Points per shot

[rrravenred]

How valuable do people feel points per shot is as a measure of offensive efficiency?

Just for fun I did some excel-ing... firstly, I put in some arbitrary values... 40 games played minimum, a minimum of 15 ppg, which yields (theoretically) the top 58 scorers in the game. Then I worked out rankings within those 58 players both in PPG and Points per shot.

The top two in PPG were KD and LBJ, both of which were reasonably efficient scorers at 1.48 and 1.5 points per shot. All well and good.

After that in PPG comes Kobe, Wade and Monta Ellis. Not forgetting that these rankings are out of 58, they ranked 28th (1.25 PPS), 21st (1.31) and 46th (1.18. ouch, Monta).

The runaway points per shot leader, if you're interested, is Dwight Howard. Despite his reputation for offensive ineptitude, he still earns 1.85 points every time the ball leaves his hand.

The question then becomes whether or not redistribution of the shots from some of these less efficient players are taking would actually result in more effective offensive outcomes for their teams. Any thoughts?

[sabi]

how are u accounting for freethrows?

[rrravenred]

All points. It doesn't refer to FTA, but the ability to convert from FTA opportunities is sort of inherent in there.

[Doctor MJ]

"points per shot" taken literally isn't nearly as good as TS%, which is essentially "points per actual shot". You have to factor in free throw attempts because otherwise a player who makes 2 free throws on a missed-but-fouled shot (2 points, no "shots", infinite points per "shot"), actually looks better than the guy who got an and-1 (3 points, 1 shot, 3 points per shot).

To be clear though, TS% isn't perfect because ideally you shouldn't be factoring in FTs at all (except when a player goes to the line on a non-shooting foul) but instead should just be counting every damn time a player shoot a non-free throw. Very frustrating that this isn't done. There are so many advanced stats that it's understandable that scorekeepers of the past didn't have the manpower to track - but this is just a decision that was made way back when that was wrong.

[AbdicatedReign]

I much prefer PC/PU.

[Paydro70]

"points per shot" taken literally isn't nearly as good as TS%, which is essentially "points per actual shot". You have to factor in free throw attempts because otherwise a player who makes 2 free throws on a missed-but-fouled shot (2 points, no "shots", infinite points per "shot"), actually looks better than the guy who got an and-1 (3 points, 1 shot, 3 points per shot).

To be clear though, TS% isn't perfect because ideally you shouldn't be factoring in FTs at all (except when a player goes to the line on a non-shooting foul) but instead should just be counting every damn time a player shoot a non-free throw. Very frustrating that this isn't done. There are so many advanced stats that it's understandable that scorekeepers of the past didn't have the manpower to track - but this is just a decision that was made way back when that was wrong.

Agreed on all points. TS% has to estimate trips to the line, which is not as accurate as it could and should be. Still the best thing out there at the moment.

[TheSecretWeapon]

Best measure of shooting efficiency from the field is effective field goal percentage, which accounts for the effect of the 3pt shot.

Best all around measure of shooting is TS%, which incorporates free throws. The .44 multiplier on free throws does a pretty good job, but doesn't hold for all players -- some get more and-one opportunities, others get fewer. As a league average, it's held up to several studies. Someone with good programming skills could likely use play-by-plays to calculate actual TS% -- using actual and-ones.

Another approach, used by at least one analyst I'm aware of, is to say 2 FTA = 1 FGA instead of using the .44 estimate. There's a logical consistency there.

Finally, the best all-around measure of individual offensive efficiency is Dean Oliver's offensive rating, which incorporates shooting from the field and from the FT line, assists, turnovers and offensive rebounds.

[tsherkin]

It wouldn't even take GOOD programming skills, just a lot of dirty work assembling data from play-by-plays.