Surebets and arbitrage: the theory, the maths and the traps

Bookmakers price independently. Occasionally two of them disagree enough that backing every outcome, at the best available price for each, guarantees a profit whatever happens. How to detect one. Convert every outcome's best available odds to implied probability and add them up. If the total is below 100%, an arbitrage exists. A worked example on a tennis match: • Site A prices Player 1 at 2.10 → implied 47.6% • Site B prices Player 2 at 2.10 → implied 47.6% • Total: 95.2% The 4.8% shortfall is your guaranteed margin. Stake correctly across both and you profit regardless of who wins. Why it exists at all: operators disagree about probabilities, they react to news at different speeds, and they shade prices to balance their own books. A site holding too much money on one player shortens the other, and that adjustment can cross another site's line. Typical size: real arbitrage opportunities usually offer 1% to 3%. Anything advertised as 10% is almost always a mistake in the data, a market misread, or a price about to be voided.

Getting the split wrong turns a guaranteed profit into a guess. The formula. For each outcome: Stake = (total investment × implied probability of that outcome) ÷ sum of all implied probabilities Worked through, with $1,000 total on the example above: • Player 1: (1000 × 0.476) ÷ 0.952 = $500 • Player 2: (1000 × 0.476) ÷ 0.952 = $500 • If Player 1 wins: $500 × 2.10 = $1,050. Profit $50 • If Player 2 wins: $500 × 2.10 = $1,050. Profit $50 A three-way example, football, with best prices of Home 3.50, Draw 3.80, Away 2.40: • Implied: 28.6% + 26.3% + 41.7% = 96.6% → a 3.4% arbitrage • On $1,000: Home $296, Draw $272, Away $432 • Every outcome returns roughly $1,036 The practical detail people miss: rounding. Bookmakers accept stakes in whole units, and rounding a $296.30 stake to $296 shifts the profit slightly between outcomes. On a 1% arbitrage, careless rounding can erase the edge entirely.

Arbitrage is guaranteed on paper. The paper does not include any of this. 1. Account limitation. This is the decisive one. Operators identify arbitrage patterns quickly — unusual stake sizes, betting only at the best price, betting immediately after a line moves. The response is to cut your maximum stake to a few dollars, or to close the account. Most arbitrage careers end here, not in a losing bet. 2. The price moves before you complete both legs. You place the first bet, and the second site has already shortened. Now you hold a one-sided position on a match you had no opinion about. 3. Bet rejection and partial acceptance. Some operators accept a reduced stake rather than your full amount, leaving you unbalanced. 4. Palpable error rules. Every operator reserves the right to void bets placed on obviously wrong prices. The most attractive arbitrages are precisely the ones most likely to be voided — and the other leg stands. 5. Different rules between operators. A tennis retirement voided at one site and settled as a loss at another turns a guaranteed profit into a guaranteed loss. Same for extra-time rules, postponements and settlement of ambiguous markets. 6. Capital and friction. Meaningful returns require substantial money spread across many accounts, with deposits, withdrawals, currency conversion and time.

It genuinely works, and the mathematics is not in doubt. The question is whether it works for you. It can make sense if: • You have significant capital to spread across many accounts, and it can sit idle • You accept that accounts will be limited and that maintaining new ones is a permanent, ongoing job • You are meticulous. Rule differences and rounding errors eat 1% margins alive • You treat it as work, with software to find opportunities, because manual searching does not find them fast enough It does not make sense if: • You have a small bankroll. A 2% margin on $200 is $4, before any friction • You want to keep usable accounts for ordinary betting. Arbitrage and a healthy account do not coexist for long • You expect it to be passive. The opportunities are short-lived and the work is constant A more accessible relative: the same skill — comparing prices and calculating implied probability — applied to line shopping produces a smaller, slower, entirely sustainable edge without triggering account restrictions. For most people that is where this knowledge actually pays.