Simple question regarding True Shot Attempts

[BasketDork]

I know the formula for TS % is = PTS÷(2*(FGA+.44*FTA)), so by definition, are TS Attempts = (2*(FGA+.44*FTA))or simply FGA+.44*FTA ? Dumb question I know, but I've seen conflicting equations, which both get you your TS%, but if I'm only interested in True Shot ATTEMPTS, which is the correct answer ? Thanks guys!

[Crow]

FGA+.44*FTA.

[bbstats]

Good question though! I frequently look at Pts/Shot Attempt rather than TS% as it can be understood in a Pts/Possession context.

[Mike G]

DeAndre Jordan last night had this shooting line:

Code: Select all

Pts  FGA  FTA   TSA    TS%
26    7   34   21.96  .592 

Did his 34 FTA use just about 15 possessions? Or did they use 17?

If it was 17 (i.e., no and-1's in there), his "true" TS% was .542 -- rather a large difference.

[The Clippers shot .600, so he was/wasn't a liability on offense.]

[mystic]

I know the formula for TS % is = PTS÷(2*(FGA+.44*FTA)), so by definition, are TS Attempts = (2*(FGA+.44*FTA))or simply FGA+.44*FTA ?

FGA + 0.44 FTA = TSA

That should be pretty obvious, because the 2 in there is just based on the fact that a typical field goal gives a team 2 points. And the formula does not compare the made shots with the total amount of shots, but with the made points. That gets corrected by using 2 in the divisor to get the TS%.

DeAndre Jordan last night had this shooting line:

Code: Select all

Pts  FGA  FTA   TSA    TS%
26    7   34   21.96  .592 

Did his 34 FTA use just about 15 possessions? Or did they use 17?

If it was 17 (i.e., no and-1's in there), his "true" TS% was .542 -- rather a large difference.

[The Clippers shot .600, so he was/wasn't a liability on offense.]

0.44 as a coefficient is supposed to work in average, not for every single player in every single game.

[Mike G]

I get that.
Once upon a time, perhaps in the first month of this forum's existence, we debated the proper term for this.
I didn't much like that it was gonna go down as TS%, when it isn't really a "true" shooting%, but rather an estimate. Play by play would come along someday, and then we'd know true shooting%.

Meanwhile, in a given game it's virtually never any player's "true" %. And on occasion, it can be off by at least .050

[mystic]

Well, maybe eTS% would be a better abbreviation, where e = estimated? But e also stands for effective in eFG% ... I understand where you are coming from, but I never seen the "true" as if it means perfect measurement anyway and always treated it as an approximation of the true shooting attempts/percentage.

[NateTG]

Well, maybe eTS% would be a better abbreviation, where e = estimated? But e also stands for effective in eFG% ... I never seen the "true" as if it means perfect measurement anyway and always treated it as an approximation of the true shooting attempts/percentage.

'Effective' or 'equivalent' could also work in that context, and there's a whole slew of pretense in the naming conventions that implicitly assume that some approximation or estimate is accurate.

[BasketDork]

I understand that the '2' is reflecting the roughly estimated shot value and that the .44 is a rough estimate of 44% of all FTA's being of the two-shot variety, but being that FG% is FG/FGA, FT% is FT/FTA, 3ptFG% is 3PM/3PA, shoot, batting average is H/AB, SLG% is TB/AB, Completion % is CMP/ATT, numerator/denominator, numerator/denominator, numerator/denominator, so it isn't that dumb of a question, being that TS% is PTS/(2*(FGA + .44 * FTA)), inferring that the entire denominator is the "attempts" portion of the formula, I'm sure, is a common trip-up.

[Statman]

I understand that the '2' is reflecting the roughly estimated shot value and that the .44 is a rough estimate of 44% of all FTA's being of the two-shot variety, but being that FG% is FG/FGA, FT% is FT/FTA, 3ptFG% is 3PM/3PA, shoot, batting average is H/AB, SLG% is TB/AB, Completion % is CMP/ATT, numerator/denominator, numerator/denominator, numerator/denominator....

