Player of the Game #9 (Nuggets @ Bullets)

[payitforward]

I missed the entire game and am just looking at the box score right now. Bertans had 35 points on 14 shots! That's a TS% of 125%. That's gotta be some type of record.

14 shots? Make that 11 shots. Plus 8 FTAs.

But, you are very close on the TS% -- 120.5%

No. I'm exactly right.

He shot 8 free throws. He took 3 shots to generate those 8 free throws (because two of his trips to the line were while being fouled shooting a 3). So it took him 14 total shot attempts to generate the 35 points.

Now, the official TS% calculation assumes .44 shot attempts were necessary to generate a FT attempt, because that's the average for all players. But in this particular case, we know the exact number of shots necessary to generate those free throws. How about we call it the true true shooting percentage?

But, we always know how many shots generated the free throws.

Whenever you use the formula, you put both ftas & fgas in as hard data, so I don't believe the formula relies on the calculated average you suggest.

Ok I just checked at captain calculator -- https://captaincalculator.com/sports/basketball/true-shooting-percentage-calculator/.

20 points on 10 FGAs/0 FTAs = a 1.0 TS%. If instead of calculating on FGAs exclusively, you calculate on FTAs exclusively, then entering 0 FGAs/22.727 FTAs (which is 10 divided by .44) again yields a 1.0 TS%.

It's invariant: if you keep points constant then no matter what figures you work with, you can demonstrate this. I.e. subtract 1 FGA & add 2.272 FTs, & the calculated TS% remains the same. True for any number of points.

[payitforward]

nate -- just looking through that post again, I see that you are doing some clever thinking, but that it isn't actually how TS% works. Keep in mind that TS% doesn't stand for "True Shooting %" -- it stands for "True Scoring %."

Imagine 2 guys each take 10 shots & produce 12 points. Whether one of them went 6 of 10, all 2-pointers, & the other went 4-10, all 3-pointers, doesn't affect their TS%. Ditto if a 3d guy went 5-10, making 3 2-pointers & 2 3-pointers. Same TS%.

If on one of those FGAs, he's fouled -- but the shot goes in anyway -- he'll get a FTA. No matter whether the shot that took him to the line was a 2 or a 3 will make no difference at all.

If he's fouled on one of those shot attempts, but the shot does not go in, then he'll go to the line for either 2 or 3 FTs. Again, what he proceeds to do at the line directly affects his TS%. You don't have to make any adjustments based on the kind of shot it was. In fact, that shot attempt isn't counted. A player who's fouled while shooting & misses the shot is not charged a FGA.

Similarly, if the opposing team is in the penalty, the same guy may go to the line independent of taking a shot.

[nate33]

14 shots? Make that 11 shots. Plus 8 FTAs.

But, you are very close on the TS% -- 120.5%

No. I'm exactly right.

He shot 8 free throws. He took 3 shots to generate those 8 free throws (because two of his trips to the line were while being fouled shooting a 3). So it took him 14 total shot attempts to generate the 35 points.

Now, the official TS% calculation assumes .44 shot attempts were necessary to generate a FT attempt, because that's the average for all players. But in this particular case, we know the exact number of shots necessary to generate those free throws. How about we call it the true true shooting percentage?

But, we always know how many shots generated the free throws.

Whenever you use the formula, you put both ftas & fgas in as hard data, so I don't believe the formula relies on the calculated average you suggest.

Ok I just checked at captain calculator -- https://captaincalculator.com/sports/basketball/true-shooting-percentage-calculator/.

20 points on 10 FGAs/0 FTAs = a 1.0 TS%. If instead of calculating on FGAs exclusively, you calculate on FTAs exclusively, then entering 0 FGAs/22.727 FTAs (which is 10 divided by .44) again yields a 1.0 TS%.

It's invariant: if you keep points constant then no matter what figures you work with, you can demonstrate this. I.e. subtract 1 FGA & add 2.272 FTs, & the calculated TS% remains the same. True for any number of points.

We don't always know how many shots generated the free throws, not when looking at a entire season of data. Basketball-reference doesn't check exactly how many FTA's were generated from 3PA's, 2PA's, And-1's and technicals. It just knows that it takes approximately .44 FGAs attempts to generate a FTA. But that's a league wide average.

An individual player might go 4/8 from the field (all 2's) and also get fouled 3 times on FGA's and go 6/6 from the FT line. TS% would say that he scored 14 points on 10.64 attempts (8 + (6 x.44)), which is a TS% of .657. But that's because it didn't follow the game log. The reality is that the player actually took 11 shot attempts to generate those 14 points and therefore had a true TS% of .636.

Another individual play might go 4/8 from the field (all 2's) and also get fouled twice on 3PA's and go 6/6 from the FT line. TS% would say he had a TS% of .657. The reality is, he generated 14 points on 10 actual shot attempts and has a true TS% of .700.

Bertans had a true TS% of 1.25 the last game. The calculated TS% actually penalized him because it assumed it took Bertans 3.52 shot attempts to generate his 8 FTA's when in fact it only took him 3 shot attempts.

[payitforward]

Actually... there is something weird about TS% at this high level. Suppose you leave out Davis's FTs altogether; he went 100% on 8 of them.

So, now it's 27 points on 11 shots -- that's a 1.227 TS%. Yet, if you add in the 8 FTAs & 8 points from making all of them... his TS% goes down!

On FTs, you have a max TS% of 100%.

[tontoz]

If there were no 3s, and1s and technicals they could just count each free throw as .5 possessions. If a guy took 6 fts they would know it took 3 possessions.

They had to make an adjustment (because of 3s, and 1s and technicals) to estimate the number of possessions. The problem is that they did this a long time ago when guys weren't taking as many 3s.

If they studied possessions today I am sure the current .44 adjustment would go down.

