Expected Value in Horse Racing — The Maths Behind Profitable Betting

Expected value is the concept that separates punters who think clearly about money from punters who think only about winners. Every bet you place has an expected value — a mathematical measure of what that bet is worth over the long run, accounting for both the probability of winning and the size of the payout. A bet with positive expected value (+EV) makes you money over time. A bet with negative expected value (−EV) loses you money over time. Everything else — the thrill of a close finish, the satisfaction of picking the right horse — is emotional decoration around that core arithmetic.

According to BookiesEnemyNo1, realistic ROI benchmarks for racing tipsters range from 2–5% for a decent long-term record to 5–10% for a genuinely strong edge. Those ROI figures are the outcome of consistently making +EV bets — hundreds of them, compounded over months. No one achieves a positive ROI by accident. They achieve it by understanding what expected value is, how to calculate it and, most importantly, how to identify situations where the market offers it.

This article breaks down the formula, walks through worked examples at different odds levels and explains why EV only reveals itself over a meaningful sample of bets — not from a single afternoon at the races.

The Expected Value Formula Explained

The expected value of a bet is calculated with a straightforward formula: EV equals the probability of winning multiplied by the profit if you win, minus the probability of losing multiplied by the amount you lose. Written out: EV = (P(win) × profit) − (P(lose) × stake).

Suppose you assess a horse as having a 25% chance of winning. The bookmaker offers odds of 5/1, meaning a £10 bet returns £50 profit if the horse wins. The probability of losing is 75% (1 minus 0.25), and the loss is your £10 stake. Plug in the numbers: EV = (0.25 × £50) − (0.75 × £10) = £12.50 − £7.50 = +£5.00. The expected value of that bet is positive: +£5 per £10 staked. Over many repetitions of this type of bet, you expect to make £5 for every £10 you risk. That is a 50% edge — an enormous number that you would rarely encounter in practice, but the arithmetic illustrates the principle clearly.

Now consider the opposite. You assess the same horse at 25% chance but the bookmaker offers 2/1 instead of 5/1. A £10 bet returns £20 profit if the horse wins. EV = (0.25 × £20) − (0.75 × £10) = £5.00 − £7.50 = −£2.50. The expected value is negative: you lose £2.50 per £10 staked in the long run. The horse might still win this race — 25% is not negligible — but the price does not adequately compensate you for the risk. Backing this horse is a −EV proposition, and doing it repeatedly guarantees a loss over time.

The formula’s power lies in its clarity. It reduces every betting decision to a single question: given my estimate of this horse’s probability and the price on offer, is the expected value positive or negative? If positive, the bet is worth making. If negative, it is not. The only question that matters: is this bet +EV?

Three Worked Examples — From Favourite to Longshot

Three examples at different price points show how EV operates across the spectrum of racing odds.

Example one: the short-priced favourite. Data from BetTurtle’s analysis shows that favourites win between 36% and 38% of races. Suppose a favourite is priced at 6/4 (decimal 2.5), implying a 40% chance. If the true probability is 37%, the EV on a £10 bet is: (0.37 × £15) − (0.63 × £10) = £5.55 − £6.30 = −£0.75. The market has overpriced this favourite’s chance — 40% implied versus 37% actual — and the bet is −EV. At 2/1 (implied 33%), the same 37% true probability gives: (0.37 × £20) − (0.63 × £10) = £7.40 − £6.30 = +£1.10. A small but positive edge. The price makes the difference.

Example two: the mid-priced contender. You assess a horse at 20% chance, and the bookmaker offers 6/1 (implied 14.3%). EV = (0.20 × £60) − (0.80 × £10) = £12.00 − £8.00 = +£4.00. A comfortable +EV bet. The market thinks the horse wins one in seven; you think it wins one in five. That gap generates profit over time. If the price were 4/1 (implied 20%), the EV would be zero — no edge in either direction.

Example three: the longshot. You rate a horse at 8% — roughly one win in twelve or thirteen attempts. The bookmaker offers 20/1 (implied 4.8%). EV = (0.08 × £200) − (0.92 × £10) = £16.00 − £9.20 = +£6.80. The EV is strongly positive, but the variance is punishing. This horse loses twelve out of thirteen times. The winning thirteenth must deliver enough profit to cover the twelve losses and still leave you ahead. That requires not just mathematical conviction but emotional resilience — the willingness to endure a long losing streak in pursuit of the occasional outsized payout.

All three examples share a common structure: assess the probability, calculate the EV, and act only when the number is positive. The skill lies not in the formula — anyone can learn it in five minutes — but in the probability estimate that feeds it. Getting that estimate right, race after race, is the craft that produces long-term profit.

Why One Bet Proves Nothing — EV and Sample Size

A single bet, whether it wins or loses, tells you almost nothing about its expected value. A +EV bet at 10/1 that loses is still a good bet — you would make the same decision again in identical circumstances. A −EV bet at 2/1 that wins is still a bad bet — it happened to land this time, but over hundreds of repetitions it would drain your bankroll.

This is the hardest part of EV-based thinking for most punters to accept. Results and quality are different things. A winning bet is not necessarily a good bet, and a losing bet is not necessarily a bad one. The quality of the bet is determined before the race, by the relationship between probability and price. The result is determined by what happens on the track, which includes randomness that no amount of analysis can eliminate.

The law of large numbers says that over a sufficiently large sample, actual results will converge on the expected value. For a +EV bettor, the larger the sample of bets, the more closely their actual profit approaches the theoretical profit predicted by the EV calculation. The operative phrase is “sufficiently large.” In horse racing, where win probabilities range from 5% to 50%, a meaningful sample is at least 200–500 bets. Anyone claiming to have proven their edge in 30 or 40 bets is claiming to have beaten the variance — and the mathematics of probability says that is not a credible claim.

Applying EV Thinking to Your Daily Betting

Applying EV thinking to your daily betting does not require a spreadsheet open during every race — though some punters do exactly that. At its simplest, it requires developing the habit of estimating probabilities before you look at the price.

Study the race. Form your view. Assign each horse a rough probability: this one 30%, that one 15%, the outsider 5%. Then — and only then — look at the odds. Where the market offers a price that implies a lower probability than your assessment, you have a potential +EV bet. Where it implies a higher probability, you do not. The discipline of forming your view before seeing the price prevents the market from anchoring your thinking. Most punters work backwards — they see the price first and then convince themselves the horse’s chance justifies it. That inversion is the root of most −EV betting.

Walk away from races where you cannot form a meaningful probability estimate. If you have no view, you have no edge, and any bet you place is likely to be −EV simply because the bookmaker’s margin guarantees it. Selectivity — the willingness to skip races rather than bet on them — is not a sign of timidity. It is the natural consequence of +EV thinking applied honestly. Not every race offers a +EV opportunity, and the punters who make money are the ones who bet only when the maths says they should.