Curry's shot selection has gone somewhat unnoticed but it's actually quite amazing to see how he approaches the game these days.
He's basically the analytics' dream with regard to what kind of shots he creates. Let's have a look at bbref's shooting numbers:
% of FGA by distance
0-3ft - 22.6
3-10ft - 8.6
10-16ft - 5.0
10ft<3 - 8.8
3P - 54.9
FTr - 0.269
It's amazing to see for two reasons. One, and this is probably the most impressive aspect, is his individual development over the course of his career. The other is how he looks compared to the rest of the league / other perimeter players in the league.
Let me focus in the first part. By now, he basically completely abolished the midrange-/longrange-game for himself. Isn't it astounding to see that a great shooter like Curry only takes 22.4% of his FGA from outside 3ft and inside the 3pt-line? From 10ft to 3P, i.e. the least efficient shots in theory, he only takes 13.8% of his FGA. Note that this isn't even taking into account the fouls he draws to get to the line, which mostly occurs around the rim. It's not that he can't make these shots, as he shoots around 47% from 10ft to 3P and slightly less if we also take the 3-10ft range (still scores at 44% from that area this season) into account; it's because these shots are clearly less efficient than those around the rim (69%) or from long distance (likewise 69% eFG).
Compared to his previous years, it's a) amazing to see when we look at his whole career but b) also nothing more than the next step of an already established path in recent years. In his first five years in the league (2010-2014), about 27.5% (+18.7) of his shots came from 16ft to 3P, about 9% (+4%) of his shots came from 10-16ft and about 10% (+1.4%) came from 3-10ft. His FTr was approximately around 0.2 (-0.069, which is more significant than it looks as it's, in other words, -6.9%). We can see some kind of development starting back in 2013/14, before Kerr arrived, but 2014/15 constitutes the most significant change in Curry's shot selection as this was the time where he clearly started to not really look for shots from 16ft to 3P any longer. His % of FGA from this are decreased from 24.5 in 2014 to 16.3 in 2015 to 8.8 in 2016.
What's also noteworthy, to me: he might not be finished yet. While 2014/15 was some sort of a turning point, we can still see further development in the direction I illustrated above. He increased the shots at the rim by 4% to 22.6% overall. It's amazing when you realize that the number - 22.6% - is basically twice as high as it has been his two early prime years (2012/13 and 2013/14). I already mentioned the drop-off in volume from 16ft to 3P. His relative FGA from 3ft to 16ft also decreased, while his 3PAr clearly increased from 0.482 (2015) to 0.549 (2016). The same is true for his FTr, although far less significantly (0.251 to 0.269). He's seems to be on a mission recently: picking the spots wisely to maximize efficiency.
What does this tell us? Well, it tells us two things. First, the offense has become far more sophisticated under Kerr. Superior coaching and development of the players/team allows Golden State to run an ultra efficient offense, which includes extensive minutes for Curry to move away from the ball in order to create good looks for himself and others in collaboration with the high post playmaker (usually Green or Bogut). But theory has to be converted into actual results and this leads us to the second part. Curry has become stupidly good at creating shots in the most efficient areas on the court and he has also become extremely wise as far as picking his spots is concerned. How did he become so good at creating these kind of highly efficient shots constantly? Well, of course the main answer lies in his shooting ability. He can not only shoot 3's at will from previosuly unheard of distances and regardless of coverage, which explains his 3pt-shooting, but he learned to leverage his shooting-threat to get to the basket more frequently and under less pressure on the first feet after blowing by his defender. Combine this with elite handles, a deceptively quick first step and a creative and impeccable finishing ability and you have an elite 3pt-shooter who's as or even more efficient around the rim - at an not-so-insignificant volume - as he's from behind the arc.
Hence, Curry's improved efficiency to unprecedented heights for volume scorers isn't sorcery. It's rather simple to explain: he decided to take the efficient shots more often and the less efficient shots less often. What is indeed amazing is the way he gets it done, since there's much more to it than merely realizing where the efficient spots happen to be. To shoot at 70% TS on shots entire game plans are supposed to take away at high costs, on incredibly high volume, and to shoot >70% TS around the rim, i.e. shots every game plan - independent of the opponent - tries to take away, on decent volume, is mind-blowing. If you're trying to understand efficient basketball, you have to look no further than Curry. Sure, nobody else has been able to do or replicate what current Curry is doing but as an individual, Curry has been maximizing his efficiency like no other player in history has ever been capable of.
