Air Force’s third-down conversion rate with Kemper Hodges and without Kemper Hodges
, ,
Against what the season’s ratings expected, Air Force was 10.8 percentage points over with Kemper Hodges and 7.8 percentage points over without Kemper Hodges over its last 25 games, 2 games against 20. The plain averages are 58.5% and 52.1%, a difference of 6.4%. All 13 home games were without Kemper Hodges. All 3 opponents outside the major colleges were without Kemper Hodges. All 4 games after a bye were without Kemper Hodges.
Kemper Hodges, quarterback: 74 passing yards, 16 pass attempts and 6 completions in 15 of those 25 games.
2 games with Kemper Hodges is still too small a sample for the statistics to say much, but maybe you can find something if you squint.
third-down conversion rate a game over expected
A +3.0% difference in third-down conversion rate over expected with Kemper Hodges vs without Kemper Hodges is about what you’d expect.
A +6.4% difference in third-down conversion rate with Kemper Hodges vs without Kemper Hodges is about what you’d expect.
The games
| Side | Date | Game | Result | 3rd downs | Expected |
|---|---|---|---|---|---|
| Nov 2, 2024 | @ Army | L | 33.3% | 43.4% | |
| Nov 9, 2024 | vs Fresno State | W | 52.6% | 36.7% | |
| Nov 16, 2024 | vs Oregon State | W | 66.7% | 40.6% | |
| Nov 23, 2024 | @ Nevada | W | 68.8% | 43.5% | |
| Nov 30, 2024 | @ San Diego State | W | 66.7% | 43.6% | |
| Aug 30, 2025 | vs Bucknell | W | 66.7% | — | |
| Sep 13, 2025 | @ Utah State | L | 44.4% | 44.7% | |
| Sep 20, 2025 | vs Boise State | L | 57.1% | 40.5% | |
| Sep 27, 2025 | vs Hawaii | L | 50.0% | 46.5% | |
| Oct 4, 2025 | @ Navy | L | 53.8% | 44.6% | |
| Oct 11, 2025 | @ UNLV | L | 40.0% | 41.3% | |
| Oct 18, 2025 | vs Wyoming | W | 42.9% | 44.0% | |
| Nov 1, 2025 | vsArmy | L | 57.1% | 47.5% | |
| Nov 8, 2025 | @ San Jose State | W | 47.1% | 43.8% | |
| Nov 15, 2025 | @ UConn | L | 40.0% | 47.6% | |
| Nov 22, 2025 | vs New Mexico | L | 38.5% | 46.3% | |
| Nov 28, 2025 | @ Colorado State | W | 73.3% | 47.4% | |
| Sep 5, 2026 | vs Duquesne | W | 60.0% | — | |
| Sep 12, 2026 | vsNNorth Dakota State | L | 42.9% | — | |
| Sep 26, 2026 | @Nevada | W | 44.4% | 36.8% | |
| Oct 3, 2026 | vsNavy | W | 50.0% | 45.1% | |
| Oct 10, 2026 | @ Northern Illinois | W | 76.9% | 47.7% | |
| Oct 17, 2026 | vsUNLV | W | 44.4% | 45.5% | |
| Oct 23, 2026 | @Wyoming | L | 53.3% | 42.6% | |
| Oct 31, 2026 | vsUConn | W | 43.8% | 48.0% |
Method
- Expected is an offense and a defense a team, fit over every game of its season between two teams in the table with the game being read left out, plus 0.6% in 2024, plus 0.7% in 2025, plus 1.4% in 2026 for the home team (the edge the fit finds once strength is in), nothing at a neutral site.
- Each expected number is then fit to the season’s games: a game bears out about 50% in 2024, 56% in 2025, 40% in 2026 of the difference the model predicts.
- The shuffle keeps the season’s shape: the games are cut in two where one date leaves the least variance, each game is read against its stretch, and the sides are shuffled only inside each stretch.
- Bucknell, Duquesne, North Dakota State have no other game in the table, so those games have no expected number and sit out the second figure.
- Who played is read from the schools’ box scores’ participation lists for 20 of these 25 games; in the other 5 a player is counted as in them when the box score has a stat for him.
- With him means he was the team’s quarterback that game, at least half of its pass attempts; a game he left early or came into late reads as without him.