Arbitrage Betting Across International Bookmakers: The Real Math Behind Surebets

Every so often a player discovers that two different bookmakers disagree enough about the same event that betting both sides guarantees a profit no matter who wins. This is arbitrage betting — arbing, or hunting surebets — and unlike almost everything else marketed as a "guaranteed" strategy, the core math is genuinely true. It is also thinner, more fragile, and far more operationally demanding than the pitches for arbitrage-scanning software ever admit. This guide walks through the real mechanics: how to spot one, how to size the stakes correctly, and — more importantly — why the opportunity usually closes before it pays for the tool used to find it.

The one condition that makes an arbitrage exist

Every bookmaker's odds encode the true probability of an outcome plus its own margin — the overround, expressed most cleanly in decimal odds. Convert any single bookmaker's full market to implied probabilities and the outcomes will sum to more than 100% — that gap is the house edge, and our margin calculator turns a set of prices into that percentage in seconds. An arbitrage exists in the rare moment two different bookmakers' individual prices, taken from opposite sides of the same market, sum to less than 100% between them. At that instant the two books are, unknowingly, competing against each other rather than against you.

Two-way arbitrage, worked through

Take a tennis match with two outcomes. Bookmaker A prices Player 1 to win at decimal odds of 2.10. Bookmaker B, pricing the same match independently, has Player 2 at 2.05. Convert each to implied probability:

• Bookmaker A on Player 1: 1 ÷ 2.10 = 47.62%
• Bookmaker B on Player 2: 1 ÷ 2.05 = 48.78%

Summed: 96.40%. Because the total is under 100%, backing both sides — Player 1 at A, Player 2 at B — locks in a profit regardless of who wins, provided the stakes are sized correctly.

Sizing the stakes

The rule is to make the payout identical whichever leg wins: each leg's stake is proportional to its own implied probability divided by the combined probability. For a total outlay of €1,000:

• Stake on Player 1 (A): €1,000 × (0.4762 ÷ 0.9640) = €494.00
• Stake on Player 2 (B): €1,000 × (0.4878 ÷ 0.9640) = €506.00

Check both outcomes: if Player 1 wins, the payout is €494.00 × 2.10 = €1,037.40. If Player 2 wins, €506.00 × 2.05 = €1,037.30. Both land within rounding of €1,037, against a €1,000 outlay — a guaranteed €37, or 3.7%, locked in the moment both bets are placed, whatever happens on court.

Three-way markets work the same way, with a third leg

Football's 1X2 market adds a draw. If Bookmaker A prices the home win at 2.60, Bookmaker B prices the draw at 3.60, and Bookmaker C prices the away win at 3.90, the implied probabilities are 38.46% + 27.78% + 25.64% = 91.88% — an unusually generous 8.8% arbitrage, split across three legs in the same proportional way. Three-way arbitrages of this size are rare in mainstream leagues; they show up more often in thin markets — lower-tier competitions, in-play moments right after a goal, niche props — precisely because fewer sharp bettors are watching to correct the mispricing. Our odds converter is worth keeping open while you check a live board, since the books involved rarely quote in the same format.

Why the margin is thinner than the spreadsheet shows

The 3.7% and 8.8% margins above are the number before real-world friction. Four costs eat into it every time, and for an international bettor moving between internationally licensed books they are larger than casual arbers assume.

