levon wrote:I think pre-analytics NBA thought is really interesting and is undervalued by blog boys all the time. Being dismissive and smart-alecy about the mental aspects of the game when you've never played NBA ball isn't more empirical; it's actually less responsible because you're actively dropping qualitative information that can either be quantified, be used to interpret the numbers, or both.
Kobe's approach to the game was deception. He would start off by imposing his will and getting the defense to react a certain way, and then do the opposite to throw them off. He would do this in something as small scale as a single move, or across possessions, or across games within a playoff series. He would be very informed by film and scouting reports, focusing on tendencies and less on local optima.
It seems to me the dominant school of thought now is to do the most locally optimal thing per possession. It's definitely less of a mental/emotional calculation and more of just referring to data and basically executing a simple min-max. But I think winning is largely a separate experiential skillset, based on very minute details. That's something Kobe loves talking about.
A "locally optimal" strategy is probably near-optimal for basketball, since temporarily ignoring things like fatigue, it isn't clear that following some other strategy actually improves long-term reward at the cost of short-term reward.
For certain games like chess, you definitely cannot be short-sighted otherwise you'll win the battle but lose the war.
But for basketball, you want to try to win each possession (again, temporarily ignoring fatigue).