Matchup calculator
2026 model · generated 2026-11-06Select two teams to compare projected margins.
Matchup projections
Team comparison
Comparison — opponent-adjusted EPA (ridge)| Stat | LSU | Notre Dame |
|---|---|---|
| Conference | SEC | FBS Independents |
| Net rating | 0.260 | 0.250 |
| Adj EPA (off) | 0.233 | 0.268 |
| Adj EPA (allowed) | -0.026 | 0.018 |
| Recent net (last ≤5 wks) | 0.205 · 0.211 · 0.262 · 0.254 · 0.260 | 0.238 · 0.279 · 0.276 · 0.265 · 0.250 |
| Home record · avg margin | 4-0 · 4 games · +33.0 | 3-0 · 3 games · +40.0 |
| Opponent EPA net (EPA/play) · FBS opponents, rated as of today | 0.067 · 7 games | -0.014 · 8 games |
- Output:
- Team A margin (points); logistic win estimate (%)
- Model:
- EPA-OLS calculator with home-field prior; separate from published picks
- Training:
- 2026 fit; exact training snapshot and cutoff unavailable
- Features:
- Opponent-adjusted EPA (ridge); feature snapshot unavailable
- Evaluation:
- Unavailable — fitted/evaluation identity not recorded. This calculator's scale cannot calibrate the current card.
Positive margin favors Team A. The win estimate is a logistic conversion of this margin, not an independently fitted win-probability forecast or a betting confidence score. Home field is worth +4.0 in this fit, shrunk toward +2.5 with the weight of 40 games (5 neutral-site games in the training set).
2026 · Through provider Week 9. Each metric uses only published FBS teams with a numeric value. Higher percentile is better; tied values receive the midpoint of their positions, rounded to a whole percentile. Missing values are unranked.
| Axis | LSU | Notre Dame |
|---|---|---|
| Off PPA | 91th · n=138 | 99th · n=138 |
| Off success | 97th · n=138 | 98th · n=138 |
| Off explosive | 34th · n=138 | 28th · n=138 |
| Off line yds | 56th · n=138 | 78th · n=138 |
| Def PPA | 95th · n=138 | 81th · n=138 |
| Def success | 92th · n=138 | 78th · n=138 |
| Def explosive | 63th · n=138 | 73th · n=138 |
| Def havoc | 98th · n=138 | 84th · n=138 |
Percentile of FBS teams in the metric's better direction (defense: lower allowed ranks higher). Raw season PPA, no opponent adjustment.
Model output, not betting advice. Historical hit rates are small-sample backtests; several edge buckets lose money at -110. Flat units only.