What are the Odds D’Angelo Russell Becomes a Star?
As soon as news broke that the Brooklyn Nets had acquired D’Angelo Russell from the Los Angeles Lakers, I was fascinated. As a Nets fan, I suddenly had a rooting interest in the development of a player I had previously never followed. The price for the Nets was steep as all three other pieces moved benefited the Lakers. Los Angeles gained one year of Brook Lopez, a late first round pick, and the opportunity to shed Timofey Mozgov’s albatross of a contract. But Russell was the second pick in the 2015 NBA draft, the type of high upside player that a team starved of draft assets would reasonably salivate over.
What immediately interested me about Russell was that I had no idea how to properly value a player with his resume. He is only 21, playing a premium position (point guard), with undeniable offensive potential as evidenced by his high selection in the draft. On the other hand, we have already witnessed two seasons from him that were at least mildly disappointing. After all, he has already been traded. Despite playing on a bad Lakers squad, his team was 4.6 points per 100 possessions worse last season when he was on the court (per NBA reference).
D’Angelo Russell represents a particular player prototype: the high lottery pick who is still very young, but whose career production to date has been mediocre to bad. How should we properly value such a player?
To answer this question, I have built a model which identifies the 20 players drafted from 1990-2007 who are the most “similar” to a young prospect. I then examined how their careers turned out to gain some insight into how our young prospect might perform.
I used a statistic called Win Shares, created by NBA Reference, to value player performance. Win Shares is a one number summary of performance that attempts to capture the overall value of a player. Of course it is not a perfect judge of value, but it is a reasonable estimator of performance and readily available on nbareference.com. To learn how Win Shares is explicitly computed, go here.
To measure a player’s career success, I use a statistic I call Prime Win Shares. Prime Win Shares are the total Win Shares accumulated over the 8 year period from their third season through their tenth season. To gain an idea of what constitutes a good or bad Prime Win Shares total, I made a histogram of Prime Win Shares.
The histogram shows Prime Win Shares is heavily right skewed, meaning that there is a “right tail” of star players who performed at a much higher level than the median player. For reference, the median value is 13.5 and the mean is 19.9. The first and the third quartiles are 1.9 and 32.25, respectively. LeBron James accumulated 133.4 Prime Win Shares, the maximum value.
To identify the past players who were most similar to the young prospect, I compute a similarity score for each player using four factors:
- A weighted average of Win Shares produced in the first few seasons which weights later seasons more heavily than earlier seasons
- Draft position
D’Angelo Russell’s Career Projection through Similarity Scores
To understand how D’Angelo Russell’s career might turn out, I plotted the Prime Win Shares of the 20 most similar players to Russell (Figure 2). For reference, I also plotted the Prime Win Shares of all drafted players from 1990-2007 who played at least three seasons.
The plot below is not very encouraging. Not only is his projected median outcome (the green square) only 15.6 Prime Win Shares, but the highest comparable player only achieved a value of 42.2. The problem for Russell is that his two season Win Share production ranks him at the 38th percentile of all players in my database, meaning 62% of players started their careers with higher production.
Of course, the fact that none of the 20 players most similar to Russell had star careers does not mean that Russell cannot reach this level. While the sample size of underachieving high draft picks in my database is small, there are a few examples of players who fit this profile and then became “stars” (if we define a star as at or above the 80th percentile in Prime Win Shares). There were 12 top five draft picks whose two season Win Share production ranked below the median. Of those, two – Antonio Daniels and Donyell Marshall – became stars. However, neither player would be considered a “superstar” (above the 90th percentile in Prime Win Shares). The sample size of underachieving stars is so small that it is hard to draw conclusions from this statistic.
What does this all mean for D’Angelo Russell’s star potential? I would say that is highly unlikely that Russell reaches the rarified stratosphere of 70 or more Prime Win Shares, which I would consider superstar level. It is nearly unprecedented for a player who starts out so poorly to rebound and become a superstar, even among top draft picks. The player comparisons indicate that, at this point, Russell’s upside potential is probably limited to being a very good starter. At least in terms of Win Shares, Russell’s upside is significantly lower than it was two years ago when he was drafted.
Projections of other Interesting Young Players
In the plots below, I use the similarity score method to project other interesting players from the 2015 draft class.
Karl-Anthony Towns Career Projection
Kristaps Porzingis Career Projection
Devin Booker Career Projection
Jahlil Okafor Career Projection
Myles Turner Career Projection
Rondae Hollis-Jefferson Career Projection
Photo via the Flickr creative commons, thanks to Keith Allison for the shot of Russell.
Statistics via Basketball Reference, unless otherwise specified.
Graphs are property of Thomas Bassine.