I use pts/(2*FGA+FTA) for that very reason. Simple, like the other known sport rates. Might undervalue FTs slightly - don't care, it's all an estimate anyway.

Slugging% or OBP% would be MUCH more complicated if we tried to incorporate relative value of each outcome. A home run would have about 5 times the value of a single (instead of 4 in Slug%), over 3 times the value of a double, over twice the value of a triple.

A walk would have a little more than half the value of a single.

But - we don't muck it all up in baseball - we keep it simple for the masses.

In basketball - that .44 as the FT factor should actually be different depending on the league and league average efficiencies as well as foul bonus rules. Less shots in bonus, higher factor. Lower league effiencies from the field relative to free throw rate - lower factor (getting to the line more relative value).

Since it's just general relative information like Slug & OBP - I say have general TS% be simple also. It still gives the type of quality info one needs (combining well 2ptFG, 3pt, & FT rates & %'s)- even if it might very slightly from the more "exact" .44 FT factor estimate.

[Mike G]

I understand that the '2' is reflecting the roughly estimated shot value and that the .44 is a rough estimate of 44% of all FTA's being of the two-shot variety,....

I use pts/(2*FGA+FTA) for that very reason. Simple, like the other known sport rates. Might undervalue FTs slightly - don't care, it's all an estimate anyway..

Same here. I call it "effective shooting%", or Eff%
Missed FT are rebounded by offense less often than missed FG, so it may be (partly) justifiable to count them all in the denominator.

It's estimated that 88% of all FTA are in 2-shot situations. The simplest TS% formula is :
Pts/(2*FGA + .88*FTA)
... unless of course you have already gotten TSA; then it's Pts/TSA/2

The '2' isn't an estimate -- it's the value of a 2pt possession. This convention puts TS% on the same scale as eFG% and FT%.

[EvanZ]

You guys are talking about *estimates* of TSA.

First, shouldn't we define what an *actual* TSA is?

For my money:

Code: Select all

TSA = FGA + (# of shooting fouls received) + 0.5*(# 3pt shooting fouls received)

An estimate of this is TSA = FGA + 0.44*FTA. On my site, I have the actual TSA (although I realize in writing this post that I need to add that last term to correct for 3pt shooting fouls).

[NateTG]

...

First, shouldn't we define what an *actual* TSA is?

Code: Select all

TSA = FGA + (# of shooting fouls received) + 0.5*(# 3pt shooting fouls received)

...

That formula -as written- seems strange to me. Why is getting fouled on a 3-point attempt considered to be 1.5 shooting attempts?

If we go by the usual definition of the "TS%" statistic, then - effectively by definition - a "TS Atttempt" is either 9/4 of a free throw, or a field goal attempt.

You can make the justifiable case that "TS%" is an approximation for how efficiently a player converts possessions into points by shooting, but there are many different sensible ways to make that approximation depending, for example, on what data is available and how the approximation will be used.

[EvanZ]

...

First, shouldn't we define what an *actual* TSA is?

Code: Select all

TSA = FGA + (# of shooting fouls received) + 0.5*(# 3pt shooting fouls received)

...

That formula -as written- seems strange to me. Why is getting fouled on a 3-point attempt considered to be 1.5 shooting attempts?

Ha. You're right. I completely agree with you. Good news is the formula on nbawowy is now correct again. (I must have realized this a long time ago, and forgot how I had reasoned it.) Should just be:

Code: Select all

TSA = FGA + (# of shooting fouls received) 

[Nate]

Ha. You're right. I completely agree with you. Good news is the formula on nbawowy is now correct again. (I must have realized this a long time ago, and forgot how I had reasoned it.) Should just be:

Do you also subtract points from technical and 'off the ball' free throws from players' totals?