NFL Win Totals: Three Regression Signals That Still Beat the Summer Market
Training camp opened July 17. All 32 teams will have reported by July 28. The Hall of Fame Game is August 6. The window to find value in NFL season win totals is not closed, but it is narrowing.
The market begins taking shape in February, within days of the Super Bowl, when books post opening totals. By July, the major moves from free agency and the draft are priced in. Teams losing cornerstone players see their totals trimmed. Teams adding a premier quarterback or pass rusher get bumped up. What remains in late July is the team-specific information training camp starts to surface: depth chart decisions, injury updates, quarterback timing, the physical condition of players who had offseason surgeries.
Sharp bettors who understand the structure of win totals do not wait for preseason results. They move before camp closes the line further. Below are three measurable signals with documented track records, each applied to the current 2026 board.
Why This Market Systematically Misfires
The academic case for this approach is documented in peer-reviewed research. In 2015, economists Bill Woodland and Linda Woodland published "The National Football League Season Wins Total Betting Market: The Impact of Heuristics on Behavior" in the Southern Economic Journal (vol. 82, no. 1, pp. 38-54). Their core finding: the NFL win totals market is "highly inefficient."
The source of the inefficiency is the representativeness heuristic. Bettors expect small samples to predict the population. A team goes 14-3 and bettors treat next year as a repeat. That assumption is statistically wrong and produces mispriced lines across the board.
Woodland and Woodland identified a specific pattern: bettors overweight prior-year records. Teams with strong recent records draw overcrowded over bets. Teams with weak recent records draw overcrowded under bets. Books adjust their lines, but the adjustment is incomplete. The residual gap is where the edge lives.
Their research covered NFL data through 2013. The market has had over a decade to absorb the finding. Based on the current 2026 board, the gap has not closed.
Rodney Paul and Andrew Weinbach documented a parallel inefficiency in individual NFL game totals in a 2002 paper published in the journal Sport Management Review. Their finding: systematic forecast errors exist across NFL betting markets when bettors anchor to expected team quality. The inefficiency is not unique to any one market type. Win totals show the same pattern, with a longer feedback loop.
Signal 1: One-Score Game Record
One-score games, decided by 8 points or fewer, are the portion of the NFL schedule most sensitive to variance. A team going 11-2 in one-score games almost certainly ran better than their underlying quality across close situations. Books account for this to some degree. The public does not.
The Denver Broncos went 14-3 in the 2025 regular season. Their 11-2 record in one-score games tied the NFL single-season mark, previously set by the 2024 Kansas City Chiefs. They also posted 12 comeback wins in 2025, a new NFL record for a single season. Their head coach, Sean Payton, acknowledged publicly that this level of success in one-score games "will be hard to replicate."
The 2024 Kansas City Chiefs went 11-2 in one-score games and led the league with 15 wins. The following year, with Mahomes going down in Week 15, they finished 6-11. Some of that collapse traces directly to injury. But the regression from the one-score record was going to arrive regardless.
Denver's 2026 win total at BetMGM sits at 9.5.
Under +105 → implied probability: 48.8%
No-vig over: 53.2% / No-vig under: 46.8%
The market prices a 53% chance Denver clears 9.5 wins.
A team whose one-score game record and comeback win total both set all-time records faces a true-win-rate baseline well below 14. The historical pattern for teams with extreme one-score game records points to regression of 3-5 wins the following season. The question is whether the market at 9.5 has priced enough regression or too little.
The over sitting at -125 implies the market still leans toward replication. That lean is precisely the bias Woodland and Woodland documented. The under at +105 is where the weight of evidence points.
Signal 2: Turnover Margin Regression
Turnover margin is the most volatile team statistic in the NFL from year to year. Teams dominating the turnover battle in one season almost always see meaningful reversal the next. Since 2016, teams posting a double-digit positive turnover margin in one season have averaged a 10.3-point decrease in margin the following year, according to analysis published by Covers.com.
A swing of 10.3 in turnover margin is worth roughly 2-3 wins, depending on the team's efficiency at converting turnover differential into scoring opportunities. When a team already runs a record-breaking one-score game record, the turnover margin signal adds a second, independent layer pointing the same direction. Close-game wins and favorable turnover luck frequently travel together.
The 2024-to-2025 season gives the clearest recent template. Three teams won 14 or more regular-season games in 2024: the Kansas City Chiefs (15-2), the Detroit Lions (15-2), and the Minnesota Vikings (14-3). All three missed the 2025 playoffs or regressed sharply. The Chiefs had a specific cause in the Mahomes injury, but the Lions and Vikings declined without catastrophic injury. Their underlying metrics from 2024, including record-level one-score game performance and elevated turnover margins, had flagged both as overperformers before the regression arrived.
The 2025-to-2026 version of this template features the Broncos and the Patriots.
Signal 3: Schedule Difficulty Delta
Books set opening totals in February using projected schedule difficulty. By July, the opponent quality projections are sharper. What often does not fully update in win totals is the specific schedule delta: how much harder or easier a team's schedule gets relative to the prior year.
The New England Patriots went 14-3 in the 2025 regular season while playing the easiest schedule in the NFL. Their opponents posted a combined .391 winning percentage, the lowest opponent win rate faced by any team since the 1999 St. Louis Rams, according to CBS Sports. New England played only three regular-season games against teams that finished with winning records.
For 2026, New England draws the sixth-toughest schedule in the NFL. That shift represents the maximum negative schedule delta: the team performing best against the weakest competition now faces near-maximum structural resistance.
Their 2026 win total across the major books is 10.5. At that number, the market expects New England to lose roughly 3.5 wins compared to their 2025 record. A team moving from the softest schedule since 1999 to the sixth-toughest slate in the league faces a headwind the 3.5-win regression does not fully account for.
