Guide
The principle fits on one line: bet when the odds pay more than the risk is worth. The maths is simple. The hard part is estimating the probability better than the bookmaker. Here is the formula, a worked example, and the trap few sites point out.
In short: a value bet is a bet whose odds are higher than the true probability of the outcome justifies. There is value when probability × odds > 1. An outcome with a 50% chance offered at 2.20 gives 0.50 × 2.20 = 1.10: an expected return of +10% of the stake, over time.
A value bet is not a sure bet, nor a bet on the favourite. It is a price that is too generous. An outsider at 5.00 can be a value bet if it wins more than one time in five; a favourite at 1.20 may not be one if it wins fewer than five times in six.
Value says nothing about the result of a match. It says what you can expect on average if you replayed this bet many times at the same price, provided the probability estimate is right.
Every price states a probability, called the implied probability: implied probability = 1 ÷ odds. It is also the hit rate you need just to break even at that price.
| Odds | Implied probability |
|---|---|
| 1.50 | 66.7% |
| 2.00 | 50.0% |
| 2.50 | 40.0% |
| 3.00 | 33.3% |
| 4.00 | 25.0% |
On a match, the implied probabilities of the three outcomes always add up to more than 100%. That surplus is the bookmaker's margin. To get the market's real estimate, divide each probability by the total.
| Outcome | Odds | Implied | No margin | Fair odds |
|---|---|---|---|---|
| Home | 1.90 | 52.6% | 50.1% | 2.00 |
| Draw | 3.50 | 28.6% | 27.2% | 3.68 |
| Away | 4.20 | 23.8% | 22.7% | 4.41 |
| Total | 105.0% | 100% |
The “fair odds” are what a bookmaker taking nothing would post. Here, the draw paying 3.50 would be worth 3.68: the 5% margin sits in that gap.
This is the step that decides everything. You estimate, through your own analysis or a model, the probability of the outcome. Then:
expected value = probability × odds − 1
If you put the draw at 30% and it is offered at 3.50: 0.30 × 3.50 − 1 = +5%. If another operator offers 3.70: 0.30 × 3.70 − 1 = +11%. The same estimate is worth twice as much with the second one, which is why it pays to compare margins first.
The calculation also works backwards: the minimum odds at which a bet has value is 1 ÷ your probability. At 30%, you need more than 3.33.
Many tools flag a value bet as soon as the estimate beats the market's margin-free probability. That is a misleading shortcut, because you never bet at the fair odds: you bet at the offered odds, margin included.
Take the draw from the example again. The market puts it at 27.2% without the margin. If your estimate is 28%, you are clearly above the market. But at 3.50: 0.28 × 3.50 − 1 = −2%. The bet still loses over time, because the margin eats the edge.
Between the fair probability (27.2%) and the implied probability (28.6%) there is a zone where you “beat the market” and still lose money. The only test that matters is probability × offered odds > 1.
On MARGE, the gap shown on a match page (“Value +3%”, for example) measures the difference between our model and the market's margin-free probability. It is a signal to examine, not a promise of positive expected value, and that is why we only flag it from a 3-point gap.
That is what our investigation into tipster profitability shows: a sound method, tested on thousands of matches, can win more than half the time and still lose money.
We compare three models (odds, form, history) with the market's margin-free probability, and we publish every prediction before the match, win or lose. Gaps are leads, never certainties.
See today's matchesNo. A value bet can lose, and often does. Value only shows over a large number of bets, and only if the probability estimate is right.
Fair odds are the market price without the bookmaker's margin. A value bet is a bet whose offered odds exceed the odds matching your own estimate. Beating the fair odds is not enough, as the draw example above shows.
Very rarely. The few bettors who manage it work on low-margin markets, with proven models, and operators often limit their stakes. It is not a source of income.
Every Friday morning, before the weekend's football: an honest record of the week, misses included, and the French-licensed operator taking the least on the upcoming matches. Free, one email a week (in French), one-click unsubscribe.
MARGE is an analysis and statistics site. No bet can be placed here, and we receive no payment from operators. No analysis constitutes advice or a guarantee of any result.
Sports betting carries risks: debt, isolation, dependency. Forbidden to minors. For help in the United Kingdom call GamCare on 0808 8020 133, or in France 09 74 75 13 13 (free of charge) or visit joueurs-info-service.fr.
Home · Understanding the margin · Margin comparison · Margin calculator · Convert odds · Are tipsters profitable? · Our results