RealGM

Hey guys, I'm doing some research work for a school regarding sports analytics and I need your help.

Essentially I need to figure out a way to capture the usage rate of any given player for their respective teams by using minutes as a variable. This could be in the form of mintues/game, total minutes/season, etc

So one approach would be:
=(player x's total minutes played in a season)/(total minutes available per season)

also if I sum the numerator for a given team for all of their players, i should be able to get the denominator of 19680 (48 minutes per game x 5 players on the court at all times x 82 games a season). However I do not get this. The denominator, following this approach, would be close but would fluctuate by roughly 100 minutes or so. I do not understand why this is so.

We considered factors such as players being traded away to other teams, but the raw data already captures this effect. Could this be just a sampling error (rounding perhaps)?

Greatly appreciate any comments in advance, thank you very much realgm

416 wrote:Hey guys, I'm doing some research work for a school regarding sports analytics and I need your help.

Essentially I need to figure out a way to capture the usage rate of any given player for their respective teams by using minutes as a variable. This could be in the form of mintues/game, total minutes/season, etc

So one approach would be:
=(player x's total minutes played in a season)/(total minutes available per season)

also if I sum the numerator for a given team for all of their players, i should be able to get the denominator of 19680 (48 minutes per game x 5 players on the court at all times x 82 games a season). However I do not get this. The denominator, following this approach, would be close but would fluctuate by roughly 100 minutes or so. I do not understand why this is so.

We considered factors such as players being traded away to other teams, but the raw data already captures this effect. Could this be just a sampling error (rounding perhaps)?

Greatly appreciate any comments in advance, thank you very much realgm

If in a game, 9 players play 26.67 minutes each (240 total), each is recorded as 27 for a total of 243. That gives you an idea of what can happen over an entire season. You might want to take a look at USG%, an already established stat that is a bit more sophisticated.

http://insider.espn.go.com/nba/hollinge ... dusageRate

There's a snippet at the end that shows how it's calculated.

Yes thank you!

We came to this conclusion as well, it seemed weird that the average totaled greater than 240 for most of the cases so we dismissed rounding errors in the beginning, but your rationale is absolutely correct.

Also the input for USG is much appreciated, unfortunately we were unable to use it for the purposes of our research.

Please lock this thread for the issue has been resolved, once again thanks!

Yes thank you!

We came to this conclusion as well, it seemed weird that the average totaled greater than 240 for most of the cases so we dismissed rounding errors in the beginning, but your rationale is absolutely correct.

Also the input for USG is much appreciated, unfortunately we were unable to use it for the purposes of our research.

Please lock this thread for the issue has been resolved, once again thanks!

Overtime game perhaps as well? That would add another 40 minutes to the total, make that 80 since there are 10 guys on a court at once.