Optimizing End Of Quarter Shot-Timing In The NBA

[jessefis]

Here is research paper I just finalized that I'd like to share:
Optimizing End Of Quarter Shot-Timing In The NBA: Everyone Knows About The 2 For 1, But What About The 3 For 2?
http://tothemean.com/2015/02/15/optimiz ... iming.html

This research is what I was referring to in this other thread:
http://www.apbr.org/metrics/viewtopic.php?f=2&t=8738

[Mike G]

It looks pretty comprehensively done, and it may well be the authoritative analysis for many years to come.
Having said that, I didn't study it closely all the way through. The conclusion is about what I anticipated:

... even when properly accounting for the reduced value of a “timed” possession, there still exists opportunities for implementing strategic timing. Interestingly, these optimal timing strategies revealed benefits outside of the traditional “2-for-1” time period, which suggests that there is a legitimate “3-for-2” strategy opportunity.
A strategy like this will only be effective in real-time situations if it is simplified so that players can execute the strategy with little instruction or deep thought...

Kinda hard to cut and paste from the article, but here is the 'cheat sheet' accompanying:

Code: Select all

Poss. start   Simple     Optimal   Pt. Advantage 
time (sec)   Strategy    End Time  Over "Normal"
30 to 40     Push pace   25 to 29     0.29
47 to 52     Slow pace   25 to 29     0.11
58 to 66     Push pace   51 to 56     0.06

Relative to gaining 2 points in a 2-for-1 strategy (top line), your odds are 1/0.29/2, or about 1 in 7.
The middle line seems to be a strategy to deny your opponent a 2:1. Consistent with the previous calculation, that works once in 18 tries.
The true 3:2 strategy (bottom line) pays off one time in 33.
These odds, I think, account for the occasions when the strategies 'backfire'.

Watching the game clock AND the shot clock, while trying to get a good shot at 25-29 sec, you might have to have code words for what's going on. Obviously, getting the ball with 52 sec. left, it may be super-optimal to shoot at 29 sec. (1 on the shot clock); but in a game, you will take a good shot before that.

[Crow]

There is also this poster / paper for end of game. Could be generalized to end of any quarter. http://www.sloansportsconference.com/?p=16381