• The moving middle. Between placing leg one and leg two, the price can shift — a few seconds is often enough at a liquid book. Miss the second leg at the quoted price and you are left with a single unhedged bet, the opposite of the point of the exercise. Software that scans and alerts exists specifically to shrink this window; it does not eliminate it.
• Currency conversion. Books licensed in different jurisdictions frequently settle in different currencies. Converting a HUF or RON bankroll into EUR to fund one leg and back again to withdraw the other routinely costs 1–3% per conversion — often larger than the arbitrage margin itself. An 8.8% three-way surebet can become a 2–3% one after two currency conversions; a 3.7% two-way surebet can go negative. The FX mechanics are the same ones we cover for ordinary multi-currency play in multi-currency bankroll management.
• Stake limiting. Bookmakers monitor for exactly this betting pattern — mirrored stakes across correlated markets, placed within seconds of each other, on accounts with unusually low variance in results. Once flagged, the standard response is not a ban but a stake limit: the maximum bet on that book quietly drops to a few euros, which kills the strategy's earning potential without technically closing the account. This practice, permitted under most operators' terms, is what ends most arbers' run on a given book — not fraud enforcement.
• Withdrawal and verification lag. Arbitrage capital needs to move fast between books to catch the next opportunity. A withdrawal stuck in verification for days is capital sitting idle rather than working — the same friction detailed in KYC and withdrawal friction on an international licence, and it bites arbers harder than ordinary bettors because their whole model depends on capital velocity.

What a realistic arbing operation actually looks like

Professional and semi-professional arbers do not work off two open browser tabs. The realistic version of this activity involves accounts at many books — often a dozen or more, because any single book's stake limit caps that book's contribution to near nothing within weeks of active use — plus scanning software that continuously compares live odds across books and surfaces gaps above a minimum threshold, because manual scanning cannot compete with the speed the opportunities close at. It also assumes a stake-limiting budget: the working expectation that every book eventually restricts an active arber, with the model viable only if new accounts and fresh limits keep replacing the ones that close. Finally, it requires a currency and jurisdiction map — knowing in advance which books settle in which currency and what a licence's payout speed actually looks like, because discovering it mid-arbitrage is how thin margins turn negative.

This is closer to running a small logistics operation than to a betting strategy, which is the honest answer to why arbitrage software is marketed as passive income and rarely performs as one for the person who bought it rather than the person who sold it.

Is it worth it for an occasional player?

Occasionally, yes — spotting one obvious mispricing and taking it costs nothing beyond the time to calculate the stakes correctly, and the profit is real. As a standing strategy that a casual international player runs weekly, the honest expectation is that stake limits arrive faster than the profits compound, and FX and withdrawal friction routinely erase margins that looked comfortable on a calculator. Anyone weighing whether the winnings are even worth pursuing across a tax year should also read our map of how gambling winnings are taxed by country — in a withholding jurisdiction, a 3.7% arbitrage margin can be reduced by tax before the bettor ever sees it.

A five-point pre-flight check before placing either leg

• Recalculate the implied probabilities from the actual displayed odds, not a screenshot from a few minutes ago — prices move.
• Confirm both books settle bets in a currency you already hold, or price the conversion cost into the margin before committing.
• Have both bet slips loaded and ready before placing either — the gap between leg one and leg two is where arbitrages die.
• Check each book's maximum stake at the odds quoted; a book that would cap the stake below what the math calls for turns the whole trade lopsided.
• Size conservatively on the first few bets at any new book — mirrored, tightly-timed stakes are often exactly what gets an account flagged.

FAQ

Is arbitrage betting illegal?

No. It is placing legitimate bets at legitimate prices; nothing about the wager itself breaks any law. What it typically breaches is a bookmaker's own terms of service on account usage, which is a commercial matter between player and operator, enforced through stake limits and account restrictions rather than courts.

Do bookmakers ban arbers?

Rarely outright. The standard response is quietly capping the account's maximum stake to a level that makes further arbing pointless, while leaving normal recreational betting on the account technically open. Outright closure is more common where a book's terms explicitly prohibit value betting or hedging patterns it deems abusive.

How much can a realistic arber actually earn?

Published figures from experienced practitioners cluster in the low single-digit percentage range on turnover, before the costs in this guide, and that figure erodes further as accounts get limited. It is a volume, low-margin, operationally heavy activity — closer to running a small trading desk than to finding free money.

Do I need software to do this?

Not to execute a single opportunity spotted by hand, but to find opportunities reliably and fast enough that they still exist by the time you act, yes — the gaps this guide describes typically last seconds to a few minutes at liquid books.