strategy · Updated 21 Jun 2026

Expected Value: The Number That Decides Everything

The whole idea in one line

Expected value (EV) is the average profit or loss a bet would make if you could repeat it forever. Positive EV, bet. Negative EV, skip — no matter how confident you feel about the outcome.

EV is a single number, and it is the only one that tells you whether a bet is worth making. Everything else — your confidence, the team's form, last week's results — is noise next to it.

Work out any bet

Enter the odds and your honest estimate of the win chance. The bar swings green when the price is in your favour and red when it is against you.

Price implies
40.0%
EV per unit
+25.0%
Return on 100
+25.0

Positive value — over many identical bets this returns +25.0% per unit. Expected profit on a 100 stake: +25.0.

The formula behind it

For decimal odds it is short:

EV per unit = (your win probability × decimal odds) − 1

Multiply your honest probability by the odds, subtract one, and you have the average return per unit staked.

Odds
2.50
Your chance
50%
EV per unit
+25%
0.50 × 2.50 − 1

A +25% EV means that, over a thousand identical bets, you would expect to be up around 250 units — even though plenty of those individual bets lose.

The normal case is negative

Confidence does not change the math. The same gut feeling can be a great bet or a terrible one depending entirely on the price.

BetOddsYour chanceEV per unitVerdict
Mispriced underdog2.5050%+25%Bet
Overbet favourite2.0045%−10%Skip
Same conviction, opposite EV. The price decides, not how sure you feel.

The second bet might win tonight. Over time it is a leak — the price simply is not paying enough for the real risk.

Why one result tells you nothing

A +25% EV bet can lose. A −10% EV bet can win. EV is a long-run average, so judge a bet by its EV when you placed it, never by whether it came in. A good bettor on a losing run still made good bets. That is exactly why Tofiko publishes its model accuracy openly, losing streaks included — variance guarantees them even when every bet was correctly priced.

The margin you have to beat

Most bets are negative EV before you even start, because bookmakers do not price outcomes at their fair probability — they add a margin. Take a two-way market priced 1.90 on each side:

Side A implies
52.6%
Side B implies
52.6%
Total
105.3%
Overround
5.3%
the bookmaker's edge

That extra 5.3% is baked into every price. So your estimate does not just have to be right — it has to be right enough to clear the margin. If the true chance is 53% but the price implies 52.6%, the margin eats the edge. You need genuine, repeatable mispricings, and those live in the games the market models least carefully. (Reading the implied probability behind any price is instant with the odds converter.)

Turning EV into a process

  1. Estimate the true probability honestly — ideally with a model, because gut estimates rarely beat the margin.
  2. Read the implied probability from the price.
  3. Compute the EV and bet only when it is clearly positive, not marginally.
  4. Size the stake so variance can't ruin you first — that is bankroll management.
  5. Track everything, losing runs included, so you can tell a real edge from a lucky streak.

The catch is always step one: EV is only as good as the probability you feed it, and most people's estimates are not better than the bookmaker's. That is the entire reason statistical models exist.

Related

Frequently asked questions

What does positive expected value mean?

A positive EV bet is one that would make money on average if you could place it many times. It doesn't guarantee the individual bet wins — it means the price is in your favour over the long run.

How do I calculate the expected value of a bet?

For decimal odds, EV per unit staked = (your true win probability × decimal odds) − 1. If that number is above zero, the bet is positive EV.

Why are most bets negative EV?

Bookmakers add a margin to their prices, so the implied probabilities of all outcomes add up to more than 100%. That built-in edge makes the average bet a loser unless your own probability estimate is good enough to overcome it.