Bet types, side by side
Every bet loses part of its stake on average, but not the same part. Bets with two nearly even sides cost little, around 3.6%. The more possible outcomes, and the rarer the one you pick, the more the bookmaker keeps, and in accumulators its cuts multiply: in New Jersey, parlays kept 17.3% of what was staked, against 4.6% on American football.
At a glance
Two things describe a bet: how often it wins, and how much of each stake the bookmaker keeps on average. The second is its cost: the bigger it is, the faster a habit loses. From the cheapest to the dearest:
| Kind of bet | How often it wins | Kept of every stake |
|---|---|---|
| Asian handicap (football) | About half the time: two nearly even sides | 3.6% measured |
| Point spread (basketball) | About half the time | 4.5% at the usual price |
| Tennis match winner | Depends on the match; two sides | 7.4% measured |
| Match result 1X2 (football) | One of three outcomes | 7.8% measured |
| Horse racing, to win | One of many runners | Grows with the runners |
| Correct score, scorers, specials | Rarely: one of dozens of outcomes | 34% in one US measurement |
| 5-match accumulator | Rarely: all five must come in | 22.6% if each match keeps 5% |
"Measured" means what bettors actually lost, in studies of thousands of matches; the sources are at the end, read on 8 November 2026. The other figures follow from the rules of the bet itself.
Two-sided bets: the cheapest
In an Asian handicap the stronger side starts a goal or more behind, so that the two sides are nearly even, and a draw often returns the stake. Over 84,230 matches in 22 European leagues, bettors there lost 3.6% of what they staked on average. The point spread in basketball works the same way: at the usual price, 1.91 on both sides, it keeps 4.5%.
Over/under goals is two-sided too. Type its odds into the margin calculator to see what it keeps where you bet.
Match result and tennis: a little dearer
On 1X2, closing odds promised a loss of 6.5%, but bettors actually lost 7.8%. The gap comes from the outsiders: the bets with the smallest chances lost 17% of their stakes, the favourites only 2%. In tennis, over 58,112 matches, the loss was 7.4% where the odds suggested 5.4%.
Double chance (home or draw, and so on) and draw no bet win more often, because they cover two of the three outcomes or return the stake on a draw. We found no published measurement of what they cost. Winning more often does not mean losing less: the odds shrink to match. See any bet's real chance in one bet as a habit.
Many possible outcomes: the dearest singles
The more outcomes, the more the bookmaker keeps. In horse racing, the odds add up on average to 105% in a two-horse race and 167% in a race of thirty. Correct score, goalscorers and specials have dozens of outcomes, each of which rarely comes in; in the US, bets on a player to score a touchdown lost 34% of their stakes on average. For correct score in football we found no measurement we could read. For each way, see the each-way calculator.
Accumulators, bet builders and system bets: the cuts multiply
In an accumulator every match must come in, and what each keeps multiplies: five matches that keep 5% each keep 22.6% together, ten keep 40.1%. Official figures show it: in New Jersey, from January to September 2026, bookmakers kept 17.3% of what was staked on parlays and 4.6% on American football bets. A bet builder, several bets on the same match, gets its own price, and no regulator counts it apart. See what your own accumulator costs in the accumulator calculator.
A system bet such as a Yankee or a Lucky 15 is many accumulators at once: every double, treble and so on from your selections, and in some the singles too. The stake multiplies by the number of bets:
| System | Selections | Bets |
|---|---|---|
| Trixie | 3 | 4 |
| Patent | 3, with the singles | 7 |
| Yankee | 4 | 11 |
| Lucky 15 | 4, with the singles | 15 |
| Canadian | 5 | 26 |
| Lucky 31 | 5, with the singles | 31 |
| Heinz | 6 | 57 |
| Lucky 63 | 6, with the singles | 63 |
| Super Heinz | 7 | 120 |
| Goliath | 8 | 247 |
Live betting and cash out
Once a match starts, the margin rises: in the Premier League by 1.6 percentage points from five minutes before kick-off to five minutes after. Cash out settles a bet before it ends, at a price below what it is worth; see how much in the cash-out calculator.
Sources
- Hegarty and Whelan, "Forecasting soccer matches with betting odds: a tale of two markets", International Journal of Forecasting, 2025
- Whelan and Hegarty, "Calculating the bookmaker's margin: why bets lose more on average than you are warned", MPRA paper 116924, 2023
- New Jersey Division of Gaming Enforcement, gaming revenue results for September 2026, sports wagering completed events
- Whelan, "Field size and margins across markets", lecture slides, University College Dublin, 2026
- Whelan, Economics of Betting Markets, lecture 14 (special bets and bet builders), lecture slides, University College Dublin, 2026
- Whelan, "In-play bets and cash out", lecture slides, University College Dublin, 2026
The University College Dublin slides are not peer-reviewed; we use them only where we found nothing better. We name no bookmaker and suggest no bet.