In this case scoring 8 points in 3 possessions on free throws gives a TS of 1.33 on those possessions. That is the correct number but the problem is that when comparing Bertans game to other games they have to use the .44 because like Nate said they don't track where the fts came from.

[payitforward]

So... you make a FT -- & your TS% goes down! Weird....

[BrianInPhilly]

Appreciate the above explanations

[payitforward]

No. I'm exactly right.

He shot 8 free throws. He took 3 shots to generate those 8 free throws (because two of his trips to the line were while being fouled shooting a 3). So it took him 14 total shot attempts to generate the 35 points.

Now, the official TS% calculation assumes .44 shot attempts were necessary to generate a FT attempt, because that's the average for all players. But in this particular case, we know the exact number of shots necessary to generate those free throws. How about we call it the true true shooting percentage?

But, we always know how many shots generated the free throws.

Whenever you use the formula, you put both ftas & fgas in as hard data, so I don't believe the formula relies on the calculated average you suggest.

Ok I just checked at captain calculator -- https://captaincalculator.com/sports/basketball/true-shooting-percentage-calculator/.

20 points on 10 FGAs/0 FTAs = a 1.0 TS%. If instead of calculating on FGAs exclusively, you calculate on FTAs exclusively, then entering 0 FGAs/22.727 FTAs (which is 10 divided by .44) again yields a 1.0 TS%.

It's invariant: if you keep points constant then no matter what figures you work with, you can demonstrate this. I.e. subtract 1 FGA & add 2.272 FTs, & the calculated TS% remains the same. True for any number of points.

We don't always know how many shots generated the free throws, not when looking at a entire season of data. Basketball-reference doesn't check exactly how many FTA's were generated from 3PA's, 2PA's, And-1's and technicals. It just knows that it takes approximately .44 FGAs attempts to generate a FTA. But that's a league wide average.

An individual player might go 4/8 from the field (all 2's) and also get fouled 3 times on FGA's and go 6/6 from the FT line. TS% would say that he scored 14 points on 10.64 attempts (8 + (6 x.44)), which is a TS% of .657. But that's because it didn't follow the game log. The reality is that the player actually took 11 shot attempts to generate those 14 points and therefore had a true TS% of .636.

Another individual play might go 4/8 from the field (all 2's) and also get fouled twice on 3PA's and go 6/6 from the FT line. TS% would say he had a TS% of .657. The reality is, he generated 14 points on 10 actual shot attempts and has a true TS% of .700.

Bertans had a true TS% of 1.25 the last game. The calculated TS% actually penalized him because it assumed it took Bertans 3.52 shot attempts to generate his 8 FTA's when in fact it only took him 3 shot attempts.

Suppose a guy takes no FGAs whatever. But, with the opponent in the penalty, he is fouled 3 times & goes 6/6 from the line. His TS% is 1.136.

Explain.

[nate33]

But, we always know how many shots generated the free throws.

Whenever you use the formula, you put both ftas & fgas in as hard data, so I don't believe the formula relies on the calculated average you suggest.

Ok I just checked at captain calculator -- https://captaincalculator.com/sports/basketball/true-shooting-percentage-calculator/.

20 points on 10 FGAs/0 FTAs = a 1.0 TS%. If instead of calculating on FGAs exclusively, you calculate on FTAs exclusively, then entering 0 FGAs/22.727 FTAs (which is 10 divided by .44) again yields a 1.0 TS%.

It's invariant: if you keep points constant then no matter what figures you work with, you can demonstrate this. I.e. subtract 1 FGA & add 2.272 FTs, & the calculated TS% remains the same. True for any number of points.

We don't always know how many shots generated the free throws, not when looking at a entire season of data. Basketball-reference doesn't check exactly how many FTA's were generated from 3PA's, 2PA's, And-1's and technicals. It just knows that it takes approximately .44 FGAs attempts to generate a FTA. But that's a league wide average.

An individual player might go 4/8 from the field (all 2's) and also get fouled 3 times on FGA's and go 6/6 from the FT line. TS% would say that he scored 14 points on 10.64 attempts (8 + (6 x.44)), which is a TS% of .657. But that's because it didn't follow the game log. The reality is that the player actually took 11 shot attempts to generate those 14 points and therefore had a true TS% of .636.

Another individual play might go 4/8 from the field (all 2's) and also get fouled twice on 3PA's and go 6/6 from the FT line. TS% would say he had a TS% of .657. The reality is, he generated 14 points on 10 actual shot attempts and has a true TS% of .700.

Bertans had a true TS% of 1.25 the last game. The calculated TS% actually penalized him because it assumed it took Bertans 3.52 shot attempts to generate his 8 FTA's when in fact it only took him 3 shot attempts.

Suppose a guy takes no FGAs whatever. But, with the opponent in the penalty, he is fouled 3 times & goes 6/6 from the line. His TS% is 1.136.

Explain.

His free throws still take up the possession for his team, so they essentially "cost" his team a FGA. So he scored 6 points on 3 FGA's.

[payitforward]

Aaah. You are simply using "FGA" to mean the same thing as "possession." Now, some possessions end without attempts to score, but... so what?

Except that 6 points on 6 FTAs produces a higher TS% than 6 points on 3 FGAs.

Yet... it's possible, nate, that we have squeezed out every bit of the juice that this orange contains!

Wait! I may have found more juice!

When I calculate Davis's TS% by entering his 35 points, 11 FGAs & 8 FTAs, the algorithm that does the calculating doesn't know he went 9-11 on the FGAs & 8-8 on the FTAs. E.g. he might have gone 10-11 on the shots & 5-8 on FTAs. Or 11-11 on the FGAs (1 of them having been a 2-pointer) & 4-8 from the line.

Does this matter to your point?