A Patriots upset in Buffalo, and why the books were happy about it
The clock is winding down at Highmark Stadium, the 3-0 Bills are chasing, and Drake Maye, pulled from a 35-6 embarrassment against Jacksonville just seven days earlier, is closing out a 29-26 road win. For Buffalo backers it was a gut punch. For the people pricing the games, it was the single best thing that happened all day. That one result is the cleanest lesson in NFL underdog betting you’ll find: an upset that costs bettors money usually saves the bookmaker money, and sometimes saves a whole Sunday.
Maye’s line was no fluke: 22-of-37 for 269 yards, three touchdowns and one interception, plus 54 rushing yards on eight carries and only one sack. ESPN put his passer rating at 97.7 and his total QBR at 91.0, which is elite territory. The Patriots became the first team to beat the Bills in their new stadium, and in doing so they wiped out one of the most heavily parlayed teams on the board.
Zoom out and the pattern held across the week. Per OddsTrader, Week 4 underdogs went 8-4-2 against the spread with five outright upsets. The Raiders, 5-point home underdogs against Kansas City, lost 30-27 and still covered, pushing them to 4-0 ATS. In the Sunday night game, Carolina won 32-26 as a three-point underdog at Detroit, after the Lions at -3 had absorbed roughly 80% of one major book’s handle by Friday. A senior trading manager at BetMGM summed the day up as “a positive day for the book,” adding that the Titans and Dolphins covering helped and that the Bills losing “broke up many of the common big favorite parlay bets.”
Bettors weren’t shut out. Indianapolis won 30-13 in London as a -225 moneyline favorite. Derrick Henry’s first touchdown against his old team cashed, along with popular Josh Allen and James Cook end-zone props. The Rams came back to beat Philadelphia 24-20, the Giants handled Arizona 36-24, and the Bears beat the Jets 23-12. But the big-ticket results, Buffalo losing, Jacksonville beating Cincinnati 22-17, Dallas edging Houston 34-30, all landed on the house’s side of the ledger.
| Week 4 result | Betting angle | Who it helped |
|---|---|---|
| Patriots 29, Bills 26 | Road underdog upset; Bills a top parlay leg | Sportsbooks |
| Panthers 32, Lions 26 | Lions -3 held ~80% of one book’s handle | Sportsbooks |
| Chiefs 30, Raiders 27 | Raiders +5 lost the game, won the bet | Underdog backers |
| Colts 30, Commanders 13 | -225 moneyline favourite won outright | Bettors |
| Cowboys 34, Texans 30 | Favoured side on the public’s board fell short | Sportsbooks |
Point spreads explained: what that number actually does
A point spread is a handicap. The bookmaker adds points to the weaker team and subtracts them from the stronger one so that backing either side becomes close to a coin flip. Without it, almost nobody would bet the Chiefs at short moneyline odds and almost nobody would bet the Raiders to win outright, so there’d be no two-sided market to trade.
What the numbers mean
Take the Raiders at +5 against Kansas City. If you back Las Vegas, your bet wins when the Raiders win outright or lose by four or fewer. Kansas City at -5 needs to win by six or more. A six-point win? Chiefs backers cash. The actual result, 30-27, meant the Raiders lost by three and covered. That’s the detail casual fans miss: the team can lose and the ticket can still win.
If the margin lands exactly on the number, say a three-point favourite winning by three, the bet is a push and stakes are returned. That’s why you see half-point lines like -3.5 and why two of Week 4’s underdog results were recorded as ties against the spread.
Why both sides pay the juice
Spread bets are usually priced at -110 on both sides. You stake 110 to win 100. That extra 10 is the vigorish, or juice, and it’s the bookmaker’s fee for making the market. Written as implied probability, -110 is 110 รท 210 = 52.38%. Price both sides at -110 and you’ve sold a total of 104.76% of probability on a 100% event. The surplus is the margin.
