We know college football is here because the pre-season polls have been released. And, we know only one fact about the polls: They are wrong. The season will not transpire as predicted. The season will not end as predicted. That is the beauty and charm of college football: Actual games will be played, we will be surprised by the results, and we will marvel at how wrong the experts were in their predictions.
Unfortunately, the polls will use these unsubstantiated starting positions to rank teams throughout the season. Losers will be demoted and winners will backfill in a nonsensical mechanical process that will confuse fans and experts alike. The reason is that apart from the Relative Performance Grading system (RPG) there is no scientific measure of how well a team plays and therefore, how good a team is. There are two reasons for this condition:
- Won/lost records are the most unreliable of all measures of team play. Wins and losses are like pass or fail grades on a school test. The winner played relatively better than the loser of any game, but W’s and L’s tell us nothing about how well either team played. A team can team play poorly and yet win if its opponent played worse. A team can play well and still lose if its opponent played relatively better. In the first case, a team gets a win that makes it look better than it is, and in the second case a team is saddled with a loss that makes it look worse than it is. Every year there are teams that are better than their records and teams that are worse than their records. The solution is a grading system that substitutes numerical grades for pass/fail grades but that leads to problem #2.
- Traditional statistics, the ones we’re fed on every broadcast or telecast of a college football game, are not useful for determining how well a team played. Traditional statistics measure effort—yards gained, first downs achieved, etc.—but not do not measure the factors that produce points on the scoreboard. Points are not awarded for yards gained or first downs. A touchdown counts the same six points no matter how long the scoring drive and no matter how scored (punt return, kickoff return, fumble return, pick six). A field goal counts the same three points no matter the distance kicked or the length of the offensive drive that preceded the kick.
To solve these problems, a new set of statistics was invented and incorporated into an algorithm that grades how well a team plays and therefore how good a team is. It is called RPG (Relative Performance Grading) and it is used to create accurate rankings that are published on www.nemosnumbers.com. You’ll find the rankings there all season. The RPG system is defined in a slim book called, Lies, Damned Lies and Statistics which is available wherever you buy books.
Let the games begin.