Overround is the amount above 100% obtained when the implied probabilities of every mutually exclusive outcome in a fixed-odds market are added together.
A fair market with no pricing margin would total 100%. A market totaling 105% has a 5-percentage-point overround. The extra percentage is built into the posted prices, but it is not a promise that the operator will retain exactly 5% of stakes.
From decimal odds to bookmaker percentage
For decimal odds (O_i), the displayed implied probability for outcome (i) is:
[ q_i=\frac{1}{O_i} ]
Add the implied probabilities for all (n) outcomes:
[ B=\sum_{i=1}^{n}\frac{1}{O_i} ]
where (B) is the bookmaker percentage expressed as a decimal. The overround is:
[ \Omega=(B-1)\times100% ]
Interpretation:
- (B=1.00): 100% book, no overround before other charges;
- (B>1.00): prices contain an overround;
- (B<1.00): the combined prices may create an arbitrage if every outcome is covered and all wagers can be executed under compatible rules.
Worked three-outcome market
Suppose a football market posts:
| Outcome | Decimal odds | Raw implied probability |
|---|---|---|
| Home win | 1.80 | 55.56% |
| Draw | 3.60 | 27.78% |
| Away win | 4.50 | 22.22% |
The bookmaker percentage is:
[ B=\frac{1}{1.80}+\frac{1}{3.60}+\frac{1}{4.50} ]
[ B=0.5556+0.2778+0.2222=1.0556 ]
Therefore:
[ \Omega=(1.0556-1)\times100%=5.56% ]
The three raw implied probabilities total 105.56%, even though the outcomes are mutually exclusive and one of them must occur under the market rules.
Estimating no-vig probabilities
A simple way to remove the overround is to normalize each raw implied probability:
[ p_i=\frac{q_i}{B} ]
For the example:
| Outcome | Raw implied probability | Normalized probability |
|---|---|---|
| Home win | 55.56% | 52.63% |
| Draw | 27.78% | 26.32% |
| Away win | 22.22% | 21.05% |
| Total | 105.56% | 100.00% |
These are often called no-vig, fair, or margin-removed probabilities.
The method assumes that the margin is distributed proportionally across the outcomes. Real pricing may be shaded unevenly because of customer demand, liability, information, market-making strategy, or different uncertainty around each selection. Normalization is a useful estimate, not proof of the bookmaker’s private probability model.
Why 5.56% overround is not a guaranteed 5.56% return
Overround is a pricing measure. It is not identical to the loss rate from backing every outcome, and it is not the operator’s final hold percentage.
If a bettor divides a total stake across every outcome at one bookmaker so that each outcome produces the same gross return, the return fraction is:
[ R=\frac{1}{B} ]
For (B=1.0556):
[ R=\frac{1}{1.0556}\approx94.74% ]
The equalized loss is therefore approximately:
[ 1-0.9474=5.26% ]
That figure differs from the 5.56-percentage-point overround because one measure is taken above a 100% probability base and the other is calculated as a percentage of the money staked.
Most customers do not bet every outcome in exact balancing proportions. Actual operator results depend on where the stakes fall, price changes, winning selections, limits, promotions, voids, commission, and risk management.
Overround, house edge, margin, and hold
These terms are related but should not be treated as interchangeable.
| Term | What it measures |
|---|---|
| Overround | Excess above 100% in the implied probabilities of a quoted market |
| Margin | General term for the operator’s pricing advantage; definition varies by context |
| House edge | Expected casino advantage on a wager under specified rules |
| Hold | Actual or reported revenue divided by a defined betting volume or drop measure |
A roulette wager is normally described through house edge because the wheel probabilities and payouts are fixed by the game. A sports or virtual-event market is more naturally described through overround because the operator quotes a set of prices across possible outcomes.
The UK Gambling Commission’s remote technical standards explicitly refer to communicating the “house edge, margin or over-round,” including for a virtual race, in its rules on describing the likelihood of winning.
Two-outcome markets can look simpler than they are
Suppose both sides of a two-outcome market are offered at decimal odds of 1.91.
Each side has raw implied probability:
[ \frac{1}{1.91}\approx52.36% ]
Together:
[ 52.36%+52.36%=104.71% ]
The overround is approximately 4.71%.
Because the prices are symmetrical, proportional normalization produces 50% for each side. In an uneven market, removing the overround does not tell you with certainty how the operator allocated the margin between favorites and outsiders.
Comparing markets correctly
A lower overround usually means more competitive aggregate pricing, but comparisons need the same scope.
Check that:
- all possible outcomes are included;
- the markets use the same settlement rules;
- overtime, retirements, dead heats, and voids are treated consistently;
- exchange commission is included where relevant;
- prices were captured at the same time;
- boosts or promotions are actually available at the required stake;
- each-way or multi-part bets are not being reduced to one headline number.
Overround can also vary within one event. A major match-winner market may be priced tightly while player props or exotic combinations carry much larger margins.
Connection to arbitrage
One bookmaker’s market will normally sum above 100%. An arbitrage search takes the best price for each outcome, often from different operators, and performs the same implied-probability calculation.
If the best-price total falls below 100%, the displayed prices may allow every outcome to be covered for more than the total stake. Execution and settlement risks still remain, as explained in Arbitrage Betting.
Overround measures the price of the market as a whole. It does not identify the winner, predict actual hold, or show how the margin is distributed across individual selections.