NCAA player rankings

I've made the first run through an NCAA player rankings system similar to my NBA player rankings. The linked post describes the system pretty well, but their are some differences in the collegiate system, mostly due to different data sets used.

The basic formula for Total Performance is exactly the same:

TP = Points scored - shots missed (FG or FT) + rebounds + assists - turnovers + steals + blocks - fouls

A very basic approach that works well as a substitute for points per game to reflect a player's overall contribution. The first college player ranking (A: Per Game) is simply the player's average Total Performace Points per game.

	Name	Team	Per Game	Total	MVP Rating
1	Michael Beasley	Kansas St.	28.2	789.6	507.7
2	Reggie Williams	Virginia Military	27.3	655.2	327.6
3	Arizona 'AZ' Reid	High Point	25.6	742.4	409.8
4	Tyler Hansbrough	North Carolina	24.5	710.5	661.5
5	Jason Thompson	Rider	23.9	693.1	478.2
6	Lester Hudson	Tenn-Martin	23.5	728.5	375.9
7	Marqus Blakely	Vermont	22.4	582.4	291.2
8	George Hill	IUPUI	22.0	638.0	510.4
9	Kevin Love	UCLA	21.7	607.6	542.6
10	Will Thomas	George Mason	21.1	633.0	422.2

This gives an interesting variety of players from top teams (Kansas State, North Carolina, UCLA) and smaller schools, similar to what the pure point averages show, but reflecting overall game contribution.

The second chart (B: Total Output) covers the entire season. As such, players who miss a lot of the season are lower in this chart, but it's the strongest reflection of what a player has done to help his team.

	Name	Team	Per Game	Total	MVP Rating
1	Michael Beasley	Kansas St.	28.2	789.6	507.7
2	Arizona 'AZ' Reid	High Point	25.6	742.4	409.8
3	Lester Hudson	Tenn-Martin	23.5	728.5	375.9
4	Tyler Hansbrough	North Carolina	24.5	710.5	661.5
5	Jason Thompson	Rider	23.9	693.1	478.2
6	Reggie Williams	Virginia Military	27.3	655.2	327.6
7	George Hill	IUPUI	22.0	638.0	510.4
8	Will Thomas	George Mason	21.1	633.0	422.2
9	Jaycee Carroll	Utah St.	21.0	609.0	420.2
10	Kevin Love	UCLA	21.7	607.6	542.6

As you can see, the chart is similar in the top ten with some shuffling. Having both charts allows players who missed half the season to show up on the Per Game ratings, and the Total Output chart gives an edge to players who have contributed to their team the whole season.

This is where the break from the NBA ratings occurs. I don't use stats that contain minutes played, so I don't deal with efficiency stats for college players.

For the final chart, MVP Ratings, each player's Total Output score is multiplied by their team's winning percentage.

	Name	Team	Per Game	Total	MVP Rating
1	Tyler Hansbrough	North Carolina	24.5	710.5	661.5
2	Kevin Love	UCLA	21.7	607.6	542.6
3	George Hill	IUPUI	22.0	638.0	510.4
4	Michael Beasley	Kansas St.	28.2	789.6	507.7
5	D.J. White	Indiana	20.3	568.4	487.1
6	Jason Thompson	Rider	23.9	693.1	478.2
7	Stephen Curry	Davidson	20.5	594.5	471.4
8	Demetric Bennett	South Alabama	18.3	549.0	457.3
9	Tony Lee	Robert Morris	17.5	542.5	437.3
10	Reggie Larry	Boise St.	19.8	574.2	435.8

Doing this gives an indication of how much a player is helping his team win in addition to racking up monster stats. Admittedly this works better for the NBA, where schedule strength is much more even. But as a general indicator it works well, and is simple to implement.

Some argue (and I agree) that even if winning percentage were a perfect reflection of team strength, it's unfair to players who play on worse teams to penalize them for lack of quality teammates; but at the same time, a great player with lesser teammates has much more opportunity to rack up the great stats.

Two comparisons show how various arguments can be made. Note that Michael Beasley, who dominated the Per Game and Total Output charts, falls to number four in the MVP Ratings. There's a big debate right now as to who should be MVP: Beasley or Tyler Hansbrough, who is at #1 on the chart right now. Beasley's stats are even better than Hansbrough's, but North Carolina's 27-2 beats Kansas State's 18-10 by so much, Hansbrough zooms to the top.

Let's put aside for now the argument that Kansas State is much better than their record indicates, as that's a whole different take on the subject. The basic question is, should Hansbrough be lofted above Beasley for being on a better team? Or would Hansbrough, on a lesser team like K State, put up even bigger numbers and look like the better player himself?

I argue for the latter, though I understand the problem. It's never going to be perfect, so I opt for simplicity. I assume that great teams have a number of contributors, whose stats will appear to be low for "sharing the wealth" with each other, and using team strength is a way to heighten their recognition. A great individual player still has to make his team win; otherwise, he's just another great player, who may or may not be garnering the points and rebounds if he had other great teammates. The best possible MVP would be someone like Larry Bird in 1979, who led Indiana State to an undefeated season while putting up unbelievable numbers.

Pitting Beasley vs. Kevin Love of UCLA is another example. Both are vying for Freshman of the Year, and the consensus has Beasley in a walk. Here, though, UCLA's record is the deciding factor in favor of Love. I guess the question is: have the two players switch teams, and how do their records change? How do their stats change? Hard to say.

If Kansas State were better represented that by record alone, say by a power rating, it would make a difference, and that's an option I may explore: using power ratings from 1 to 341 on a scale where #1 = 100%, down to 0%. It would certainly eliminate the current situation whereby IUPUI rates higher than Kansas State, putting George Hill ahead of Beasley.

Comments

[Basketball Picks]

Thanks for the information on topics.I was excited for this article.
Thank you again.