NBA Combine invites validate statistical NBA Indicators and DraftExpress, while NBADraft.net strikes out with snubs of Crowder, DJO and others

The announcement of the 60 players invited to the NBA combine this week confirmed something I’ve observed for the last several weeks; while the mock draft on Draft Express  is certainly missing a few players who will contribute at the NBA-level, it is DraftNet that bears almost no resemblance to a statistical analysis of which college players will make the jump. (scroll to the bottom for the table below of which mock draft missed which players)

The combine list is the first true indicator of who the NBA teams believe will be drafted.  There are 60 players invited to the combine and 60 players are drafted, so except for a few extra European players who will make the draft without going to the combine – this is basically the list until some underperform at the combine itself.

While I am sure both mock drafts will become much more accurate after being adjusted for who is invited to the combine and how they perform there, DraftNet was off on an incredible 19 players heading into the mock draft, more than twice as many as DraftExpress, whose mock much more closely mirrored my statistical analysis.

Adaption of Statistical Analytics to NBA Drafts a proven winner

A decade ago stat-skeptics may have challenged this proposition when Mark Cuban paid Jeff Sagarin and Wayne Winston over $100,000 ayear to start guiding moves with statistics.  However, in the ten years before that decision the Mavericks had a .303 winning percentage and after hiring Sagarin and Winston they had .691 winning percentage capped by a World Championship.

Maybe a few stat-skeptics were still around when the Indiana Pacers hired another of the true statistical masterminds, Kevin Pelton, in March of 2010 while suffering through a 32-50 season that would be their fourth straight out of the playoffs.  The answer to the headline, “Can Numbers Turn Around the Pacers?,” after that hire was also yes, as just two years later they have gone from doormat to giving the Heat a scare as the 3-seed in the East.

Why we can’t print statistical NBA indicators like we do Value Add

The Spurs in-house analytic team is one of the best in the NBA, and their ability to evaluate college players that are undervalued has made them the virtual Moneyball team of the NBA but with equal resources.  The difference between the Spurs and Billy Beane is that Beane lost all competitive advantage when Moneyball the book was printed and the teams with more money could adapt his statistical analysis, while the Spurs analytics effort has been kept close to the vest and therefore enabled them to maintain a huge advantage in statistical analysis over most teams even though they have received few high draft choices due to such consistent great play over the years.  The Spurs get the most wins per dollar spent year after year.

So while we are all happy to make public who we believe are the best/most valuable college players on Pomeroy, Value Add, or other sites, the public and draft sites are not aware of the statistical measurements of NBA Indicators.

If Sagarin ran a site updating the public and other teams on the draft order he was recommending to the Mavs and Pelton did the same team for who his stats said the Pacers should pick, then their statistical analysis would lose all competitive value because other teams would grab the identified undervalued players in the middle of the first round right before the Mavs, and later in the round right before the Pacers.

Likewise when I meet with an NBA team and they ask me to run numbers on three specific players they are looking at as surprise picks, I keep that in confidence.  I don’t go to the next team and say, “hey, you may want to look at these three guys I just heard about.”  In fact, I never even let an NBA team know what other NBA teams I’ve talked to. 

I can only pass on historical questions such as one I recently got from an NBA team about how our system projected David Lee to be a strong NBA starter by his fourth NBA season while every other statistical analysis said he was a likely bust.  In that case, I can convey to  other teams the flaws in competing evaluations that kept them from seeing Lee’s potential, and how we handle them, but even in a historical case I’m not going to print the formulas like I do with Value Add, because then every other team would simply run our NBA Indicator numbers and know our evaluation of each players’ NBA potential before I even told the team with whom I was meeting.  In the case of Lee, our system was the only one that works, as from his fourth through eighth seasons he has averaged over 10 rebounds a game while his rookie points per game of 5.1 has increased all but one year to 10.7, 10.8, 16.0, 20.2, 16.5 and 20.1.

So while there may be a small gap in statistical knowledge of how good a college player based on whether a fan believes Crowder is the 2^nd best player in the nation based on his Value Add, or the 8^th best player as rated at www.kenpom.com or one of the top 10 players as selected by the AP All-American voters or a little lower based on some other evaluation, the knowledge is all public.

