Measuring the D...

[Axel]

Let's debate the same numbers at least.

Actually, debating the value of the different numbers is the real intent of this thread. Debating Ross vs DD subbing should go to DD's thread or Ross's. Take it to Ross's, it needs the page count.

[Barolt]

Actually, debating the value of the different numbers is the real intent of this thread. Debating Ross vs DD subbing should go to DD's thread or Ross's. Take it to Ross's, it needs the page count.

I meant more that nba.com and Basketballreference.com calculate Drtg and Ortg differently, and if one person is pulling from each site, the entire discussion becomes somewhat useless.

[JWash]

On that note. NBA.com uses a more simplified formula for possessions than BasketballReference, and they also calculate their numbers using team pace not game pace, so basketballreference's numbers are more accurate with respect to oRTG/dRTG which is why you see them used more often here for those numbers. NBA.com is better for things like play type stats and such because basketball reference doesn't really track them.

[Axel]

I meant more that nba.com and Basketballreference.com calculate Drtg and Ortg differently, and if one person is pulling from each site, the entire discussion becomes somewhat useless.

And I mean that debating the value of one calculation over the other would be far more relevant to the topic that what you are trying to talk about.

[Barolt]

On that note. NBA.com uses a more simplified formula for possessions than BasketballReference, and they also calculate their numbers using team pace not game pace, so basketballreference's numbers are more accurate with respect to oRTG/dRTG which is why you see them used more often here for those numbers. NBA.com is better for things like play type stats and such because basketball reference doesn't really track them.

Interesting. But if the pace changes with different lineups for a team, as it sometimes can, wouldn't nba.com's numbers be more accurate if, say, a team played faster when said player is on the court?

[JWash]

Interesting. But if the pace changes with different lineups for a team, as it sometimes can, wouldn't nba.com's numbers be more accurate if, say, a team played faster when said player is on the court?

I'm not a stats guru but logically the way basketballreference calculates oRTG/dRTG makes more sense. It doesn't make sense to calculate them based only on your team's possessions because both teams are contributing to the flow and pace of the game not just one.

[JWash]

Axel, shit's back on topic now lol

[Barolt]

I'm not a stats guru but logically the way basketballreference calculates oRTG/dRTG makes more sense. It doesn't make sense to calculate them based only on your team's possessions because both teams are contributing to the flow and pace of the game not just one.

http://nyloncalculus.com/2015/12/21/...1-possessions/

Seems to indicate there's value to both ways of calculating it, because individual pace is far more accurate than game pace for individual statistics, but basketball-reference accounts for rebounds better than nba.com. So no truly right answer.

[DanH]

On that note. NBA.com uses a more simplified formula for possessions than BasketballReference, and they also calculate their numbers using team pace not game pace, so basketballreference's numbers are more accurate with respect to oRTG/dRTG which is why you see them used more often here for those numbers. NBA.com is better for things like play type stats and such because basketball reference doesn't really track them.

Accuracy has nothing to do with the differences in the numbers, depending on what you are looking at. The two sites, in their most commonly presented forms of ORTG and DRTG, are measuring entirely different things.

Basketball-reference's ORTG and DRTG are individual numbers (individual points scored per individual possession used). NBA.com's are on-court numbers (how the team performs with the player on the floor).

You can find the on-court numbers on basketball-reference as well but it is buried in a separate page (on-off data) than most of their team and player statistics. There you see the differences in the way the stats are calculated (for example, by BR, DD's on-off split for DRTG is 106 on, 97.3 off, while by NBA.com it is 102.8 on and 95.6 off).

They are both actually estimates, just different ones. Basketball reference's is actually a little less accurate (at the player level), as BR uses the generic pace numbers while NBA.com is the one that uses the correct on-court pace for each player. But BR does factor in offensive team rebounds better, so at a team level their ratings are a little more accurate (NBA.com tends to overestimate pace/possessions used - and therefore underestimate both ORTG and DRTG - by about 4%.

[tDotted]

Thread derailed in under 2 hours. Nice work team.

Just gotta keep moving posts!

[JWash]

Accuracy has nothing to do with the differences in the numbers, depending on what you are looking at. The two sites, in their most commonly presented forms of ORTG and DRTG, are measuring entirely different things.

Basketball-reference's ORTG and DRTG are individual numbers (individual points scored per individual possession used). NBA.com's are on-court numbers (how the team performs with the player on the floor).

You can find the on-court numbers on basketball-reference as well but it is buried in a separate page (on-off data) than most of their team and player statistics. There you see the differences in the way the stats are calculated (for example, by BR, DD's on-off split for DRTG is 106 on, 97.3 off, while by NBA.com it is 102.8 on and 95.6 off).

They are both actually estimates, just different ones. Basketball reference's is actually a little less accurate (at the player level), as BR uses the generic pace numbers while NBA.com is the one that uses the correct on-court pace for each player. But BR does factor in offensive team rebounds better, so at a team level their ratings are a little more accurate (NBA.com tends to overestimate pace/possessions used - and therefore underestimate both ORTG and DRTG - by about 4%.

Why is NBA.com's formula here correct and BasketballReference's isn't? Just wondering.

[hotfuzz]

Very misleading title....

[Barolt]

Why is NBA.com's formula here correct and BasketballReference's isn't? Just wondering.

Because bref pulls pace for the game, applies that pace equally to all players. Nba.com generates pace based on possessions while that player was on the floor only.