Will Analytics Trump Subjective NBA Draft Picks Again?

As soon as an NBA team announces a first-round pick June 26 they are promising the player $3 million to $14 million. Before investing that money, each team may want to look back at a packet Rob Lowe and I provided them two years ago that recommended a better player/investment based on www.valueaddbasketball.com and other factors than they chose. In our live coverage of the 2012 NBA Draft (click for Round 1 and Round 2) we were right 72 percent of the time we said a player would be BETTER than the normal production for someone picked at that point of the draft, and we were right 70 percent of the time we said the team should have used that pick on someone much better.

In summary, 2013-14 performances of the 60 players taken in the 2012 NBA draft indicate that our analytical recommendations to teams were a more accurate predictor of how much a player would actually produce than the actual team pick in 70 percent-plus cases in which the two were at odds. 

Our grade	Performance > draft position	Exactly met expectation	Worse	Our% correct
A or B	13	0	5	72%
C+ or C	5	3	5	almost perfect
C-, D or F	6	0	14	70%
Foreign	2	0	7	didn’t rate

While my colleague and Breitbart Sports contributor and NFL draft expert Cole Muzio has finished his radio appearances on the draft that has become bigger than most sports as a whole, NBA teams have just over six weeks to prepare for a draft that can have a much more immediate impact than an NFL Draft due to each pick potentially playing one of five starting positions on the court vs. one of 22 in football (offense/defense).

The following is a breakdown of each player, with the spot in which they were picked in the 2012 Draft.

72% of picks we graded as great have backed us up

(Hits) 13 picks we graded as an A or B have produced more than a player should if picked at their draft spot: Will Barton, 40th pick; Bradley Beal, 3; Jae Crowder, 34; Anthony Davis, 1; Draymond Green, 35; John Henson, 14; Terrence Jones, 18; Quincy Miller, 38; Kyle O’Quinn, 49; Miles Plumlee, 26; Mike Scott, 43; Jared Sullinger, 21; Tyler Zeller, 17.

The consensus was that Anthony Davis would be a relatively weak first pick compared to other years, while we countered that he was not only the best player in this draft but one of the best prospects going back to 2006. We went so far as to project that a team of Anthony Davis’ would win 91% of their games in the NBA. In just his second year, Davis is already the fourth best player in the league according to ESPN’s Value Added PER Rankings. Bradley Beal has joined John Wall to form the best young backcourt in the league.

But the bigger test is that six players we said should have gone in the first round (Barton, Crowder, Green, Miller, O’Quinn, Scott) have been either starters or played major roles off the bench despite not being taken until the second round.

(Misses) 5 picks we graded as an A or B have produced less than their draft spot: Marcus Denmon, 59; Kim English, 44; Meyers Leonard, 11; Arnett Moultrie, 27; Thomas Robinson, 5.

However, to admit our mistakes, we had five picks rated as an A or B that have not measured up. We thought Denmon would sneak onto the court a little as the next-to-last-pick, but he did not play this year. We also calculated that Leonard might be as good a center as Andre Drummond (see below), but he has not been. We could argue that only four of these have underperformed since Robinson has actually been very good but he was picked fifth and there were more than five players from this class better than him.

70% of those we ranked as duds were just that two years later

We gave 20 picks a C-, D or F because we thought they had picked players way too high.

(MISSES) We were only wrong in six players we rated as bad picks: Justin Hamilton, 45; Damian Lillard, 6; Khris Middleton, 39; Darius Miller, 46; Robert Sacre, 60; Tony Wroten, 25.

Lilliard has obviously been phenomenal and worth the 6-pick, and teams found value in the second round with the other five on this list.

(HITS) In 14 cases we were right as the player in question has NOT produced enough to justify where he went in the draft:  Harrison Barnes, 7; Jared Cunningham, 24; John Jenkins, 23; Orlando Johnson, 36; Kris Joseph, 51; Michael Kidd-Gilchrist, 2; Kendall Marshall, 13; Kevin Murphy, 47; Andrew Nicholson, 19; Austin Rivers, 10; Terrence Ross, 8; Tyshawn Taylor, 41; Dion Waiters, 4; Royce White, 16.

Of Barnes we said, “All of his college stats are below NBA-level…,” of Jenkins “…really looks borderline to even be in a rotation,” of Kidd-Gilchrist, “…NBA GMs almost always draft NCAA champs too high … Kidd-Gilchrist is way behind the No. 2 and 3 pick…,” and of Rivers, “we really have a lot of red flags on him.”

All 14 teams who went with one of those picks ahead of players we recommended in our pre-draft packets to teams would have almost certainly won more games this season with our picks.

13 picks we rated as C or C+ have broken out about as expected. Five have been better than their draft spot, five worse, and three have been almost exactly as expected.

5 C+/C picks have been better than their draft spot: Quincy Acy, 37; Andre Drummond, 9; Robbie Hummel, 58; Doron Lamb, 42; Perry Jones, 28.

Two NBA team officials told me during one-on-one meetings prior to the 2012 draft that the one projection that was way off was Drummond, and they were correct. One of them told me that occasionally a player is so physically dominant that you have to throw out the stats–and that Drummond was one of them.

Drummond was outplayed by many during his one year of college. For example, the 6-foot-5 Crowder dominated the 6-10 Drummond at UConn, putting up 29 points and 12 rebounds while drawing five fouls to send Drummond to the bench with just 7 points and 4 rebounds, but by the draft workouts it was obvious the sky was the limit for the then 19-year-old Drummond.

3 C/C+ picks have been almost exactly what you would hope for at their draft position: Bernard James, 33; Darius Johnson-Odom, 55; Jeremy Lamb, 12.

Lamb is a solid starter, Jones has made contributions on the bench, Johnson-Odom has gotten on the court – almost exactly what is expected given each of their positions in the draft.

5 C/C+ picks have been worse than their draft pick: Festus Ezeli, 30; Moe Harkless, 15; Fab Melo, 22; Jeffery Taylor, 31; Marquis Teague, 29.

We were very surprised to see big men Ezeli and Melo go when they did but left open the door that maybe they both had a little Drummond in them to produce well beyond our projections–but neither produced anything this year.

Foreign players (not graded), 2 of 9 have overproduced: Ognjen Kuzmic, 52; Tornike Shengelia, 54.

The seven that have not produced yet are: Furkan Aldemir, 53; Evan Fournier, 20; Ilkan Karaman, 57; Kostas Papanikolaou, 48; Tomas Santoransky, 32; Izzet Turkyilmaz, 50; Tomislav Zubcic, 56.

I have moved one step closer to rating foreign players as well by rating the level of play in each country vs. the D-League and NBA, but we did not offer evaluations of these picks in 2012. In all fairness on the seven “misses” above, teams often pick a foreign player and wait more than two years to bring him over–letting his current team pay his salary during that time.

While there are certainly things scouts can determine that analytical methods cannot, based on the 2nd year players this season, it looks like analytics would have suggested better picks than were actually made in the 2012 draft.

The projections were based partly on “Senior Value Add,” which is what a player’s Value Add (see players from 2003-2014 here) would have been projected to be if he stayed through his senior year. The rest of the calculation included a calculation of combine results and specific statistical combinations by stat guru Lowe that were proven “NBA Indicators” of future performance. As Lowe and I said at the time, we wanted to be judged not on if we nailed where players were actually picked, but by where they should have been picked based on how they actually performed. Based on their PER ratings, our system nailed future performance.