Every betting price carries a cost that is easy to overlook: the bookmaker’s margin. Understanding how that margin works, and how it differs between markets and operators, is one of the most practical skills in sports betting analysis.
What the margin is
In a fair, two-outcome market, the implied probabilities of both outcomes would add up to exactly 100%. In practice, bookmakers shade their prices so that the total comes out higher. That excess is the margin, often called the overround. It represents the built-in advantage the operator retains across all outcomes, regardless of who wins the match.
For example, if Player A is priced at 1.80 (implied probability 55.6%) and Player B at 2.10 (implied probability 47.6%), the total is 103.2%. The extra 3.2% is the overround.
Why margins vary
Margins are not uniform across tennis. They tend to differ according to several factors:
- Event profile: high-profile matches with heavy betting volume usually carry tighter margins because of competition between operators.
- Market type: match-winner markets generally carry lower margins than more complex markets such as set betting or correct-score style markets.
- Operator model: traditional fixed-odds bookmakers and exchanges price differently, with exchanges typically charging commission rather than building margin into the price.
- Timing: prices released early, before the market has settled, can carry wider margins than those quoted close to the start of play.
| Market Type | Typical Margin Range (Illustrative) | Comment |
| Match winner, major event | Low | Heavy competition between operators |
| Match winner, lower-tier event | Moderate | Less competition, thinner information |
| Set betting | Higher | More outcomes, each carrying a share of margin |
| Total games / handicap | Moderate | Depends on how many lines are offered |
| Outright / tournament winner | Highest | Many outcomes, large cumulative overround |
(Ranges are illustrative and vary by operator and period.)
Removing the margin to estimate fair probability
To compare a market price with an independent probability estimate, the margin should be stripped out. The simplest method is proportional normalisation: divide each outcome’s implied probability by the total.
| Outcome | Odds | Implied Probability | Normalised (Margin Removed) |
| Player A | 1.80 | 55.6% | 53.9% |
| Player B | 2.10 | 47.6% | 46.1% |
| Total | 103.2% | 100.0% |
Proportional normalisation is a simplification. Some analysts argue that margin is not distributed evenly, and that favourites and longshots may carry different proportions of it. More advanced methods exist, but the proportional approach is a reasonable starting point for most purposes.
Why margin matters for long-run results
A margin of around 3% might sound small, but it compounds. Anyone placing a large number of positions at prices that include a built-in margin starts every decision with a handicap. To achieve a long-run profit, an independent probability estimate has to beat the market by more than the margin just to break even.
This is one reason that comparing prices across several operators matters. Even a modest difference in the margin between two quotes for the same outcome can add up meaningfully across hundreds of selections.
Margin on longshots and favourites
Research across many sports has suggested that the way margin is distributed can differ between favourites and outsiders, a pattern sometimes called the favourite-longshot bias. In some markets, longer prices have been observed to carry a larger share of the overround than shorter ones. Whether and how strongly this shows up in tennis varies by period and operator, so it is best treated as something to test against data rather than assume.
Exchanges and commission
Betting exchanges work differently: users offer and take prices from each other, and the platform charges commission on net winnings. The “margin” in this case is effectively the commission rate plus the spread between the best available back and lay prices. For liquid tennis markets, that combined cost can be lower than the margin built into a traditional bookmaker’s price, though thinner markets can show wide spreads that erase the advantage.
Practical implications
- Always convert odds to implied probability before comparing with a model estimate.
- Remove the margin to get a fairer baseline.
- Compare prices across operators and across exchanges where possible.
- Be aware that some markets are structurally more expensive than others, even when they look more appealing.
A worked example
Suppose a model estimates Player A at 56% to win, and the market offers 1.85 on Player A and 2.05 on Player B. The implied probabilities are 54.1% and 48.8%, which total 102.8%. After proportional normalisation, the market’s “fair” view is roughly 52.6% for Player A and 47.4% for Player B.
Against that normalised figure, the model sees a gap of about 3.4 percentage points. But the price actually available is 1.85, which implies 54.1%, so the realistic edge over the price on offer is closer to 1.9 points. The expected return per unit staked is (0.56 × 1.85) − 1, or about +3.6%, if the model’s estimate is correct. The difference between the two gaps is the margin, which is why it should never be ignored.
Summary
The bookmaker’s margin is the cost of doing business in any betting market, and it varies with event profile, market type and operator model. Understanding how to calculate it, remove it and compare it across sources is a foundational step in assessing whether any price genuinely reflects value.
