Here We Go Again! Records and Eyeballs Deceived the Committee

Bill Belichick once said that you are what your record says you are. While that may be true in professional football, it is far from true in college football. College football schedules vary wildly from one conference to another and the College Football Playoff Committee admits on its Website that it has no tool to measure schedule strength or otherwise distinguish one schedule from another.

While pro teams may play consistently and follow a trajectory of steady improvement throughout a season, a plot of any college team’s week-to-week performances would look like a map of the Rocky Mountains—all peaks and valleys.

Won/lost records mask these variations in schedule strength and playing performance, leaving the false impression that records can be compared. On last night’s College Football Playoff Committee rankings reveal show, the topic of most interest was how the committee would rank one loss teams because it was assumed that undefeated teams would be at the top of the rankings simply because they were undefeated.

Experts should realize that a college football team can play poorly and still win a game if its opponent plays relatively worse. Notre Dame played very poorly against Ball State and Vanderbilt but managed wins in both cases. The Irish played mediocre football against Virginia Tech, Pitt and Navy but walked away with victories in all three cases. At the same time, Michigan played very well in a string of wins over Nebraska, Northwestern, Maryland, Wisconsin and Michigan State. Michigan proved itself better than Notre Dame because it played better football in all five games than did Notre Dame in the five games used as an example above.

Yet, the committee ranked Notre Dame one spot above Michigan, possibly because Notre Dame squeaked past the Wolverines in the first game of the season. It seems logical that a head-to-head winner should be ranked ahead of the loser, and it’s very difficult for people inside the game to give up old ways of thinking, but Notre Dame has regressed to its playing performance mean since that game while Michigan has progressed to its higher mean.

It is also undeniably true that a team can play well and still lose if its opponent plays relatively better in an individual game. Washington played better in its losses to Oregon and Cal than Notre Dame did in its wins over Ball State and Vanderbilt but the Huskies losses make their record look worse than Notre Dame’s and therefore the Huskies are unranked by the Committee while Notre Dame is illogically ranked No. 4.

The Relative Performance Grading System (RPG) follows a simple principle: A team is precisely as good as it plays the game, without regard for the win or lose outcome. RPG grades each playing performance in the same way that a teacher puts a numerical grade on a paper, a test or a project by a school student.

As a result, our rankings look different than Playoff Committee rankings. While we agree that Alabama and Clemson are head and shoulders above all other teams, we have Georgia at No. 3 and LSU at No. 4 (Yes, Georgia has played better overall than the team that beat them!) and Michigan at No. 5 and Notre Dame at No. 9 (Yes, Michigan has played better overall than the team that beat them!). We do not care who beat whom; we only care about how well the teams have played the game.

Each year the Committee has invited one team to the playoffs that didn’t deserve an invite. This year Notre Dame could be that team.

Waive goodbye to Wisconsin and Stanford who’ve been dropped from our rankings and say hello to Kentucky (No. 10) and Washington State (No. 11) which have been added to our rankings.

For a complete list of ranked teams and weekend playing performance grades, visit www.nemosnumbers.com. If you want to know how the numerical grades are calculated, find a copy of Lies, Damned Lies and Statistics wherever you buy books.

By |2018-10-31T09:33:46+00:00October 31st, 2018|Blog|