Arbitrage Betting Explained: A Practical Guide For 2026

Best Leagues and Markets

Some leagues are simply more ‘arb-friendly’ than others, mostly because pricing behaviour differs.

The ideal league is usually a high-liquidity one (it’s popular). This means there are multiple bookmakers reacting to the same information at slightly different speeds.

The Premier League and NBA tend to offer smaller but cleaner discrepancies. Tennis is mechanically tidy because it’s two-way. Cheltenham and other major horse racing meetings are messier, with more runners and more maths, but occasionally wider price gaps.

Calculating Stakes & Guaranteed Profit

If you’ve got as far as spotting the arb, then you’re already doing great, but this is the part where most people slip up. Allocating your stake correctly is what really makes arbitrage betting work.

For a two-way arb, this is my framework:

1. Calculate your implied probabilities and confirm the total is less than 1
2. Decide your total stake
3. Allocate your stake proportionally using the formula below

Stake A = (Total Stake × Odds B) / (Odds A + Odds B)
Stake B = Total Stake − Stake A

Example:

£50 on Team A at 2.20
£60 on Team B at 1.95

Team A wins = £110
Team B wins = £117

Profit depends on allocation accuracy. If those stakes weren’t proportionally calculated, you’re exposed on one side.

Three-way markets follow the same logic, but you divide proportionally using each outcome’s implied probability share, like this:

Check if the arb exists (1X2 market)

Pick your three prices (Home / Draw / Away) and calculate implied probability:

• Home: 2.80
• Draw: 3.60
• Away: 3.00

Implied probabilities:

• 1 / 2.80 = 0.3571
• 1 / 3.60 = 0.2778
• 1 / 3.00 = 0.3333

Total = 0.9682 (96.82%)

Because that’s under 1, a three-way arb exists.

 Calculate the target “equal return” payout

Payout = Total Stake / (1/Odds 1) + (1/Odds 2) + (1/Odds 3)

If the total stake is £100:

Payout = 100 / 0.968253968… = £103.28 (rounded)

That means you’re aiming for ~£103.28 return, whichever outcome lands.

Allocate stakes per outcome

Stake on outcome i = Payout / Odds i

• Home stake = 103.28 / 2.80 = £36.89
• Draw stake = 103.28 / 3.60 = £28.69
• Away stake = 103.28 / 3.00 = £34.43

Total staked: £100.01 (rounding drift is normal)

Worst-case return: £103.28
Worst-case coverage: ~£3.27 before any fees/limits/void weirdness.

Tiny-but-important habit: after rounding, I always sanity-check the lowest return minus total stake. If it’s tight, I’ll shave a penny or two off the ‘oversized’ leg to keep the worst-case side positive.

Multi-leg arbs (rare, but they exist) require proportional distribution across all outcomes based on implied probability weights. These are firmly in spreadsheet territory, not mental maths!

Settlement Rules

Settlement is where your theoretical profit meets reality.

A full arb assumes both legs are fully accepted and settled normally. But partial bets (annoyingly) happen. If one side is accepted at £100 and the other at £60, you’re exposed unless you rebalance quickly.

Voids are another consideration. If one side of a two-way arb is voided due to a rule interpretation, your coverage disappears.

Example: £50 at 2.20 and £52 at 2.05. If the 2.05 leg is voided, you’re left with £50 at 2.20 exposed. That’s no longer an arb, it’s a standard bet.

Live markets add timing risk and aren’t a good idea for a beginner (probably not for a pro to be honest!) Pre-match, you have breathing room. I always double-check rules before placing both sides.

Here’s how it works in practice:

Actual Odds:
Home: 2.40
Away: 1.80

Implied Probability Less Than Zero?

1 / 2.40 = 0.4167
1 / 1.80 = 0.5556
Total = 0.9723

£100 Stake Distributed Proportionally:

Outcome        Odds        Stake        Return
Home              2.40         £42.85        £102.84
Away               1.80         £57.14         £102.85

(Adjusted proportionally for equal return target in spreadsheet.)

The numbers tell me whether it works. My experience tells me whether it’s worth executing.

And sometimes, in fact often, the correct decision is simply: pass.

Arbitrage Betting Match Patterns

Over time, I’ve noticed arbs tend to appear in fairly predictable situations. There’s not a constant flood of them, but enough to notice a pattern.

• Early lines before market correction

This is the classic one. A bookmaker posts early prices on a Premier League fixture, money comes in elsewhere, and one book is slower to adjust. I’ll play these if liquidity is solid and limits aren’t restrictive. I’ll pass if it’s a niche league with £20 caps.

• Live in-play discrepancies

In-play can throw up short windows when one site suspends and reopens slightly off-market. I’ll only touch these if I’ve already calculated the range beforehand. If I’m scrambling mid-point, it’s far too easy to make a mistake, so I’ll leave it.

• Multi-bookmaker promotions

Occasionally, boosted prices, Premier League bet offers, or enhanced odds create temporary distortions. I’ll consider it if the boosted leg doesn’t distort the maths elsewhere. If it relies on awkward rollover terms or unclear settlement rules, then I’ll pass.

The key isn’t chasing every type. It’s knowing which patterns suit your process and acting only when you’ve run through every stage of the checklist, and you’re confident in the maths.

Arbitrage vs Other Markets

Arbitrage sits in a very different category from most betting markets, because the result doesn’t actually matter at all. You’re backing a pricing problem, not an outcome.

Singles / System Betting

With singles or systems, you’re backing your opinion. You accept variance in exchange for upside. Arbitrage removes opinion entirely.

Example: Instead of backing Arsenal outright at 2.10, you’re distributing stakes across outcomes to remove exposure. Far less upside, far less volatility.

Accumulators

Accas rely on multiple outcomes landing. One leg fails, and the whole thing collapses. Arbitrage is the opposite; it’s designed so that any outcome settles profitably if calculated correctly. You sacrifice potential big payouts for small, controlled returns.

Spread or Totals Markets

Occasionally, arbs appear in spreads when one site moves from say -6.5 to -7.5 and another lags. If the implied probability gap is genuine, it’s playable. But spreads move quickly, so the window can close in seconds.

Each has its place. Arbitrage is about coverage, while pretty much everything else is about prediction.

When I Avoid Arbitrage Betting Completely

There are plenty of situations where I simply don’t get involved.

• Extreme line volatility: If prices are swinging every few seconds, there’s every chance of placing the first half of the bet only for the price to leap and me to be left with a single.
• Low liquidity & tiny percentage gaps: A 0.5% theoretical edge disappears quickly with rounding or stake limits.
• Restrictions: Bookmakers do not like arbitrage bettors. If one firm is already limiting me heavily, the second leg may not be placeable at the scale I need.

Example: Two-way market shows 0.99 implied probability. Looks close. But one side caps me at £15 and the other at £200. Sometimes discipline simply means walking away, and in this kind of bet, you’d be having two singles, not an arb.