If you're a basketball fan — particularly an NBA fan — you've undoubtedly heard by now that the Golden State Warriors point guard Stephen Curry won this year's MVP award. The only other real contender was James Harden of the Houston Rockets, but also in the mix were LeBron James (as usual), Anthony Davis, and Russell Westbrook. Personally, I think Curry was the right pick, but no one could have argued all that bitterly had Harden won.

Needing to learn a little R, though, and inspired by a series of density charts looking at the game-by-game distribution of QB ratings for NFL quarterbacks, I decided to look at the distribution of GameScores (GmSc) among the top 5 MVP candidates for this year's NBA season. GmSc is based on John Hollinger's Player Efficiency Rating statistic, and while it has its problems, namely overvaluing scoring, it's good enough for my own modest purposes. The equation for GmSc is as follows:

GmSc = PTS + 0.4 \times FG - 0.7 \times FGA - 0.4 \times (FTA - FT) + 0.7 \times ORB + 0.3 \times DRB + STL + 0.7 \times AST + 0.7 \times BLK - 0.4 \times PF - TOV

The statistic is standardized so that a GmSc of 10 represents an average game.

The plots below were created using the ggplot2 package, which, as a n00b, I'm a fan of for its simplicity and the beauty of its graphical outputs.


Data: Basketball-Reference.com

As you can see, Curry consistently puts up a very good GmSc, with very few awful games to speak of, but also fewer of the cosmic rampages that Harden, Davis, and Westbrook produce. Both of the latter players are prone to having a few stinkers, though. Davis and Curry seemingly are the most consistent players, according to GmSc. The distribution for James looks to be bimodal, with one peak around 12 or 13, and another around 25.

Scatterplots by player show much the same, but here I think it's easier to get a sense of each player's spread. You can see the Westbrook and Harden flops, as well as the range and distribution of performances you could expect for each player. James looks a little more consistent in this view.


Data: Basketball-Reference.com

Nothing complex, and certainly not a full analysis of the candidates' seasons, but I figured I'd throw them up in case anyone was interested.

[Update 5/7: I failed to remark in the initial version that James looks more consistent in the second plot.]