Using the Memphis Tigers To Critique the Pomeroy Ratings

Jameson FlemingSenior Writer IFebruary 25, 2009

I want to preface my article by stating what my intended goals are for this article. I am critiquing a statistical system that gauges the strength of college basketball teams using tempo free statistics. I'm going to attempt to prove that an adjusted strength of schedule for an opposition's offensive and defensive efficiencies doesn't go far enough in Ken Pomeroy's efficiency models.

Pomeroy statistics have been very accurate in the past. Trends show teams generally must be in the top 25 in defensive efficiency if they want to be a legitimate contender to reach the Final Four. Typically teams that make the final weekend in college basketball are rated very highly overall by Pomeroy.

When checking out the current version of Pomeroy's ratings, you can find Memphis ranked first, West Virginia sixth, UCLA ninth, and BYU 14th. In reality, it's hard to justify by the results on the court the rankings of those teams. Those rankings provided the inspiration for this critique.

In my personal subjective view, I acknowledge Memphis has been a much better team since Coach John Calipari moved freshman Tyreke Evans to the point guard position. The Tigers have quickly become a top 15 team (again in my own personal opinion).

I have become skeptical this season that Memphis deserves to be ranked in the top five after beating up on a conference ranked eighth overall. I thought the Tigers' rise through the rankings was due mainly to major conference teams beating each other up.

I've been tying to find a way for about two weeks to objectively prove my subjective opinion. This is where my critique of Pomeroy's system comes in.

Using the performance of the team I follow the closest Syracuse and the game data of Memphis' conference results, I found my way to prove how the Tigers' weak schedule against Conference USA teams has inflated their adjusted scoring margin or position in Pomeroy's rankings.

Entering the Orange's game against St. John's on Tuesday in Madison Square Garden, Ken Pomeroy's system ranked Syracuse 34th overall. The 'Cuse also stood 24th offensively with an offensive rating around 113.5. Defensively, SU was much worse with a 60th place ranking and defensive rating of 94.8.

After whooping the Red Storm 87-58, Syracuse's rankings improved...dramatically.

By beating St. Johns, who Pomeroy ranked 120th as of Wednesday by 29 points in a 74 possession games, the 'Cuse suddenly jumped into the top 25 at 24th for essentially the first time all year. Th Orange moved up 10 spots with just one victory with a mere three regular season games left in the season.

Post St. John's, the Orange's offensive rating improved 114.7, a jump of .12 points per possession, and moved up to 19th overall. Defensively, the 'Cuse improved even more dramatically from 60th to 52nd and now SU allows .11 points per possession less.

So now it's time to enter in Memphis. It has been hard to get a gauge on the Tigers since they started playing better in conference play. John Calipari's club has fittingly faced eight opponents that rank better and eight opponents ranked worse than 120th, the rank of St. Johns. 

So I got thinking, with St. John' right in the middle, where would Syracuse rank overall if it played the Tigers' conference schedule.

In the non-conference slate, SU scored 1056 points and allowed 874 in 953 possessions. Using Ken Pomeroy's formula for Pythagorean winning percentage (substitute 11.5 for the exponent of two), Syracuse's non-conference winning percentage was .8980, good for 32nd currently.

I debated what way would be the best method to determine the 'Cuse's success if they played in Conference USA. I decided that I would project the Orange's Pythagorean winning percentage using four games against four of Syracuse's worst opponents. Those four games are 86th-ranked Seton Hall, 120th-ranked St. John's, 135th-ranked Rutgers, and 220th-ranked DePaul. Two of those games are road games, and two are home games.

In those four games, Syracuse scored 1.1879 points per possession and allowed 0.8657 points per possession. To make that an adjusted efficiency rating, I multiple the offensive efficiency by national average (1.01) and divide by the opponents average defensive efficiency (99.2). I do the same for the defensive, so (.8657*1.01)/1.023  The adjusted offensive efficiency is 1.209474. The adjusted defensive efficiency is .85477.

In a 16 game schedule played at a pace of 74.5 possessions (the average of the four games), SU would score 1442 points and allow 1019 points. If you add those two totals to the non-conference totals, you get 2497 points scored and 1893 points allowed. SU would have 2145 possessions in the season, the offensive efficiency for the season would be 1.1614 points per possession. The defensive efficiency for the season would be .8825 points per possession.

Entering those totals into the Pythagorean winning percentage theorem, Syracuse's winning percentage is .960253. That winning percentage would rank the Orange eighth overall as opposed to 24th.

Memphis doesn't have results to determine what their Pythagorean winning percentage would be using Syracuse's schedule, but the Tigers have been more dominant then the proposed SU theory.

With that said, against the Orange's schedule, the Tigers would likely drop 10-13 spots instead the 16 spots to mirror SU's ascension through the rankings.