APEX Team Rankings
Experimental weekly models. These are reproducible calculations from observed results, not approved final APEX formulas. The 0–100 indices are not win probabilities.
| Rank | Team | W–L | Index |
|---|
Methods, sources & missing data
Résumé: Colley schedule-adjusted win/loss baseline. Opponent results matter; winning margins do not. The raw Colley rating r is transformed to the bounded index 50 + 50 × tanh(2r − 1).
Power: the existing opponent-adjusted ridge model with diminishing score margins, displayed as 50 + 50 × tanh(raw rating). This baseline is not a point-spread forecast.
Moneyball: current performance against recovered estimated resources. Financial estimates are incomplete and independently unverified; a covered-cohort rank is not a national rank.
SOS: mean fitted opponent strength through the selected retrospective cutoff, displayed as 50 + 50 × tanh(raw strength). Full FCS schedules remain incomplete.
APEX gives SOS a direct contribution and caps Moneyball at 5%. SOS also influences opponent-adjusted Résumé and Power, and Moneyball reuses these scores. Direct weights are not independent causal contributions.
Where movement is available, the same current model is recomputed now on the previous game set. That comparison is not a previously published ranking or a verified historical forecast.