Objective

It is NOT the goal of Logical College Football Rankings (LCFR) to provide power rankings for the purpose of predicting future outcomes of college football games.  However, LCFR is laser-focused on creating rankings that most accurately balance the wins and losses of each competing school. 

 

LCFR is strongly opposed to subjective evaluations of teams, such as the "eye-test" and human polls.  LCFR has invented two statistics, Transitive Winning Percentage and Earned Winning Percentage, that when utilized in conjunction with other traditional, objective measures and logical reasoning, produce an arrangement of the ranked teams that is both equitable and explainable.

 

Stage 1: Three Winning Percentages      

The multi-step process of ranking teams begins with calculating the Rated Winning Percentage (RPCT) of each team.  The Rated Wins (RW) of each team include only victories versus FBS schools.  RPCT = RW/(RW+L).

 

The Transitive Winning Percentage (TPCT) of each team is computed next.  The concept of Transitive Wins (TW) is explained in the Glossary  

​

Earned Winning Percentage (EPCT) is calculated next.  The derivation of the formula for Earned Wins can be illustrated with the following example.  Suppose a group of 13 teams only plays against teams within the group and that each team plays every other team exactly once.  This results in every team playing 12 games (G). Further suppose that there are no upsets in any of these 78 games.  If a team earns 1 WinPoint for defeating a 1-win team, 2 WinPoints for defeating a 2-win team, 3 WinPoints for defeating a 3-win team, etcetera, and LossPoints are accumulated in like manner, the final standings would be those listed in the table below.

    ​

​

​

​

​

​

​

​

​

​

​

 

 

 

 

​

For each team, Wins = Difference/(12-1) + 12/2.  The generalized rule is the formula for Earned Wins: EW = (WP - LP)/(G-1) + G/2.

​

Stage 2: Projecting Plus-Minus

The Plus-Minus (PM) for each team = Upset Wins - Upset Losses.  The ultimate goal of Stage 2 is to use the three winning percentages calculated in Stage 1 to identify which game outcomes are most likely to be upsets.

​​​

The example below comparing Air Force vs Wyoming in 2025 will serve to Illustrate how the likelihood of an upset is computed.​​​​​

​

​

​

​

​

​

​

​

 

 

 

 

 

 

 

Wyoming has the 9th highest RPCT, the 9th highest TPCT, and the 9th highest EPCT among all Air Force opponents.  Air Force has 8 Losses, so Air Force has a PM = 0 if ranked just above Wyoming and a PM = 1 if ranked just below Wyoming.  Air Force has the 7th highest RPCT, the 9th highest TPCT, and the 7th highest EPCT among all Wyoming opponents.  Wyoming has 8 losses, so Wyoming has at worst a PM = -2 and at best a PM = 0 if ranked above Air Force.  This is very strong evidence that the Air Force victory over Wyoming was not an upset. 

 

The comparison above is repeated for every team with each one of its opponents.  Initially, all comparisons that do not produce a unanimous verdict are decided by the Head-to-Head outcome.

 

Stage 3: Forming Clusters

Clusters of teams are formed when the comparisons done in Stage 2 create contradictions.  A team will also be added to a cluster if its EPM (EW + PM) is greater than the average EPM of the teams already in the cluster.  The clusters are then ranked according to this average.  Some teams may not be members of a cluster and are ranked as a one-team cluster. 

 

Stage 4: Ranking Teams in a Cluster 

​​Final adjustments to the rankings are executed only within each cluster.  These alterations are performed beginning with five objective rules and conclude with an optional, subjective procedure.​

• Consecutively ranked teams trade places if this reduces the standard deviation of the PM's of the teams.

• Consecutively ranked teams that did not play each other trade places if the team with the greater winning percentage is ranked beneath the team with the lesser winning percentage.

• Consecutively ranked teams that did not play each other with equal winning percentages trade places if the team with the greater EPM is ranked beneath the team with the lesser EPM.

• Consecutively ranked teams that played each other trade places if the two teams are logically tied and the team ranked below has the greater winning percentage AND the greater EPM.  

• Consecutively ranked teams that played each other trade places if the two teams are logically tied, have equal winning percentages, and the team ranked below has the greater EPM.

• Optionally, teams within the same cluster may be rearranged in order to balance the PM's of the conferences.  This must be accomplished in a way that does not trigger any of the five rules of adjustment listed above.