The practical consequence for a bettor: you need to win about 52.4% of your -110 spread bets just to break even. Hitting 50%, exactly what you’d expect from flipping a coin, loses money over time. That gap is the house edge of sports betting, and it never switches off.
How sportsbooks make money when the book is balanced
The textbook model is simple. Suppose a book takes 110,000 on the Lions -3 and 110,000 on the Panthers +3. Whichever side wins, the book pays out 210,000 and keeps 10,000. That’s a theoretical hold of about 4.55% of total handle, collected with zero interest in the final score.
Three things follow from that math:
- A perfectly balanced market makes the result irrelevant. The bookmaker is a toll collector, not a gambler.
- Hold percentage is measured against handle, not against the money at risk, so volume matters enormously. Small margin, big turnover.
- Parlays hold far more than single bets, because the margin on each leg compounds. That’s precisely why books watch parlay exposure so closely, and why a popular leg losing is such a relief.
In the real world, balance is an aspiration rather than a state. Public money clusters on favourites, prime-time teams and household-name quarterbacks. Books rarely end a Sunday evenly split, which is exactly how a week like this one happens.
Why NFL underdog betting results can lift a sportsbook’s margin
When 80% of the handle on a game sits on one side, the bookmaker is no longer a toll collector. It has a position, and it is effectively rooting for the underdog. Carolina winning 32-26 at Detroit didn’t give the book a 4.55% theoretical hold on that game. It handed over close to the whole pool, minus the small payout owed to Panthers backers.
Parlays amplify this. Thousands of tickets that Sunday included the unbeaten Bills as a “safe” leg alongside other chalk. Those tickets are correlated: one upset kills them all at once. A Caesars football trader called Buffalo “the biggest result for the book,” noting the Patriots took down one of the most heavily bet parlay teams of the day and that the upset steadied the book through the morning slate.
This is also where sharp money and public money separate. Sharp bettors tend to take underdogs and unpopular numbers, often early, in modest-looking amounts that move prices. Public betting arrives later, in volume, on favourites and overs. When underdogs cover at an 8-4-2 clip, the small, informed side of the ledger gets paid and the large, recreational side doesn’t. Totalled across fourteen games, that’s a profitable Sunday for the industry even though plenty of individual bets won.
Be clear about the direction of the logic, though: books don’t want underdogs to win on principle. They want whichever side holds less money to win. On a week when the public loads up on a dog, a favourite blowout is the result they’re hoping for.
Setting the number, then defending it
An opening spread comes from the trading team’s power ratings, adjusted for home field, injuries, rest, weather and travel. London kickoffs, short weeks and a quarterback’s status all get priced in. The opener is a best estimate of a 50/50 line, deliberately offered with tighter limits so the market can tell the book whether it got the number right.
After that, risk management takes over. Line movement does two jobs at once:
- Repricing. If respected money hits a side immediately, the book treats that as new information and shifts the number, the same way a trading desk moves a quote.
- Rebalancing. If liability is piling on one side, moving from -3 to -3.5, or nudging the price from -110 to -120, makes the unpopular side more attractive and slows the flow.
Books also manage exposure directly: lowering limits on games where they feel outgunned, laying off risk with other operators, pricing correlated parlay legs more conservatively, and tracking total liability across a weekend rather than game by game. A bookmaker can lose a single Sunday badly. What it is engineered to do is survive variance long enough for the 4-5% theoretical margin to show up in the accounts.
The honest takeaway from Week 4 is not a strategy. Underdogs going 8-4-2 against the spread is the kind of swing that happens regularly and reverses just as often, and nobody, including the traders quoted above, saw those five upsets coming. What’s durable is the structure: the spread turns a mismatch into a near coin flip, the juice means you need roughly 52.4% to break even, and the book’s edge is baked in before a ball is kicked. Understand that and you’ll read a betting market far better, but understand too that the math points one direction over time. Bet only what you can comfortably lose, set deposit and time limits before you start, and use self-exclusion or cool-off tools if betting stops being entertainment. Support is available through national problem-gambling helplines.
