APBRmetrics

[Mike G]

We still haven't come to a consensus on what to call this stat, or how it should be displayed.

B-R.com has PSA (pts per shot attempt), and BR calls it PPFGA. In both places, the formula is = Pts/(FGA + .44*FTA). This gives numbers like 1.04, just twice what we expect from a shooting %.

At least B-R places this column right next to the Adj% column (which is itself separated from the relevant FG% and 3P% columns, by the FT% -- implying, perhaps, that FT are part of the consideration.)

DanR calls it "true FG%", I think. In each case, I believe it's confusing to have the term "field goal" in a calculation that includes FT -- even from non-shooting fouls. On first glance, I'd think "true FG%" means good old FG/FGA.

The defense may have designated you as the shooter by fouling you before any FG attempt can occur. But you are still the shooter, at the FT line. Therefore, "shooting %" is not misleading.

In an effort to be accessible, I wish we could agree on an acceptable term.


Meanwhile, what about the factor .44 that's applied to FTA? With a significant increase in fouls and FTA this year, is that still the right number? Intuitively, I'd guess there are fewer FGA because of more fouls. And maybe therefore the % of and-1 FTA has decreased, relative to all FTA.

I don't remember how the number .44 was settled on. I've used .45, and I suppose the difference is miniscule.

Anyway, I'm going to check out the "Add a Poll" feature...

[Dan Rosenbaum]

Well, I have been dividing this by 2 and calling it true field goal percentage.

True field goal percentage = (PTS/2)/(FGA+0.44*FTA)

I think that I adopted that naming convention from Kevin Pelton. At times I have called this effective field goal percentage (which I like better), but effective field goal percentage seems to have become established as the following.

Effective field goal percentage = [(PTS-FTM)/2]/FGA

Not divided by two, I am not sure what to call it. Points per shot seems established as PTS/FGA, which I think is a very hard statistic to interpret.

Divided by two, I have a very slight preference for leaving the "field goal percentage" in the name, despite the fact that some of the points are generated from non-shooting fouls. I like placing this statistic in direct competition with field goal percentage and calling it true field goal percentage or effective field goal percentage does exactly that.

I am still thinking about how to vote, because I am not sure what I would want to call this number when it is not divided by 2.

[Dan Rosenbaum]

And the 0.44 multiplier does not necessarily change as more fouls are called. It seems to be the multiplier that best equates the number of offensive and defensive possessions, so that is how I came up with it. I think others have arrived at the number in similar fashions.

The multiplier would be exactly 0.5 if there were no And-1s, technicals, flagrant fouls, clear-path fouls, or three shots after a foul attempt on a three point attempt. As the fraction of those things rise relative to regular two-shot free throw attempts, the multiplier should fall. I suspect the fraction of those things has risen in the past couple of seasons, but I am not sure.

[gabefarkas]

I think PPSA should just be Pts/FGA.

Of all the choices, I like Effective Shooting Percent the best. However, I would suggest "Effective Scoring Percent" because it shows how effective a player is at scoring, and it's presented as a percent.

[HoopStudies]

[Kevin Pelton]

The problem with points per shot attempt is that people who don't follow advanced statistics (and I'm going to go ahead and assume that's a pretty dramatic majority) think of it as meaning, as Gabe says, literally dividing by field goals attempted, which happens to be an even crappier method of divining efficiency than just looking at field goal percentage.

I think using the word true in there has a nice pyschological effect. The existence of slugging "percentage" in baseball leaves me unconcerned about going over 100 "percent". So put me down for either true shooting percentage (which I've always used) or true scoring percentage. I don't really care about the middle word.

[Kevin Pelton]

Let me correct myself. When I say I've "always" used it, that's a lie. I used about a dozen different names in the early going. Then I held this poll with myself and, to quote David Stern, "The vote was 1-0 in my favor."

[Ed Küpfer]

[kjb]

I thought Kevin's term was "true shooting %" which is a name I liked. A stat guy in another forum asked me about it and got the name wrong, and he runs around touting "true scoring %".

Without the 2 factor, I've called it pts per scoring opportunity or PSO.

My concern with using "effective" in this name is that it could be easily confused with efg%, which is a different measure. "fg" should not be part of the name since it includes free throws.

[KnickerBlogger]

I would have voted for True Shooting % (TSP), but since that wasn't an option I went with PSA. Admittedly I did so since that what it's called in Forecast, but the name PSA is a bit deceiving, since you would think that it's PTS/FGA or PTS/(FGA+FTA).

[Dan Rosenbaum]

Sorry KevinP for not noticing that you called it true shooting percentage and not true field goal percentage.

With the 2 factor, I vote for true shooting percentage (TSP).

And without the 2 factor, I like WizardsKev's suggestion of points per scoring opportunity (PSO).

[HoopStudies]

[Ben]

BTW, .45 or .44 is an average. Has Roland or anybody looked at the 82games data to see if there's much variation?

[Kevin Pelton]

So now we just need John Hollinger to reprint all his books.

Where is John anyways? Why hasn't he come over here?

[HoopStudies]