sample to look like a guy who might emerge this season.  Louisville’s incoming freshmen may be surprisingly dominant, though neither is projected on a large enough sample to warrant too much excitement.  This statement would have been even more true if Blackenley did not choose to de-commit and head to LSU.  Jalen Brunson and Tyler Dorsey both look like solid prospects on better samples than I have for most players.