In the 29 fully audited games, stable/strong programs were shaded about 1.6 points at open. The broader opening sample is less conclusive.
The Ledger · Roster continuity
Early-Season System and Program Strength
Do program tier, prior quality, retained systems, and roster talent explain FBS-FBS opener gaps beyond the independent OnVarsity Line?
Research candidateContinuity alone was null; a small opening program-strength signal requires a prior-versus-Factor decision.Updated August 31, 2026
Decision
Do not publish generic continuity points. Test whether the preseason-known program-strength relationship belongs in the OV Rating prior before considering it as a separate opening Factor.
- Historical seasons
- 2022–25
- Primary sample
- 76 FBS-FBS openers
Start here
What the evidence says
The paired sample does not support an additional continuity adjustment after the opener.
Program-strength models improve kickoff-gap accuracy in the small sample but do not improve actual results against the market.
Because program strength is known before the season, we must test whether it belongs in the OV Rating prior rather than on the matchup card.
2026 prospective implementation
The frozen 2026 candidate states can be calculated from existing coach, quarterback, recruiting, returning-production, and prior-quality files; they are not connected to matchup pages.
Current inputs
- Coach and quarterback continuity
- Prior quality
- Recruiting and blue-chip depth
- Returning PPA
- Additional source need
- No new provider for the current candidate definition. Coordinator-specific expansion would require a new maintained source.
- Refresh cadence
- Preseason and after verified staff or starter changes.
- Coverage and limitations
- The historical improvement is small and uncertain, so candidates remain zero-point prospective observations.
Study results
| Test | Games | Result | Validation |
|---|---|---|---|
| Opening-gap MAE | 75 | 2.433 to 2.310 | Program strength without roster; interval includes no gain |
| Kickoff-gap MAE | 29 | 2.800 to 2.043 | Program-strength improvement; bootstrap interval +0.36 to +1.31 |
| Opening-to-kickoff MAE | 29 | 1.758 to 1.870 | No later-movement gain |
| Stable/strong paired opening shade | 22 | 1.57 points toward stable/strong side | 77% in that direction |
| Actual vs kickoff MAE | 29 | 13.877 to 14.531 | Program-strength model worse |
| System continuity alone | 75 opening / 29 kickoff | No dependable gain | Generic continuity remains null |
What we tested
Coach/QB continuity alone does not improve the audited baseline. In the 29 games with both established-opening and kickoff observations, stronger stable programs are shaded about 1.6 points at open and remain shaded at kickoff. There is no dependable additional movement or ATS improvement.
- Coach and quarterback continuity
- Historical conference tier and prior-year advanced-stat quality
- Roster recruiting score and blue-chip depth
- Returning PPA and prespecified interactions
Leakage and validation safeguards
- Direct comparison against no adjustment to the independent line
- Complete-season holdouts with inner ridge selection
- Unsupported extrapolation outside the historical feature range rejected
- Opening-week, Weeks 1–4, and all-season generalization compared
What we learned
- The program-strength difference is already present at the opener rather than created by later weekly news.
- Generic coach/QB continuity did not explain the market gap.
- Current-season results quickly swamp any static continuity signal.
What this does not prove
- The opener sample is small for a multivariable interaction model.
- Recruiting and roster-profile coverage is uneven.
How uses this finding
No points are displayed or summed. Program strength remains a research candidate while we test whether its preseason-known information belongs in the annual prior rather than on the matchup card.
Historical data record
Revalidated August 31 under market-line contract 2026-08-31.2. The panel contains 75 established-opening games and 29 kickoff-observed games. The downstream audit uses the validated game-flow baseline without subtracting the unsupported broad QB model.
Audited game-level table: being regenerated under the current market-line contract