NBA Draft Prospects in the Final Four

Feb 13, 2016; Villanova, PA, USA; Villanova Wildcats guard Josh Hart (3) shoots the ball past St. John
Feb 13, 2016; Villanova, PA, USA; Villanova Wildcats guard Josh Hart (3) shoots the ball past St. John /
facebooktwitterreddit
Villanova Wildcats guard Josh Hart
Villanova Wildcats guard Josh Hart /

As an NBA-centric basketball fan, the chance to watch potential draft prospects against high level competition is always part of the draw of March Madness for me.[1. The craziness of a high stakes single elimination tournament combined with college player decision making is also compelling in its own way.] Last year, draft fans were spoiled in the Final Four with a loaded Kentucky team featuring prominently, with Duke, Wisconsin and Michigan State all sporting potential first round picks as well. This year’s Final Four teams are more sparse in terms of potential high end NBA, but there are a few players that could hear their names called either by the commissioner or the assistant comish this June.

Below is a graphic I have used in the past to visualize how prospects are evaluated by my PAWS draft model. It breaks the statistics evaluated in the draft model into representative categories and standardizes them to the same scale with average scores centered at zero. So, if a player is average at scoring compared to the other prospects in this class, he gets a zero in scoring. If he is in the top thirty-three percent in rebounds he is approximately standard deviation above average and that colored bar shows up as plus one, in the bottom third he would show up as below the center line at about negative one.

In addition, there is a plot for the overall draft model score for each player[1. A score of 4.0 is approximately average for a drafted player in the training model and indicates an end of the first round, beginning of the second round player, 6.0 and above indicates top ten talent and in the 5.o indicates late lottery/teens].

image (3)
image (3) /

Some of the patterns are simply the result of the player’s position. For example, we expect point guards to be below average rebounders compared to all prospects and above average at assists. Often I use this style of chart to compare positionally prospects for just that reason, but there are still things to be seen in the comparisons above.

For those who prefer tables, here’s the chart information in table format with additional information on each player’s best and worst rank in different model iterations as well as the median rank.

Final Four prospects
Final Four prospects /

Buddy Hield, for example, is the most high profile prospect left in the tourney and the highest in most draft boards. He performs less well than the current hype might indicate in my draft models. In some others, but not all, mine included his score looks like a second rounder not a lottery pick[2. Worth noting that this year looks to have a particularly low correlation between draft boards and analytic models after a high degree of correlation last year. This is probably due to some of the more highly rated freshman underperforming this year.] In the chart it is apparent why Hield is rated as mediocre in the model – he is very well-rated as scorer but is below average in every other category. The particular problem with evaluation is that scoring is an immensely important skill but that scoring is the statistical aspect that, historically, is the most difficult to project moving from college to the NBA. I am not at all at the #NeverHield point, but taking him top ten is a gamble that the aspect of his game that historically have translated better will improve while the least reliable stat will hold steady.

Tyler Lydon, on the other hand, is generally rated lower by mock drafts than my model and others have indicated. Lydon is young and at least average in every category, including being well above average in the blocks and steals category. Potentially Lydon could stretch the floor and protect the rim, as Kevin Pelton termed an NBA unicorn, with one significant caveat is that he plays in the Syracuse zone, a scheme that likely inflates the block and steals group of stats.

Brice Johnson rates similarly in draft boards and my models, a likely mid-first round selection if he enters this year’s draft. His only below average category is distribution, though perfectly respectable for a center.

Michael Gbinije and Marcus Paige are rated as potential second round picks, at best, given their ages and lack of dominance. Paige is only above average in distribution a red flag for a 23 year old point guard. Gbnjie suffers primarily due to his age.

Justin Jackson is scored at the beginning of the second round in my primary PAWS model, and lower when scouting ranks are given greater weight.

Lastly, Josh Hart is rated as a marginal prospect by draft express, but the junior shooting guard has a solid all around game push him to rate as a potential late first round selection via my models. He’s the player I will probably have my on the most to try to gauge whether the model is picking up something real or not.

As high drama as the national championship itself is, for these players, amateurism aside,with the attention of the basketball world on them over the next couple of days they’ll be playing for real money this weekend.