Expected Value in Betting, Explained (with a Calculator Walkthrough)

Expected value (EV) is the average amount a bet would return per attempt if you could place it many times over. A positive EV means the bet is expected to make money in the long run; a negative EV means it is expected to lose. EV is the single number that separates a disciplined bettor's decisions from guesswork — and it is built from one short formula anyone can work by hand.

This is educational maths, not betting advice. A positive EV does not promise a winning bet, and no calculation removes the risk to your stake.

The Formula

EV = (probability of winning × profit if it wins) − (probability of losing × amount staked)

That is the whole thing. Three inputs:

• Probability of winning — your own estimate that the bet lands, written as a decimal (55% = 0.55).
• Profit if it wins — for decimal odds, that is stake × (odds − 1). A £10 bet at 2.50 wins £15 profit.
• Amount staked — what you put at risk and lose if the bet fails. The probability of losing is simply 1 − (probability of winning).

Multiply the upside by its chance, subtract the downside by its chance, and you have the bet's expected value in pounds.

Why Expected Value Beats "Gut Feel"

A bet can feel good and still be a bad bet. EV forces the two things that actually matter into the same sum: how likely the outcome is and how much you are paid for it. A 90% shot at short odds and a 30% shot at long odds can have identical EV — or wildly different EV — and only the calculation tells you which. Feelings track the *chance* of winning; EV tracks whether the *price* pays you enough for that chance. Long-term results follow the price, not the feeling.

Crucially, EV is not a promise about the next bet. It is an average over many similar bets. A +EV bet still loses regularly. The value is in repeating good decisions, not in any single outcome.

Calculator Walkthrough: A Positive-EV Bet

Work through a concrete case, one line at a time.

Suppose you have analysed a match and rate a team's win at 50%. The best available price is 2.20. Your stake is £20.

1. Probability of winning: 0.50
2. Probability of losing: 1 − 0.50 = 0.50
3. Profit if it wins: £20 × (2.20 − 1) = £20 × 1.20 = £24
4. Plug in: EV = (0.50 × £24) − (0.50 × £20)
5. Solve: EV = £12.00 − £10.00 = +£2.00

The bet has a positive expected value of £2.00. As a share of stake, that is +10% (£2 ÷ £20). Interpreted honestly: *if* your 50% estimate is accurate, placing this same bet many times would return an average of £2 profit per £20 staked. Any one attempt could still lose the full £20.

Calculator Walkthrough: A Negative-EV Bet

Now the more common case — the bet that looks tempting but should be passed.

Same team, same 50% estimate, but this time the price has shortened to 1.80, stake £20.

1. Probability of winning: 0.50
2. Probability of losing: 0.50
3. Profit if it wins: £20 × (1.80 − 1) = £20 × 0.80 = £16
4. Plug in: EV = (0.50 × £16) − (0.50 × £20)
5. Solve: EV = £8.00 − £10.00 = −£2.00

Identical opinion on the match, opposite conclusion — because the price no longer pays enough for the risk. The disciplined action is to pass. This is why the price, not the pick, is the object of the exercise. The same logic underpins finding value in the first place; see how to find value bets.

The Break-Even Point

There is a fast shortcut for deciding whether a price *can* be value at all: the break-even probability.

Break-even probability = 1 ÷ decimal odds

This is the same figure as implied probability. At odds of 2.20 the break-even is 1 ÷ 2.20 = 45.5%. So any genuine estimate above 45.5% makes the bet positive EV, and any estimate below it makes the bet negative EV. Memorising this turns EV into a two-second check: *is my probability higher than 1 ÷ the odds?* If yes, there is value; if no, pass. The full mechanics of odds and implied probability are in how to read sports betting odds.

EV Per Bet vs EV Over a Season

A single +£2.00 EV bet is trivial. The point of EV is what it does repeated across a season. If you place 500 bets a year and each carries an average +£2.00 expected value, the expected total is +£1,000 — *on average, if your estimates hold up*. Two honest cautions sit on top of that:

• The "if" is enormous. Real long-run EV depends on your probability estimates actually being accurate. Overestimate your edge and the entire projection is fiction.
• Variance dwarfs the signal in the short term. Across 500 bets the results still swing wildly around the average. Losing months are normal even with a real edge. Anyone promising smooth, steady returns is describing something that does not exist.

Turning EV Into Stake Size: The Kelly Criterion

EV tells you *whether* to bet. It does not tell you *how much*. One well-known answer is the Kelly criterion, which sizes each stake in proportion to the edge:

f = (b × p − q) ÷ b

where b = decimal odds − 1, p = your win probability, and q = 1 − p. The output f is the fraction of your bankroll to stake.

Using the first example (odds 2.20 so b = 1.20, p = 0.50, q = 0.50):

• f = (1.20 × 0.50 − 0.50) ÷ 1.20 = (0.60 − 0.50) ÷ 1.20 = 0.10 ÷ 1.20 = 0.083, i.e. about 8.3% of bankroll.

Two things practitioners stress. First, if Kelly returns a negative number, the bet is negative EV and the formula is telling you not to bet at all. Second, *full* Kelly is aggressive and produces large swings, so most disciplined bettors use fractional Kelly — a quarter or a half of the figure — to smooth the ride. In the example above, half-Kelly would stake roughly 4% rather than 8.3%. Staking discipline is covered in depth in bankroll management, and the hands-on arithmetic is in how to calculate the value in a bet.

Where the Probability Comes From

Every EV calculation stands or falls on the probability you feed it. That estimate can come from your own analysis, from base rates, or from a model — including our Oracle match predictor, which outputs a probability you can drop straight into the formula. Whatever the source, treat the number sceptically: a model that is confident is still only estimating, and a confident wrong number produces a confident wrong EV. How these models reach a probability, and where they fail, is explained in how AI predicts sports results.

Responsible Gambling

Expected value is a tool for making less-bad decisions, not a route to guaranteed profit — a positive EV can lose, and your stake is always at risk. Only bet money you can afford to lose, set deposit and time limits first, and treat betting as entertainment rather than income. LokeNessiSport is an independent information platform for adults aged 18 and over and does not accept bets; whether online betting is available and lawful depends on where you live, so check your local rules and eligibility before registering with any operator. If gambling is causing harm, free confidential support is available at BeGambleAware.org and in our responsible gambling resources.