Football odds explained: decimal prices, returns and implied probability

Read a decimal price as a return multiplier

A decimal price such as 2.50 describes the total return per rand staked if the selection wins under the assumed full win settlement. Multiply the stake by that price. At 2.50, a hypothetical R20 stake produces R50 gross return when successful. The decimal already includes the returned stake; adding another R20 would count it twice.

Smarkets’ explanation of decimal odds sets out this multiplication. Keep the outcome condition beside the calculation: R50 is a possible return if the selection wins, not money earned when the price appears on screen.

Here, the arithmetic concerns an ordinary single selection with only a full win or full loss. Different settlement structures need their own calculation. Our football betting markets guide explains why identifying the selected market comes before applying a price. A number without its market label leaves part of the question unanswered.

The displays 2.5 and 2.50 represent the same decimal value. Neither means 2.5%. Reading the format correctly comes before multiplication; an extra displayed zero changes presentation, not the return multiplier.

Separate stake, gross return and net result

The stake is the amount committed to the selection. Gross return is the complete amount returned after settlement. Net result subtracts the original stake from that gross figure. These labels prevent a common misunderstanding: receiving R50 after staking R20 leaves a R30 net gain, not a R50 gain.

If the same R20 selection loses in this simplified example, gross return is R0 and net result is minus R20. The stake is not subtracted a second time from an already stated net loss. Read labels such as potential return and potential profit carefully; similar-looking amounts can answer different questions.

The table holds the fictional stake at R20 to make the distinction visible. Its winning rows describe conditional returns, while the losing row shows the separate downside. No additional deductions or charges are assumed in these teaching examples.

Scroll sideways to read the full table.

Understand what changes when the stake changes

A larger stake changes the rand exposure, not the likelihood of the football outcome. At the same fictional price of 2.50, R10 produces a possible R25 gross return and R15 net gain; R40 produces a possible R100 gross return and R60 net gain. A loss costs R10 or R40 respectively. Both sides of the monetary calculation scale together.

Changing the price is different. Keeping the hypothetical stake at R20, moving from 2.50 to 2.75 increases the possible gross return from R50 to R55. The extra R5 comes from the price difference, not from increasing the stake. This distinction helps explain why comparing possible returns without comparing stakes can mislead.

Do not treat a desired rand return as evidence about a match. Working backwards from a target amount merely rearranges a calculation; it supplies no information about which team will win. Likewise, increasing a stake after a loss does not make a later selection more likely to succeed.

Also name the base when describing an improvement. The R5 difference between possible gross returns of R50 and R55 is 10% of R50. The same R5 is about 16.67% of the corresponding R30 net gain. A percentage comparison becomes clearer when its starting amount is visible.

Convert decimal odds into implied probability

The reciprocal of a decimal price gives its implied probability: divide one by the price, then multiply by 100 to express the answer as a percentage. For the hypothetical 2.50 price, one divided by 2.50 equals 0.40, or 40%. Smarkets’ implied probability guidance confirms this conversion.

Notice the inverse relationship. A larger decimal price corresponds to a smaller implied percentage. A price of 5.00 implies 20%; a price of 1.25 implies 80%. The larger conditional payout accompanies the smaller price-derived percentage. The number 5.00 therefore does not mean a selection is five times as likely to win.

Call the result implied probability because it is extracted from a price. It is not an independently measured chance for that fixture. In particular, the conversion alone does not remove a bookmaker’s margin or reveal which assumptions produced the price. Those are separate questions.

Scroll sideways to read the full table.

Round the display after doing the calculation

Some reciprocals continue beyond two decimal places. A hypothetical price of 3.50 gives 28.571428…%, often displayed as 28.57%. Preserve the extra digits while adding several outcomes, then round the final total. Small discrepancies between a sum of rounded cells and a rounded overall total can otherwise appear to be mistakes.

Distinguish percentage points from a percentage change. Moving from an implied 40% to 50% adds ten percentage points. Relative to the starting 40%, the increase is 25%. Both descriptions can be mathematically correct, but they describe different comparisons. Writing only that probability rose by ten per cent leaves an ambiguity.

Precision also has a limit. Reporting a price-derived percentage to six decimal places does not make the underlying assessment that accurate. Keep enough digits for a sound calculation, then use a readable rounded figure. For rand examples, show the stake and the price used so that another reader can reproduce the arithmetic.

Use a complete market to calculate overround

For this example, imagine one fictional 1X2 market covering a home win, a draw or an away win over the same defined match period. Exactly one of those outcomes applies. Convert every quoted price to an implied percentage and add the three results. Using unrelated selections or mixing match periods would make the total meaningless.

The invented prices below produce a book percentage of about 103.57%. Its excess over 100% is approximately 3.57 percentage points, commonly called the overround. Smarkets’ margin calculation guide explains this complete-market method.

Keep two labels clear: 103.57% is the total book percentage, while 3.57 percentage points is the excess. Neither means the home side has a 103.57% chance. An individual outcome’s probability cannot exceed 100%; the larger total here concerns quoted prices across the whole market.

Scroll sideways to read the full table.

