Stephen Curry’s Three-point Barrage

April 2, 2013

From Mark Price to Ray Allen, Dennis Scott to Mookie Blaylock, and Reggie Miller to Peja Stojakovic, the NBA has seen its share of great three-point shooters. Of course, there are those that shoot obscene percentages, but could never do so if they were a focal part of the offense. Then there are those that huck it up a lot, but aren’t the most efficient. To be truly great at three point shooting, I’m looking for the combination of volume and efficiency.

Here are the ten best seasons in terms of total threes made:

Rank Player 3PT Season 3P% Multiplied
1 Ray Allen 269 2005-06 0.4119 110.8
2 Dennis Scott 267 1995-96 0.4252 113.5
3 George McCloud 257 1995-96 0.3791 97.4
4 Jason Richardson 243 2007-08 0.4057 98.6
5 Peja Stojakovic 240 2003-04 0.4332 104.0
6 Stephen Curry 236 2012-13 0.4547 107.3
7 Mookie Blaylock 231 1995-96 0.3708 85.7
Peja Stojakovic 231 2007-08 0.4408 101.8
9 Ray Allen 229 2001-02 0.4337 99.3
Reggie Miller 229 1996-97 0.4272 97.8

Stephen Curry currently sits at 236 triples this season, just 33 shy of Ray Allen’s single-season record of 269 set back in 2006. Perhaps more impressive, Curry has made them 45.5% of the time, and that’s best on the list. The three-point makes record will be tough for Curry. With eight games to play, he would have make more than four threes per game to break the record, a tall order for a man that has made 3.37 per game this year and just 2.64 per 38 minutes on his career.* But regardless, he will arguably become the most efficient three-point volume shooter for any season ever.

Now, let’s get back to breaking Ray Allen’s record.

Stephen  Curry Three Point ShootingAssessing the probability distribution of Curry’s three pointers is tricky. I would be inclined to go with a Poisson distribution, but it doesn’t fit all that well as you can see in the picture (blue curve). Even if I remove his 11-three outlier performance, it it still a bad fit (yellow curve).** It’s probably most appropriate to use a compound distribution like a good Bayesian would, but I have little time for such things. We’ll go ahead and evoke the Central Limit Theorem for eight games, if you guys don’t mind.

If we assume that his standardized distribution over the next eight games will follow a T-distribution, then the probability of Curry making at least 33 triples is about 17.8%.

Good luck, little Warrior.

*Curry is currently averaging 38.2 minutes per game this season.

**The astute basketball fan will notice my use of Warriors colors.