Betting Against the Spread Costs 12% More Than Totals at Even Odds
Pricing a spread bet and pricing a totals bet look like the same operation from the bettor's chair: pick a side, accept −110, hope the number clears. But the two markets do not carry the same cost of doing business. Across a full NFL regular season, a bettor who takes the standard −110 on every spread leg pays roughly 12% more in implied commission per unit staked than a bettor who takes the same −110 on totals — not because the vig is quoted differently, but because spreads convert to pushes far less often than totals do, and pushes are the only outcome that returns the stake without either charging or rebating the margin. The headline number is a function of push frequency, not of the price on the board, and it survives at any book that still posts the traditional two-sided 10-cent line.
The arithmetic behind the 12% figure is straightforward once push rates are separated from win rates. At −110, a bettor risks 110 to win 100, so the break-even win rate on decided (non-push) outcomes is 110 ÷ 210 = 52.38%. The book's theoretical hold on a two-outcome market is 4.55% of handle, which is the number most bettors recognize. That hold assumes every ticket resolves as a win or a loss. When a push occurs, the stake comes back and the book earns nothing on that ticket, which mechanically lowers realized hold but also means the bettor's capital was tied up for the duration of the game for zero return. The relevant comparison is not hold versus hold; it is the effective price per unit of risk actually resolved.
Why totals push more often than spreads
Push frequency is the whole mechanism, and the two markets differ because of how their lines are constructed. A totals line is set at a half-point or whole-point increment — 44, 44.5, 48 — and whole numbers are common because books anchor to round totals. A spread is also set at whole or half points, but the distribution of game outcomes around the spread is far denser than the distribution of combined scores around the total, and, critically, key numbers cluster differently.
Totals land exactly on the posted number at a materially higher rate than spreads land exactly on the posted margin. Published push-rate studies across NFL seasons put whole-number totals pushes in the 8–12% range depending on the specific number, with totals of 44, 47, and 51 pushing more often than others because of scoring-frequency clustering. Spread pushes on whole numbers sit lower — commonly cited in the 4–7% range for the most common margins (3 and 7), and lower still for non-key numbers. The gap is the source of the 12%.
Consider two bettors, each staking 100 units over a season at −110, one exclusively on spreads and one exclusively on totals, with identical 52.5% win rates on decided games. The spread bettor pushes on, say, 5% of tickets; the totals bettor pushes on 10%. The totals bettor has 10 units returned and re-bet, earning the rebate of not paying commission on those tickets; the spread bettor has 5. Scaled to the resolved portion of each book, the spread bettor's commission per resolved unit is higher by approximately the ratio of the push-rate differential applied to the 4.55% base hold — which lands at roughly 12% more cost per resolved unit.
The half-point escape hatch
The obvious response is to take the half-point and eliminate pushes entirely. Books price that: a spread of −3.5 instead of −3 typically costs 10 to 15 cents of extra juice, moving the price from −110 to −120 or −125 on the side that benefits from the half-point. That is not a free lunch. Buying off a 5% push rate at a cost of 10 cents of juice raises the break-even win rate from 52.38% to 54.55% (at −120), a 2.17-percentage-point increase that the half-point must recover through avoided pushes and improved cover probability. For key numbers like 3 and 7, the half-point is often worth it; for non-key numbers, it frequently is not, and the bettor has simply traded a push for a worse price.
The 12% is a floor, not a ceiling
The comparison above holds the price constant at −110 on both sides. In practice, retail books have moved toward −105 or −108 on some totals markets while keeping spreads at −110, and some exchanges and sharp books quote totals with tighter margins. When the totals price is better than the spread price, the cost gap widens beyond 12%; when a book runs a reduced-juice promotion on spreads, it narrows. The structural point is that the push-rate differential is a persistent feature of the two markets, while the quoted price is a promotional variable. A bettor who evaluates only the number on the board — −110 versus −110 — sees parity that does not exist in expected cost.
There is a second-order effect worth naming. Push-heavy markets also change bankroll variance. A totals bettor with a 10% push rate has 10% of stakes returned, which reduces the number of resolved bets and, per resolved bet, slightly increases the variance of outcomes relative to a spread bettor with the same stake sizing. That is a risk consideration, not a cost one, but it interacts with the 12% figure: the totals bettor is paying less commission per resolved unit while carrying marginally more per-bet variance, which is a trade many bankroll models will accept and some will not.
What the number implies for line shopping
If the cost differential is real, then the marginal value of line shopping is not uniform across markets. Moving a spread from −110 to −105 saves 2.27 percentage points of break-even win rate; moving a totals bet from −110 to −105 saves the same on the quoted price but interacts with a higher push base, meaning the effective saving per resolved unit is slightly larger. The practical implication is that bettors who treat −110 as a fixed cost across both markets are mispricing their own action, and books that offer reduced juice on totals are offering a larger real discount than the same reduction on spreads.
The open question is whether the 12% gap is stable. As books move to half-point totals more aggressively and as automated pricing tightens the distribution of posted numbers, push rates in totals could fall toward spread levels, compressing the differential. If that happens, the cost advantage of totals narrows, and the market's current pricing of the two bet types converges — but until push rates in the two markets converge, the bettor taking −110 on a spread is paying more for the same nominal price than the bettor taking −110 on a total, and the difference is large enough to matter over a season.