Historical NFL game outcomes · 2010–2025
How often does an NFL favorite lose?
A team can be a strong favorite and still lose often enough to end a survivor entry. We grouped completed regular-season games by each favorite’s no-vig recorded-moneyline estimate and counted what happened.
In the 70–75% no-vig moneyline-probability band, the favorite lost 151 of 598 games from 2010–2025. That is about one loss in four in this sample.
Historical results by no-vig moneyline-probability range
| Favorite estimate | Games | Won | Lost | Observed win rate | 95% interval |
|---|---|---|---|---|---|
| 50–55% | 542 | 288 | 254 | 48.9%–57.3% | |
| 55–60% | 740 | 426 | 314 | 54.0%–61.1% | |
| 60–65% | 779 | 464 | 315 | 56.1%–63.0% | |
| 65–70% | 619 | 424 | 195 | 64.7%–72.0% | |
| 70–75% | 598 | 447 | 151 | 71.1%–78.1% | |
| 75–80% | 452 | 357 | 95 | 75.0%–82.5% | |
| 80–85% | 225 | 183 | 42 | 75.7%–85.9% | |
| 85%+ | 185 | 170 | 15 | 87.1%–95.0% |
Each probability range includes its lower bound and excludes its upper bound; the final range is 85% and above. The number of games differs by range because games were grouped by their market estimates.
What this means for a survivor pick
A 70–75% estimate describes one game, not the chance an entry survives the season. In this historical sample, favorites in that range lost 25.3% of the time. A team’s estimate is a risk measure, not a promise.
Survivor choices also depend on teams already used, future matchups, pool rules, and—if your payout rewards finishing alone—how many other entries may pick the same team. This chart measures none of those factors and does not show the performance of Thrive ’N’ Survive’s model or optimizer.
Data and method
Source: nflverse’s public NFL games dataset, retrieved 2026-10-14. The analysis covers regular-season games in 2010–2025; 4,175 source rows were available in those seasons.
For each game, we converted both recorded American moneylines to raw implied probabilities, then divided each side by the sum of both probabilities to remove the market margin. The side with the higher resulting probability was the favorite. A game counted as a win when that team scored more points.
We excluded 1 game with incomplete scores or two-sided odds, 13 tied games, and 21 exact even-market games where neither side was a favorite. This leaves 4,140 decisive games. Each row reports the observed win rate and its 95% Wilson interval.
Limits: nflverse documents game scores and game types but does not document when these moneyline prices were captured. Treat them as recorded market estimates, not verified closing odds; they may differ from prices at a pool’s pick deadline. The Wilson intervals are approximate binomial intervals and do not adjust for repeated teams or season effects. The market is not our model, the sample is not a strategy backtest, and results cannot guarantee future outcomes.
Reproducibility: source snapshot SHA-256 1ab47132ce926dd4c53032f1e744463629f132ddf9eeb5bdbca983d990588649.