(03-01-2013 09:15 AM)Eastside_J Wrote:  

(02-28-2013 06:08 PM)Bearhawkeye Wrote:  

(02-28-2013 04:17 PM)bearcatmark Wrote:  I mean doing anything well is going to have a positive correlation to winning basketball games. Quite the revelation.

I agree those stats don't really tell us much, but it would be interesting to see some regression analysis on which stats correlate most to winning and by how much.

That would be tremendous and probably deserves its own thread to figure out how it could best be done.

My initial impression is that it might work best to start the study using NBA statistics.

Strength of schedule is such a massive variable in college ball. So great that I think it makes it very tough to make apples to apples stat comparisons. For instance Davidson is the number 1 team last time I looked in FT shooting percentage. They pretty much shoot the lights out all around. And they win. But if they played a Big East schedule and had to face big east defenses, most all those numbers would decrease, as would their winning percentage.

Maybe it is possible that you could take the stats and winning % and then apply a + or - strength of schedule factor and then back test your results against past seasons game/season outcomes.

Someone has to have done this right?

I've often seen these listed as the "The Four Factors of Winning"

effective FG% (adjust for 3-point FG being worth more)
turnover %
offensive rebounding %
FT rate (how often you get to the line)

The Four Factors based on importance:
Shooting (40%)
Turnovers (25%)
Rebounding (20%)
Free Throws (15%)

http://statsheet.com/mcb/games/2013/03/0...cincinnati
http://www.basketball-reference.com/about/factors.html

(03-01-2013 11:35 AM)Bearcat04 Wrote:  

(03-01-2013 09:15 AM)Eastside_J Wrote:  

(02-28-2013 06:08 PM)Bearhawkeye Wrote:  

(02-28-2013 04:17 PM)bearcatmark Wrote:  I mean doing anything well is going to have a positive correlation to winning basketball games. Quite the revelation.

I agree those stats don't really tell us much, but it would be interesting to see some regression analysis on which stats correlate most to winning and by how much.

That would be tremendous and probably deserves its own thread to figure out how it could best be done.

My initial impression is that it might work best to start the study using NBA statistics.

Strength of schedule is such a massive variable in college ball. So great that I think it makes it very tough to make apples to apples stat comparisons. For instance Davidson is the number 1 team last time I looked in FT shooting percentage. They pretty much shoot the lights out all around. And they win. But if they played a Big East schedule and had to face big east defenses, most all those numbers would decrease, as would their winning percentage.

Maybe it is possible that you could take the stats and winning % and then apply a + or - strength of schedule factor and then back test your results against past seasons game/season outcomes.

Someone has to have done this right?

I've often seen these listed as the "The Four Factors of Winning"

effective FG% (adjust for 3-point FG being worth more)
turnover %
offensive rebounding %
FT rate (how often you get to the line)

The Four Factors based on importance:
Shooting (40%)
Turnovers (25%)
Rebounding (20%)
Free Throws (15%)

http://statsheet.com/mcb/games/2013/03/0...cincinnati
http://www.basketball-reference.com/about/factors.html

Good stuff. I'd seen this before. Getting to the line being more important than percentage (though percentage is factored in)