On the other side of the schedule ledger, the Cleveland Browns, Cincinnati Bengals, and New Orleans Saints all project the easiest 2026 slates in the NFL. Teams posting low current win totals with meaningful schedule relief are where the over bets are worth examining before camp closes the line.
The 2026 Board: Where Each Signal Points
The three signals do not always align. When they do, the case is strongest.
| Team | 2026 Total | One-Score Signal | TO Margin Signal | Schedule Signal | Direction |
|---|---|---|---|---|---|
| Denver Broncos | 9.5 | 11-2 (NFL record) → UNDER | High-variance → UNDER | Neutral (.514 SOS) | UNDER |
| New England Patriots | 10.5 | Strong but not record-level | High-variance → UNDER | 6th-toughest (was easiest) → UNDER | UNDER |
| Kansas City Chiefs | 10.5 | Prior 11-2 record was 2024 | Mahomes injury distorts TO stats | Neutral | COMPLEX |
| Seattle Seahawks | 10.5 | Defending champion headwind | Neutral | Standard champion rotation → harder | LEAN UNDER |
| New Orleans Saints | Low | N/A | N/A | Easiest schedule in NFL → OVER | LEAN OVER |
The Counter-Case: When Bounce-Back Is Real
The Woodland and Woodland paper applies in both directions. Teams coming off unusually weak seasons have their totals set too low, because public bettors anchor to last year's results just as strongly as they do for recent winners. Kansas City went 6-11 after Mahomes tore his left ACL and LCL in Week 15 against the Chargers. Their 2026 total is 10.5, reflecting a projected 4.5-win improvement.
In this case, the public bet on the over may be directionally correct.
The distinction matters. Denver and New England overperformed their underlying quality in 2025 through mechanisms that do not sustain: extreme one-score game variance, the softest schedule in modern NFL history. Their regression is a reversion to true baseline.
Kansas City underperformed their underlying quality in 2025 because their franchise quarterback had season-ending surgery in Week 15 against the Chargers. A team going 15-2 in 2024 and then 6-11 in 2025 after losing Patrick Mahomes mid-season is not a team that got worse. An ACL and LCL tear is an identifiable, correctable event. A healthy Mahomes leading the same organizational structure that produced a 15-2 season two years prior is a plausible path to 10-11 wins in 2026.
The lesson: regression signals require understanding why a team's record diverged from expectations, not only that it diverged. Variance in close games is repeatable as a regression predictor. A season-ending quarterback injury is not.
What to Watch Before Camp Closes
Training camp depth chart news moves win total lines within hours. When a starter goes down in a padded practice, the total trims half a point before evening. When a backup quarterback outperforms expectations, books respond.
Preseason game results, on the other hand, carry almost no signal for win totals. Starters play one quarter at most in the early weeks of the preseason. The Hall of Fame Game on August 6 features the Arizona Cardinals and Carolina Panthers almost exclusively with backups. Do not adjust win total positions based on preseason results. Adjust them based on the practice reports that accompany those preseason games.
That is where the information arrives fastest: practice injury reports, depth chart updates, and the specific physical condition of players returning from offseason surgeries. For Denver, the key update is Bo Nix's status after his knee cleanup procedure in the spring. His timetable was reset to late July, and any setback changes the 9.5 line immediately.
For the Patriots, the signal to watch is offensive line depth, which was the difference between their 14-3 pace and the six-loss pace their underlying schedule metrics suggested.
Win total markets are live now at DraftKings, BetMGM, Caesars, and FanDuel. Vig levels differ by book. The Broncos under at +105 on BetMGM reflects better no-vig math than the same bet at even money elsewhere. Shop the number before placing the bet.
The three signals above are structural. They do not expire when the season starts. But the prices will be different by then, because training camp closes the window every year.
Sources
- Woodland, Bill M. and Woodland, Linda M. "The National Football League Season Wins Total Betting Market: The Impact of Heuristics on Behavior." Southern Economic Journal, vol. 82, no. 1, July 2015, pp. 38-54. academia.edu
- Paul, Rodney J. and Weinbach, Andrew P. "Market Efficiency and a Profitable Betting Rule: Evidence From Totals on Professional Football." Journal of Sport Management, Sage Publications, 2002. journals.sagepub.com
- Denver Broncos 2025 season results, StatMuse and Pro-Football-Reference. 2025 Denver Broncos season, Wikipedia
- Denver Broncos 2026 win total: 9.5, Over -125, Under +105. BetMGM sportsbook. BetMGM blog
- New England Patriots 2025 strength of schedule: .391 opponent winning percentage, lowest since 1999 Rams. CBS Sports. cbssports.com
- Kansas City Chiefs 2025: 6-11 regular season, Mahomes ACL/LCL in Week 15. ESPN. espn.com
- 2026 NFL training camp schedule: all 32 teams report by July 28. Hall of Fame Game August 6. NFL.com. nfl.com
- Turnover margin regression: double-digit positive TO margin averaged 10.3-point decrease the following year (since 2016). Covers.com. covers.com
- 2026 NFL schedule difficulty: Bears, Dolphins, Packers toughest. Browns, Bengals, Saints easiest. FOX Sports. foxsports.com
- Sean Payton on Broncos one-score game success: "will be hard to replicate." Yahoo Sports. sports.yahoo.com
- Broncos 12 comeback wins: NFL record for a single season, 2025. AToZ Sports. atozsports.com
- 2024 regression precedent: Chiefs (15-2), Lions (15-2), Vikings (14-3) all missed or regressed in 2025. Yahoo Sports. sports.yahoo.com
- Patriots 2026 schedule sixth-toughest, per CBS Sports regression candidates analysis. cbssports.com