APBRmetrics

[m_c_meyer]

I love all the NBA analysis presented by the members represented in this forum. Great Stuff!

I'm curious if anyone has applied the same methodology\analysis to the college game? And if so, is any of this available for on-line viewing?

[HoopStudies]

[SGreenwell]

I've tried my best to translate A-10 numbers to a PER formula, along with usage rate, PSA, rebound rate, etc. I did it for last year, but only for A-10 games. By doing this, I ran into sample size problems - It's only a sample of 16 games. Nonetheless, the top ten seems pretty stable:

1) Jameer Nelson, St. Joe's, 0.54
2) Delonte West, St. Joe's, 0.53
3) David Hawkins, Temple, 0.47
4) Pops Mensah-Bonsu, George Washington, 0.43
5) Tony Dobbins, Richmond, 0.40
6) J.R. Pinnock, George Washington, 0.38
7) Mike Hall, George Washington, 0.37
Steven Smith, La Salle, 0.37
9) Michael Haynes, Fordham, 0.36
10) Kieron Achara, Duquense, 0.36

I have not yet calculated what a league average player is yet, hence no standardization to 15. I also haven't entered the data for UMass yet.

Still, the list seems solid - Barely missing are some players people probably consider better, like Romain Sato, Lionel Chalmers and Mardy Collins.

I only did A-10 games to eliminate the noise from teams like St. Joe's beating up on D2 colleges by 50 points. Every team played every other team, although you will see some discrepiances. Some teams only had to play the St. Joe's juggernaut once, for example.

This unfortunately cuts down on sample size. I think you could include all 30-so games on a team's schedule if you did the *entire* NCAA and included some sort of conference or schedule adjustment, but that's too much of a project for me. I'm a full-time college student and managing editor at my college paper, and a weekend reporter, with little background in stats, so anything beyond applying the work of Hollinger and others is beyond me.

I also calculated things like defensive and offense efficiency and expected wins via pythagoreon theorm. I'll be doing the same thing with this year's numbers once the season ends, although I have no idea how long the process will take me. My interest was only recently revived because I was forced to write a column to fill up an empty sports page one night.

[BigSumo]

I also have an interest in analyzing NCAA Basketball. I have calculated stats per 40 minutes, foul shot effectiveness for both makes and attempts, floor percentage, effective shooting percentages, and a player and team offensive rating. I also have calculated a team defensive rating using opponents stats, but there is one area which I would like to improve and that is matchup probablilties and predictions using the above statistics.

Here are a few of the numbers I have calculated:

Team A Offensive Rating 116.7
Team A Defensive Rating 90.5
Team A Floor Percentage .525
Team A Floor Percentage Defense .455
Team A Play Percentage .414
Team A Play Percentage Defense .320

Team B Offensive Rating 112.7
Team B Defensive Rating 89.1
Team B Floor Percentage .565
Team B Floor Percentage Defense .438
Team B Play Percentage .418
Team B Play Percentage Defense .330

Given these ratings, how would I go about calculating either the percentage of the time that Team A would defeat Team B, or a possible point spread of Team A over Team B?