Tennessee vs. WSU at Intrust Bank Arena 1pm 12/14 ESPN2

[Snaggletooth]

fixed the hall and clemente, but Blair was D1 Transfer (E. Carolina)

[Wufan]

fixed the hall and clemente, but Blair was D1 Transfer (E. Carolina)

Ben Smith didn't regress.

Also, Paul Miller isn't listed and he made huge improvement...Cousinard? Were these guys not imcluded or just not listed?

[Snaggletooth]

fixed the hall and clemente, but Blair was D1 Transfer (E. Carolina)

Ben Smith didn't regress.

Also, Paul Miller isn't listed and he made huge improvement...Cousinard? Were these guys not imcluded or just not listed?

Added those players.

Might read again - their performance has been normalized for 40 minutes. What you think as improvement is just getting more playing time but their relative contribution may not be more.

Ben Smith - yes he went from 6.7 ppg to 9.8 ppg, but his minutes went from 16.7 to 25.4. Normalized for 40 minutes dropped from 16.4 to 15.6.

If you look at point differential from year to year you will see something like the following plot (this is from a slightly different data set).

Capture.JPG (47.24 KiB) Viewed 2608 times

If you use a straight point differential method, then you will get a different answer

[rlh04d]

I'm glad we're using per minute numbers instead of per game, although I'd also suggest we'd have to factor in other aspects such as rebounds, assists, turnovers, etc., on a per minute basis. Figuring out how to weigh those against each other would be tough, though.

[Wufan]

fixed the hall and clemente, but Blair was D1 Transfer (E. Carolina)

Ben Smith didn't regress.

Also, Paul Miller isn't listed and he made huge improvement...Cousinard? Were these guys not imcluded or just not listed?

Added those players.

Might read again - their performance has been normalized for 40 minutes. What you think as improvement is just getting more playing time but their relative contribution may not be more.

Ben Smith - yes he went from 6.7 ppg to 9.8 ppg, but his minutes went from 16.7 to 25.4. Normalized for 40 minutes dropped from 16.4 to 15.6.

If you look at point differential from year to year you will see something like the following plot (this is from a slightly different data set).

Capture.JPG

If you use a straight point differential method, then you will get a different answer

Thanks for clarifying. With Smith, I had already run the numbers and considered it statistically insignificant. However, normalized over all players, I think your conclusion is accurate.

[Snaggletooth]

I'm glad we're using per minute numbers instead of per game, although I'd also suggest we'd have to factor in other aspects such as rebounds, assists, turnovers, etc., on a per minute basis. Figuring out how to weigh those against each other would be tough, though.

Yeah, I've been thinking about that because you are right - scoring is not the complete picture but it is a tougher job. I have tried it a little but have never been satisfied with my weighting factors I have came up with.