How Sportsbook Margin Works

Margin is the amount by which a market's implied probabilities exceed one hundred percent, built into prices so the book expects to profit regardless of outcome. It is calculated by summing the reciprocals of all outcomes' decimal prices and subtracting one. Hold is the share of total stakes a book actually keeps, which differs from margin.

Calculating margin

Take each outcome's decimal price and compute its reciprocal, which is the implied probability. Sum those across every outcome of the market. Subtract one. The remainder is the margin, usually quoted as a percentage.

For a two-way market priced at 1.91 on both sides, each implies about 52.4 percent, summing to roughly 104.7 percent, so the margin is about 4.7 percent.

Multi-way markets. The same calculation applies but must include every outcome. A three-way market omitting the draw produces a meaningless number, and markets with many outcomes, such as an outright winner, typically carry much larger margins that are spread across dozens of prices.

Margin versus hold. Margin is theoretical, assuming money is distributed exactly in proportion to implied probabilities. Hold is what the book actually keeps after settlement, which depends on where money landed and which results occurred. A book can hold more or less than its margin over any period, and the two converge only across very large volumes.

Why it exists. Margin pays for operating the market and for the risk of being wrong. A market with no margin would be a coin flip for the operator.

Removing margin from prices

To recover the market's probability estimates, divide each implied probability by their sum. This proportional method is simplest and assumes margin is loaded evenly in probability terms. Alternatives, including removing it in odds terms or fitting a margin parameter, distribute it differently and produce meaningfully different estimates for heavy favourites and long shots. Whichever you use, apply it consistently and record which it was.

Why margin varies

Market type. Main markets in major leagues carry the tightest margins because they attract the most volume and the most competition between books. Derivatives such as halves and team totals are wider. Player props are wider still, and outright winner markets with many runners are among the widest.

Liquidity and attention. Where sharp money participates heavily, a book that prices loosely gets picked off, so margins compress.

Book positioning. Operators differ deliberately. Low-margin, high-limit books accept smaller theoretical margin in exchange for volume and for the information their order flow provides. Recreational-focused books run wider margins and lower limits.

Number of outcomes. More outcomes usually means more total margin, since each price carries some.

Timing. Margins often start wider when a market opens, then tighten as the event approaches and limits rise.

Promotions. Boosted prices and bonuses can produce individual markets with negative theoretical margin, funded as customer acquisition rather than priced as markets.

Comparing like with like

Margin comparisons are only meaningful within the same market type and competition. Comparing a book's main-market margin against another's prop margin says nothing. Build a baseline per market type from your own captures before concluding that any book is wide or tight.

Why this matters for analysis

Comparing books. A book with better-looking prices may simply have a wider margin on the other side. Comparing implied probabilities after margin removal is the only fair comparison.

Evaluating a model. A model's probability must be compared against the market's margin-free probability, not against the raw price. Skipping this step is one of the most common ways a modelling project produces an apparent edge that does not exist.

Measuring closing line value. The comparison between the price you recorded and the closing price should be made on a consistent basis, with margin handled the same way at both points.

Segmenting analysis. Because margin varies so much by market type, aggregating props and main markets in one analysis mixes very different baselines.

Monitoring your own data. Computing margin per market per book continuously reveals stale prices, mismatched markets, and normalisation bugs. A market showing implausibly low or negative margin is almost always a data problem rather than an opportunity.

Comparing across sports. Typical margins differ by sport and competition, so a margin that looks wide in one league may be normal in another. Build the baseline from your own data before judging.

This page describes data and method and is not betting advice.

Margin and closing line comparisons

When comparing a recorded price against the closing price, remove margin at both points using the same method. Comparing a raw early price against a margin-free closing probability, or the reverse, produces a bias that can look like skill or like failure depending on direction.

Frequently asked questions

What is margin in betting?
The amount by which a market's implied probabilities exceed one hundred percent. It is calculated by summing the reciprocals of every outcome's decimal price and subtracting one, and it represents the book's built-in expectation of profit regardless of which outcome occurs.
What is the difference between margin and hold?
Margin is theoretical, derived from prices assuming money is distributed in proportion to implied probabilities. Hold is what the book actually keeps after settlement, which depends on where money landed and which results occurred. The two converge only across very large volumes.
Why do player props have higher margins?
Because they attract less volume and less sharp participation than main markets, are harder to price accurately, and carry more uncertainty about availability and usage. Wider margins compensate for that pricing risk, and limits on these markets are usually lower as well.
How do you remove margin from odds?
The simplest method divides each implied probability by the sum of all of them so they total one, distributing margin proportionally. Alternatives remove it in odds terms or fit a margin parameter, producing different estimates for heavy favourites and long shots. Apply one method consistently and record it.