RealGM

why is it in the equation at all?

lets say you want players "actual" ts% why not use:

(fgm + ftm)/(fga + fta)

it comes more or less to the same value without the arbitrary 0.44

top 5 scorers from lats year with both ts%

(ts% with 0.44)

melo 56.0
durant 64.7
kobe 57.0
lebron 64.0
harden 60.0

now ts% without 0.44

melo 54.7
durant 64.4
kobe 57.0
lebron 62.1
harden 59.0

islamovic wrote:why is it in the equation at all?

Because in that way you get the best approximation of the "true shooting attempts". The 0.44 is not arbitrary at all. It is derived from game data.

To expand, its based on the fact that FTs aren't always taken in pairs. From my understanding its an estimate of how often FTs end a possesion.

so lets say a guy has 10 fta n makes all 10

so his ts% = 10/(2 * (0.44*10)) = 1.14

thats 114%

how does that make sense?

with 0.5 instead of 0.44 his % would be 100%

so why 0.5 isnt used in ts%?

islamovic wrote:so lets say a guy has 10 fta n makes all 10

so his ts% = 10/(2 * (0.44*10)) = 1.14

thats 114%

how does that make sense?

Only a fool would use TS% in such a fashion. If you have just one game, count the amount of true shooting attempts instead of using that formula, makes more sense.

islamovic wrote:with 0.5 instead of 0.44 his % would be 100%

so why 0.5 isnt used in ts%?

TS% is an approximation for a big enough sample. Using 0.44 gives a MORE ACCURATE number for the true shooting attempts than 0.5. Seriously, that was answered twice already, why did you ask that again?

Free throws in the NBA are not just taken as pairs, you might have seen an And1 here and there, or someone taking 3 ft due to being fouled while attempting a 3pt shot. So, no, 0.5 is not a good coefficient, 0.44 is clearly better. If you don't believe that, go through the pbp and count that by yourself. ;)

I believe that NBA.com actually uses the true TS% from Pbp, not the approximation.

colts18 wrote:I believe that NBA.com actually uses the true TS% from Pbp, not the approximation.

I was wondering why the NBA.com figures differed a great deal from basketball reference. I am focusing on the offensive rating stat.



http://stats.nba.com/playerStats.html?P ... ed&PerMode

In 2000, the Lakers offense was +5.0 (106.3) with Shaq on the floor, and the defense was -6.0 (95.3). But according to NBA.com, the league average was 101.3, while basketball reference has it at 104.1. Are the basketball reference estimates to be taken with a grain of salt?

Dipper 13 wrote:

colts18 wrote:I believe that NBA.com actually uses the true TS% from Pbp, not the approximation.

I was wondering why the NBA.com figures differed a great deal from basketball reference. I am focusing on the offensive rating stat.



http://stats.nba.com/playerStats.html?P ... ed&PerMode

In 2000, the Lakers offense was +5.0 (106.3) with Shaq on the floor, and the defense was -6.0 (95.3). But according to NBA.com, the league average was 101.3, while basketball reference has it at 104.1. Are the basketball reference estimates to be taken with a grain of salt?

NBA.com and B-R use different pace estimations which is why they have different league averages.

colts18 wrote:

Dipper 13 wrote:

colts18 wrote:I believe that NBA.com actually uses the true TS% from Pbp, not the approximation.

I was wondering why the NBA.com figures differed a great deal from basketball reference. I am focusing on the offensive rating stat.



http://stats.nba.com/playerStats.html?P ... ed&PerMode

In 2000, the Lakers offense was +5.0 (106.3) with Shaq on the floor, and the defense was -6.0 (95.3). But according to NBA.com, the league average was 101.3, while basketball reference has it at 104.1. Are the basketball reference estimates to be taken with a grain of salt?

NBA.com and B-R use different pace estimations which is why they have different league averages.

How nba.com calculates pace?
Because sometimes differences are huge. For example 2013 finals game 1:

nba.com 87.7 pace, heat 102.9 ortg, 102.3 drtg
b-r 85 pace, heat 103.5 ortg, 108.2 drtg

(And b-r numbers are much closer to real - based on pbp - numbers)