Do not read overround as an exact profit figure

Overround describes a relationship between quoted prices. It does not record how much money was staked on each outcome or which outcome actually occurred. Those missing quantities matter when calculating a bookmaker’s realised result. The fictional 3.57-point overround is therefore not proof that every customer loses exactly 3.57% of each stake.

Consider an invented book accepting R100 only on the home side at 2.00. A home win requires a R200 gross return, giving the book a R100 loss on those bets before other costs. A draw or away win leaves R100 from the losing stakes. The same three displayed prices can accompany these different realised results.

That deliberately unbalanced example separates price arithmetic from actual cash flows. It does not describe a bookmaker’s usual liabilities. A meaningful percentage also needs a stated denominator: total stakes, gross returns and account deposits are different quantities. Avoid comparing figures simply because each carries a percent sign.

Recognise the assumption behind removing margin

One way to make the fictional market’s percentages add to 100% is proportional normalisation. Divide each unrounded implied percentage by the unrounded total, then multiply by 100. The example becomes approximately 48.28% home, 27.59% draw and 24.14% away. Display rounding makes these three printed figures total 100.01%; the unrounded calculation totals 100%.

This operation assumes the excess is allocated proportionally across outcomes. It does not establish that the resulting percentages are the true chances. Removing a numerical excess cannot recover information about team strength or identify an incorrect assessment. Another allocation assumption can produce different adjusted percentages from exactly the same prices.

Research by Clarke, Kovalchik and Ingram on adjusting bookmaker odds for overround examines several methods. The practical lesson here is modest: a label such as margin-adjusted describes a calculation, not a verified football forecast. Always ask which method was used before interpreting an adjusted figure.

Keep probability separate from a match result

Even a well-supported probability between zero and 100% leaves uncertainty. Suppose, solely for illustration, an outcome’s actual chance were known to be 40%. That would not mean it must occur four times in the next ten matches, or that a loss makes the next selection due to win. A probability is not a timetable.

OpenStax’s probability terminology distinguishes possible outcomes and long-run frequencies. Applying that language to football requires care: different fixtures involve different teams and conditions. Ten selections with a similar displayed price are not automatically ten identical experiments with a known shared chance.

Likewise, one unexpected result does not prove that a probability was wrong. Judging the quality of forecasts requires more than checking whether a favourite won once. For the evidence behind a football opinion, read our match analysis guide; converting a price supplies arithmetic, while assessing a match requires separate evidence.

A favourite need not be more likely to win than not win. In an invented probability model assigning 45% to a home win, 30% to a draw and 25% to an away win, home is the largest single outcome. Yet the combined chance of no home win is 55%. This is a hypothetical model, not a conversion of the earlier market or a forecast for a real match.

The expected-goals guide explains a different kind of probability: the modelled quality of an individual shot, rather than a market price.

Understand what a break-even percentage assumes

For the simplified full win or full loss bet, the reciprocal price is also the theoretical break-even win probability before other deductions. At fictional odds of 2.50, a successful R20 selection gains R30 net and a losing one costs R20. Using an assumed 40% win chance gives 0.40 × R30 minus 0.60 × R20, which equals R0.

Change that assumed chance to 35% and the same calculation becomes R10.50 minus R13, or minus R2.50. The price and stake stayed fixed; the assumed probability changed. OpenStax’s expected value explanation provides the underlying probability-weighted arithmetic.

Expected value is a model average, not a settlement amount. In this example one bet returns either the R30 net gain or the R20 loss, not the average. Most importantly, the price does not prove the assumed actual chance. A confident personal estimate can still be inaccurate, so a spreadsheet result does not establish a profitable method.

Interpret a price change without inventing certainty

Imagine a fictional selection moving from 2.50 to 2.20. Its implied percentage rises from 40% to about 45.45%, while the possible gross return on R20 falls from R50 to R44. Saying the price shortened describes that change. It does not explain why it happened or guarantee the selection will win.

Smarkets’ discussion of changing odds identifies new information and the flow of money among relevant influences. A price movement alone does not reveal each influence’s contribution. Avoid turning an observed change into an unsupported story about an injury, inside information or a certain result.

Record which price a calculation refers to. A later screen showing 2.20 is not evidence that an earlier hypothetical calculation at 2.50 was arithmetically wrong. They use different inputs. When comparing prices, keep the selection and market definition the same; our market guide helps explain those labels.

When viewing BravoBet, distinguish any displayed price from the fictional examples in this guide.

Check the calculation and its limits together

A useful final check is to explain a fictional price in three separate sentences. State what the return would be if the selection won, what would be lost if it failed, and what percentage the reciprocal implies. Keeping those statements separate prevents a possible payout from quietly becoming a prediction.

For example: R20 at 2.50 could return R50 gross, giving R30 net gain; a full loss would cost R20; the price implies 40% before any attempt to adjust the complete market. None of those sentences establishes the actual chance of a home win. That remaining uncertainty belongs in the interpretation.

• Use the stake and decimal price actually specified in the example.
• Subtract the stake once when converting gross return to net result.
• Calculate overround from every outcome in one complete, consistent market.
• Label hypothetical figures and explain any margin-adjustment assumption.
• Treat a percentage as a quantity to interpret, never a promise of income.