About the Model
Algorithmic Rankings
TJC Rankings uses a transparent, iterative algorithm to rank college football teams based on actual game data. No human bias, no conference favoritism — just math.
The model runs up to 10 iterations, recalculating team strengths based on opponent quality until rankings stabilize (typically 4-7 iterations).
On-Field Results Matter
Head-to-head results are used as a tiebreaker when two teams have played each other. If Team A is ranked above Team B despite having a lower composite score, the ↗ badge indicates that Team A won their head-to-head matchup.
This ensures that on-field results are respected — a team that won the game gets the benefit when resumes are otherwise close.
7 Metrics (Weighted)
Each metric is scored on a 0-100 scale and color-coded for quick reading:
Green = Strong (75+) • Yellow = Average (50–74) • Red = Below Average (<50)
- Win/Loss Record 10%
- Strength of Schedule 25%
- Strength of Record 20%
- Point Differential (capped at ±28/game, opponent-aware) 5%
- Offensive Efficiency 7%
- Defensive Efficiency 7%
- Quality Wins 13%
- Championship Behavior 8%
- Special Teams 3%
- Ball Control 2%
Championship Behavior
A composite score measuring how a team performs in high-leverage situations:
- Close wins (margin ≤ 7): +1 point — shows grit in tight games
- Road wins: +1 point — shows you can win away from home
- Loss to unranked (opponent composite < 60): −3 points
- Blowout loss (margin > 21): −5 points — worst penalty
- Upset loss (you were ranked 15+ spots higher): −4 points
Normalized to 0-100 (center at 50). Above 50 = good championship behavior. Below 50 = concerning losses.
Data Sources
All game data comes from CollegeFootballData.com, a free API for college football statistics.
Rankings are updated weekly during the season and finalized after bowl games.
Iterative Methodology
Because Strength of Schedule and Quality Wins depend on opponent rankings, the algorithm must iterate:
- Start with neutral rankings (all teams at rank 65)
- Calculate all 10 metrics for each team
- Generate composite scores using weights
- Re-rank teams by composite score
- Repeat until top 25 stabilizes (≤2 position changes)