APBRmetrics

If you look at the top 100 3FGA games by individuals -- http://bkref.com/tiny/mMCJ8 -- sort by 3fg% and find the median is .438

It's possible that just by shooting more of them, their accuracy improved. Perhaps more likely: because they were hitting them, they shot more of them.
Teams with these prolific ( >14 3fga ) long-range shooters won 47 of the 100 games.

Page 2 is all guys with 14 3fga. The median % is .429, and they won 49 of 100.

schtevie wrote:

Crow wrote:I am supportive of thinking about a high frequency 3 pt attack (based on mid-range shot reduction) but I'd want to research it deeper and see more positive team results rather than just adopt it based on the logical consideration of simple comparative 2pt and 3 pt shot success / efficiency alone.

Crow, I don't want to misrepresent your views, but I take this comment as questioning the notion that the 3PA margin is a profitable one.

Alright schtevie, I guess your read of my comment and my intention are not exactly the same. I did say that the 2pt / 3 pt comparison had a logical basis so to me that is not questioning the notion that the 3PA margin is a profitable one. What I was trying to think about more carefully has the strategy of a high frequency 3 pt attack, a related but different question to me. I have been an advocate of a high frequency 3 pt strategy especially if accompanied by a high frequency of FTAs, ORs and good defense. I was just trying to dig deeper and see if my logic based advocacy of a high frequency 3 pt attack was well supported by game data or less well supported. I am not ready to close the book yet but it seems less well supported than I expected / hoped.

Analyzing a bigger data set of high frequency 3 pt attempt games (than the 48 40+ 3pt FGA games you looked at) would seem worthwhile to me to calculate how much that boxscore element contributed above league average expected shot returns and how much that contributed in win shares comparable to other boxscore elements to the actual team wins and losses of those games. Of course a high frequency 3 pt attack also affects how open other shots are, FTAs, OR%, etc. so there may be additional indirect impact / value positive and negative from the high frequency 3 pt attack. In addition to this level of analysis (perhaps using the 30 attempt cutoff, which is still 50% than the norm), it might be ideally desirable to look for a regression-based statistical plus minus value for 3 pt attempts (and I guess 3 pt makes too) just in games where the attempts exceeded 30 to isolate the value of 3 pt shots under a high frequency use strategy to find that "adjusted" value distinct from the general value of 3 pt attempts and makes in all games calculated using average expected returns as stated above. I am unlikely to do to either, especially right away; but if any one else wanted to tackle it, that could be helpful and would be appreciated.

I looked at 30+ 3pt FG attempt regular season games for several periods: thru 1993-94, '94-04, '05-09, '10-14 and 2012-14. The frequency of such games kept increasing but the average number of attempts in such games stayed quite steady at 33 and the FG% stayed just short of 38%. We will probably continue to see more 30+ 3pt attempt games. It will take a change in behavior to increase the average 3 pt FG attempts in such games. It might not be possible to raise the 3 pt FG%.

The current league average eFG% is just short of 50%. In a 33 attempt game at 38% FG%, the 3 pt attack will yield points more than the league average 2 / 3 point shot distribution assuming the 2pt FG% stays the same (it may not). There would be 33 3pt attempts - the average of 21 = 12 extra 3 pt attempts each yielding an eFG% of 57% or approximately 7% above the average yield = 0.84 * 2 = 1.68. It probably should be a little more since 3 pt shots are yield more than the overall average but I'll leave out calculation out to simplify a step. Then the 2% FG% edge on the other 21 3 pt shots (0.84 pts) and you get about + 3.5 - 4 pts from using this strategy and getting average results.

Not drawing a conclusion, just reporting:

Vipers 3 pt FG% for January games is below 35%, down from 37% in December. Team is 4-3.

Averaging 46 attempts per game for season. Almost 5 more in losses than wins. More at home than road.