However, there is a huge gap between public knowledge and the statistical evaluation of a player’s NBA potential and where he should go in the draft since anyone providing that knowledge to NBA teams has to keep it out of the public domain and even secret from other teams, and therefore you end up with a situation like this year where prior to the combine the intel of www.nbadraft.net was nowhere close to reality, while the www.draftexpress.com was pretty close.

So while I keep strict confidence on what NBA clubs tell me in meetings – never divulging it publicly or in conversations with other NBA clubs – there is a consensus out there on a couple of things pertaining to this draft.  Every team I’ve met with knows before I can tell them that Anthony Davis is BY FAR the best player in this draft if not in several drafts and Jae Crowder is the top statistical sleeper in this draft no matter which exact methodology is used by a particular club.  Even if a particular analytics team isn’t weighting a particular stat properly to measure the top prospects, they would have to really mess up the math to conclude that there was a better player than Davis or a better “sleeper,” as in player who would not have been considered a prospect coming into this season, as Crowder.  Analysis is getting better and better, and that means fewer Jeremy Lins are going to be missed, just as fewer college stars slip to Division 2 or 3 anymore.

But because this info is kept close to the vest until the combine list is reveals, DraftNet didn’t even have Jae Crowder or Darius Johnson-Odom in their mock draft, or for that matter Jared Cunningham (Oregon State), Marcus Denmon (Missouri), Tu Holloway (Xavier), Bernard James (Florida State), Orlando Johnson (UC Santa Barbara) or Kyle O'Quinn (Norfolk State) – all of whom were tabbed for the combine.  Likewise, their mock draft included Robert Sacre of Gonzaga, Maalik Wayns of Villanova and five other players on whose stats indicate they are great college players but unlikely to be able to contribute at the pro level.  While any of those players could play poorly in the combine and drop out to create a spot for a lesser prospect like Sacre or Wayns at the end of the draft, the fact is that as of now DraftNet’s intel has been way off.

While we can certainly give them a pass to them on four combine participants that both mock drafts missed in Hollis Thompson (Georgetown), Miles Plumlee (Duke), Robbie Hummel (Purdue) and Tony Mitchell (Alabama), the fact is that even without breaking confidences and revealing the specific players my program and the internal analytics of several teams indicate will be NBA contributors, DraftExpress certainly appears to be operating on much better intel.

That is good news for Crowder, who is a 1^st rounder based purely on statistical NBA Indicators and still a solid pick at 43^rd pick in Draft Express, and DJO, who has slipped a few spots but is still a solid 47^th pick of 60 in Draft Express heading into the combine.  While admittedly other considerations such as height and the potential to move down a position in the pros will be balanced against stats and the incredible strength and near zero percent body fat on both players – the fact that DraftExpress was so much more accurate overall is a good sign that both have a chance to shore up their draft status at the combine rather than having to leapfrog other players to get into the draft.

While the majority of combine invites were easy to predict, the following is the complete list of any player that was missed by either of the mock drafts:

Player	Team	DraftNet	Express	Combine	Who missed?
Quincy Acy	Baylor	yes	no	yes	express
Jae Crowder	Marquette	no	yes	yes	net
Jared Cunningham	Oregon State	no	yes	yes	net
Marcus Denmon	Missouri	no	yes	yes	net
Kim English	Missouri	yes	no	yes	express
Terrance Henry	Mississippi	yes	no	no	net
Tu Holloway	Xavier	no	yes	yes	net
Robbie Hummel	Purdue	no	no	yes	both
Bernard James	Florida State	no	yes	yes	net
Orlando Johnson	UC Santa Barbara	no	yes	yes	net
Darius Johnson-Odom	Marquette	no	yes	yes	net
Greg Mangano	Yale	yes	no	no	net
Khris Middleton	Texas A&M	yes	no	yes	express
Tony Mitchell	Alabama	no	no	yes	both
Cameron Moore	UAB	yes	no	no	net
Kyle O'Quinn	Norfolk State	no	yes	yes	net
Miles Plumlee	Duke	no	no	yes	both
Robert Sacre	Gonzaga	yes	no	no	net
Henry Sims	Georgetown	yes	no	yes	express
Hollis Thompson	Georgetown	no	no	yes	both
Casper Ware	Long Beach State	no	yes	no	express
Mitchell Watt	Buffalo	yes	no	no	net
Maalik Wayns	Villanova	yes	no	no	net
Alex Young	Denver	yes	no	no	net