"My rule was I wouldn't recruit a kid if he had grass in front of his house.
That's not my world. My world was a cracked sidewalk." —Al McGuire

Thursday, June 28, 2012

NBA Draft pick-by-pick calculations of how many wins each player will get his team … updating throughout draft

By Rob Lowe and John Pudner

NEWARK, NJ - This table will be updated throughout the NBA Draft tonight as players are selected, but we are going to take the leap of faith that Anthony Davis is picked 1st and list him ahead of time.  We calculate that a team of players as good as Anthony Davis would win 91.3% of their games in the NBA within a few years, and a team of players as good as the 2nd best player available in this draft would win 68.3% of their games.  This table is followed by a key to the terms and then notes on why Rob and I are just revealing each projection AFTER the selection is made. (first round complete below)

DraftNameGradeMockOur ProjWin%Notes
1Anthony DavisA1191.30%While an easy choice, they get “A” for not only the best choice available, but one of the best choices going back through our 2006 calculations. His 91% is first, and 2nd is 68%.
2Michael Kidd-GilchristD4461.00%NBA GMs almost always draft NCAA champs too high, and while Davis is certainly No. 1, Kidd-Gilchrist is way behind the No. 2 and No. 3 pick.  Still a very strong player, but they gave away wins by going again with an NCAA title winner too early (Kemba last year).
3Bradley BealA3268.30%A huge drop from Davis, but stats show he should be the 2nd best player to come out of this draft, with dominance in every category that predicts NBA greatness.
4Dion WaitersC-81254.50%Numbers not quite there, but with his phenomenal steals and very good Value Add and Usage, you can see him as a 4th pick.
5Thomas RobinsonA2368.20%We project he will be 3rd best NBA player, but in a virtual tie for 2nd as his incredible defensive rebounding.  With the gap between the top 3 and the rest, he is a steal.
6Damian LillardD71950.60%A lot of teams liked him here, but the amount he handled the ball (too much) actually predicts AGAINST truly excelling in the NBA - but we do have him as a starter and his vertical was almost as good as DJO. If he had his numbers as a freshman this would be a good pick, but as a junior they are too low.
7Harrison BarnesD52841.20%All of his college stats are below NBA-level, projecting a solid reserve, BUT his physical combos were unreal.  So far we've found stats a better predictor than combine - but if he excels it could move the future model a little heavier on combine results.
8Terrence RossC-161453.50%We don't quite have him there, but the combo of rebounding, steals and an incredible vertical/speed combination make him a solid starter with upside (also hope that combo plays out for DJO).
9Andre DrummondC61155.10%His Value Add is about average for a freshman bound for the NBA, and physicals are great - just such a weak defensive rebounder and weak on the bench press project him outside the top 10. We advised strongly against him as a top 5 or 6, but 9 is reasonable.
10Austin RiversC-102939.30%I may lose my Marquette credentials, but we really have a lot of red flags on him.  But we assume Doc's lineage is more important than our stats, so best of luck.  
11Meyers LeonardA13856.90%Huge get.  Drummond a little better in most things, Leonard's huge edge in the bench and defensive rebounding predict better career.
12Jeremy LambC91751.80%Can't put this below a "C" because his physicals are SO good (wingspan, speed and agility) and running the model last year he looked like a top 10, so I have to attribute the dropoff to UConn's wacky year and say he is solid at 12.
13Kendall MarshallC-182345.30%Again, it appears owners are putting high stock on vertical/speed combos.  While we have him lower and outside of his incredible assists the stats aren't good, but point guards are the hardest to predict so I can only go C-.
14John HensonA11658.70%A fantastic pick-up for the guys who share the Bradley Center.  While Value Add is low for a junior who expects to play in the NBA, his defensive rebounding and reasonable usage more than offset them to point to a strong NBA starter.
15Moe HarklessC201851.40%Very solid pick.  He is better than average in 9 of the 10 biggest predictors of NBA talent - only a very poor bench dragged him down a few spots.
16Royce WhiteC-212050.10%Right there as a borderline NBA starter.  His Value add was only 10th percentile for draftees, but other than that his overpowering rebounding and good steals for a big man make this a solid pick if just a couple of spots high.
17Tyler ZellerA12757.60%Very similar to Henson - only negative is a senior bound for the NBA should have a higher Value Add, but Usage and Def. Rebounding predict strong starter.
18Terrence JonesAC-956.30%Almost sure strong starter within a few years.  His Value Add and defensive rebounding is average, but every other physical and statistical measure predicts a great NBA player.
19Andrew NicholsonC-222444.80%Can be a strong rebounder off the bench, but a senior playing against lower competition with his numbers is just a notch below and NBA starter, and they could have done a little better with the 19th choice.
20Evan FournierNA305423.00%We haven't calibrated European stats yet, so can't pass judgment either way.
21Jared SullingerA15559.80%Stats say he will be 5th best, so we hope the back is ok and he becomes a steal
22Fab MeloC252741.30%You can't draft size, but on every other indicator Melo looks like just a good NBA reserve.  So little data though with part of one season, he could surprise.
23John JenkinsD313632.80%Again high vertical seems to put someone ahead of where he projects - but otherwise really looks borderline to even be in a rotation.
24Jared CunninghamC-343334.10%Steals are so off the charts that you can't dismiss this pick, even though overall this looks a few spots too high.  Obviously teams are filling roles now, and Cavs may need some pressure defense.
25Tony WrotenD274526.80%Looks like a longshot overall, but again it's a verticle/speed guy with a ton of steals in this case, so those three seem to have been weighted the most heavily by teams this year.
26Miles PlumleeB372246.60%Vertical, strength, agility, great speed for a center, physically unreal and domaint rebounding - he is at least a very strong backup within a few years and they might even have a starter at the 26th pick - which doesn't usually happen.
27Arnett MoultrieA192149.10%Again, speed and vertical (87th percentile and 88th percentile), but in this case he is a dominant rebounder with wingspan who is a borderline starter.  Clearly the best available - surprised they traded him to 76ers.
28Perry JonesC172641.40%He was viewed as slipping from what the public thought, but this is only one spot behind where we project him.  Speed and vertical at 92nd percentile for a big guy and 90th percentage in steals.
29Marquis TeagueC234031.10%Again, picked higher than how well we project he will perform, but with 98th percentile in vertical and 93rd percentile in speed you can see the appeal.  With that potential for a freshman on a team with 6 NBA guys, hard to say a PG pick here isn't solid.
30Festus EzeliC324427.10%A little high - doesn't project to be in a rotation, but as strong and long as he is and being out injured this year common sense says a decent pick.

