Played like a No. 16–30 offense in this game
84th percentile among completed 2017 FBS offense performancesThe verdict
How each unit played in this game
These are single-game performance grades, translated to a neutral field against an average FBS opponent. They describe this game as played; none of the four ratings is pushed toward average.
Played like a No. 76–100 defense in this game
32nd percentile among completed 2017 FBS defense performancesPlayed like a No. 31–50 offense in this game
67th percentile among completed 2017 FBS offense performancesPlayed like a No. 101+ defense in this game
18th percentile among completed 2017 FBS defense performancesOpponent + location
A simple translation, not a rewrite.
- Keep the game.Start with every observed play and non-garbage drive.
- Account for the opponent.Compare each output with what that opponent produced or allowed across 2017, leaving this game out.
- Move to neutral.The home team gives back 1.7 percentile points; the road team receives the same allowance.
The evidence
Six views, one common scale
A percentile of 50 is an average FBS game. Higher is always better for the named team, including defense. The raw number sits beside each percentile so the chart never hides the box-score result.
Team view
Idaho
OffenseProduced
DefenseAllowed
Points per driveOffensive points across non-garbage possessions33.7% weight
561.75
442.33
Line yards per rushRushing value credited mostly to blocking and front play26.0% weight
993.26
103.11
Yards per playTotal scrimmage yards divided by offensive plays26.5% weight
595.58
645.56
Late-down successCombined third- and fourth-down conversion rate13.8% weight
8710/19 · 52.6%
98/13 · 61.5%
Adjusted drive successDrive success above expectation from the starting field positionCross-check
72-0.8 pts
73-13.3 pts
EPA per playExpected points added on the average scrimmage playCross-check
80+0.11
43+0.10
Team view
Troy
OffenseProduced
DefenseAllowed
Points per driveOffensive points across non-garbage possessions33.7% weight
542.33
441.75
Line yards per rushRushing value credited mostly to blocking and front play26.0% weight
883.11
73.26
Yards per playTotal scrimmage yards divided by offensive plays26.5% weight
375.56
385.58
Late-down successCombined third- and fourth-down conversion rate13.8% weight
908/13 · 61.5%
1110/19 · 52.6%
Adjusted drive successDrive success above expectation from the starting field positionCross-check
18-13.3 pts
33-0.8 pts
EPA per playExpected points added on the average scrimmage playCross-check
61+0.10
22+0.11
How the summary is built
Four forward-tested signals set the overall grade.
The offense and defense summaries start as weighted averages of four opponent-adjusted game percentiles. Forward selection kept a factor only when it improved later-game accuracy out of sample. The weighted result is then ranked against completed FBS unit performances from the same season, so a 95 means top 5% for that year. This changes the scale, not the order, and the observed game is never pushed toward 50.
- Points per drive
- 33.7%
- Line yards per rush
- 26.0%
- Yards per play
- 26.5%
- Late-down success
- 13.8%
Adjusted drive success
A touchdown or field-goal try counts as a successful drive, then the historical success probability from that starting yard line is subtracted. A short-field score therefore earns less credit. This replaces pure drive success everywhere on the page.
EPA per play
EPA was the strongest requested statistic when used by itself. Readers who distrust any four-factor weighting can use it as an independent check. Neither EPA nor adjusted drive success added reliable accuracy after the four selected factors, so they are shown without double-counting them in the summary.
Validation and calculation notes
Candidate statistics were trained on 2023, selected only when they improved 2024 out-of-sample results, then evaluated on 1,070 untouched 2025 team-games. The current weights are 33.7% points per drive, 26.0% line yards, 26.5% yards per play and 13.8% late-down success.
This is a retrospective application of those locked weights. Opponent quality comes from the opponent’s other games in 2017; only that opponent baseline is stabilized toward the national average. The performance observed in this game is not shrunk.
The adjusted-drive idea follows the same field-position logic as BCF’s drive-value and available-yards measures; All Takes computes its own historical success expectation by starting yard line. BCF glossary ↗