Phil Steele's "National Championship Mold" is Scattershot and Arbitrary

Phil Steele's "National Championship Mold" is Scattershot and Arbitrary


Phil Steele's "National Championship Mold" is Scattershot and Arbitrary

Phil Steele pored over the stats from the seasons before every national title winner and contender over the past 20 years to determine which teams “fit the mold” of a title winner. The results are reasonable, though the methodology is problematic.

Raw Data: Steele whittles teams down using raw stats from 2011. These numbers are only relevant for 2012 as a departure point for projection. For instance, Stanford’s offensive numbers with Andrew Luck in 2011 say nothing about what will happen in 2012. The 2011 Stanford numbers with the projected rate of decline with a new quarterback after losing one drafted in the Top 10  would be far more relevant.

There is No Mold: A mold implies there is a clear correlation between sets of statistics and winning a national title. Steele is not constructing one. He is using stats of varying relevance to eliminate outliers. He keeps sculpting not to a logical conclusion, but to the point where he feels comfortable with the list.

Yards Per Game: College football is a diverse sport. It has evolved radically the past 20 years with the advent of the spread offense. It is hard to tell what a stat such as “rushing yards allowed per game” means across conferences much less across eras. It is hard to tell how that correlates to winning a national title. Yards per play would have been far more useful.

Wins: Not all wins are created equally. Teams play tougher and weaker schedules. Teams are lucky and unlucky. An 8-9 win team can have bad luck and win six. A five or six win team can have great luck and win eight. A win is also a product of factors far more indicative of a team’s true ability. Scheduling has also changed over the past 20 years with the expansion of bowl games and the 12 game schedule. “Seven wins” in 1994 is different from “Seven wins” in 2011.

Arbitrary Exclusions: Using various numbers, Steele arrives at 20 teams. His numerical analysis suggests Arkansas State, BYU, FIU, Louisiana Tech, Ohio, Southern Miss and Cincinnati all fit the mold of a national title contender. Then Steele fudges things to eliminate the undesirable teams. Hey, no non-AQ winner has won/competed for a national title (good thing Boise missed those field goals). That knocks six of the seven. Hey, Cincinnati lost eight games two years ago. No title-winner has done that. Eliminate them. TCU and West Virginia fulfill every other criterion but we should eliminate them because they are moving to a tougher conference. That leaves the 11.

Steele comes up with a list – Alabama, LSU, South Carolina, Michigan, Michigan State, Nebraska, Oklahoma, Texas, Oregon, Stanford and Florida State – that is not wholly unreasonable. One might argue USC, Georgia and, perhaps, narrow Big Ten second favorite Wisconsin should be included. One might argue a couple of those teams should not be on there. The problem is the method to obtain it is scattershot and arbitrary. It tells us nothing.

[Photo via Presswire]


Latest Leads