Key to Column Headers:

Grade – this is the Grade the TEAM gets for the pick, not the player.  If the 30th best player is picked 5th, the team will get an “F” because they had much better options.  If that same player were picked 55th, the team would get an “A” because they identified a steal who will likely produce many more wins than a normal 55th pick.  If the Hornets picked someone else they would obviously get an “F” for the pick since Davis projects as competing for the top draft option since our system tracks back to 2006.

Mock – we averaged 11 mock drafts together, and this was where the public thought the pick would go.

Our projection – Keep in mind we have built a model that projects how GOOD a player is going to be in the NBA, not where we believe a team will draft him.  If we were calculating that, we would base it more on points per game and players on NCAA champions, since things like that lead teams to draft players too high.  If we project a player as #9, that means he projects to be the 9th best NBA player of this group within a few years – not that we actually expect him to go 9th in the draft.

Win% - this is the percent of games a team would win if all of their players were as good as this player will be within the next few years. A 50% player projects to start within a few years for a sub-.500 team and be a strong reserve for a team over .500.  A 50% is equal to a .100 Position Adjusted Wins for those of you used to that figure.


The last couple of months have been tough for us, because we love to calculate and write.  We have built a statistical model that has been accurate at predicting how many wins a player will create for his NBA team based on his college stats, what year he was in school, and his physical combines.

However, we realize statistical projections will always be only PART of the evaluation process. In addition, we have to maintain confidentiality for groups that provided feedback about the evaluation process.  Therefore, it is only fair to reveal the projections tonight AFTER the picks are made – and remember we are predicting how good the players will be, not when they will be picked, so you can judge us on the players performance over the next four years. (John's post script - by "we" I mean 65% Rob and 35% me - my colleague is a genius!)

Hopefully seeing a young DJO fan is a good omen for the second round.  If you don't believe Cracked Sidewalks has clout with the NBA, look at the view from my seats (taken an hour before the draft, the seats did fill in later.)

No comments: