OddStream.io
CoperturaPrezziDocs
ENFRESDEITPT
Accedi
Ottieni una chiave API

Prodotto

  • Copertura
  • Prezzi
  • Docs
  • Mercati
  • Blog

Bookmaker

  • Bwin API
  • Winamax API
  • Betclic API
  • Unibet API
  • Daznbet API
  • Vbet API
  • PMU API
  • Betsson API

Sport

  • Calcio API
  • Tennis API
  • Basket API
  • Rugby API
  • Pallamano API
  • Hockey su ghiaccio API
  • Tutti gli sport →

Competizioni

  • Ligue 1 API
  • Premier League API
  • Champions League API
  • NBA API
  • Top 14 API
  • NFL API
  • Tutte le competizioni →

Confronti

  • vs The Odds API
  • vs OpticOdds
  • vs BetsAPI
  • vs Sportradar
  • vs Sportmonks
OddStream

OddStream

Powered by StatsnBet.
Il feed in tempo reale per scommettitori professionisti.

© 2026 OddStream.io. Tutti i diritti riservati.

Tutti gli articoli

Pubblicato il 10 luglio 2026 · OddStream Team

Arbitrage Betting on French Bookmakers: How It Works and Its Limits

Sure-bet math with a worked 2-way example, why arbs appear across the 8 ANJ bookmakers, the speed it takes to catch them, and the honest limits and risks.

The mechanics of a sure bet

An arbitrage, or sure bet, exists when the odds on all outcomes of an event, taken at different bookmakers, imply probabilities summing to less than 100%. Cover every outcome with the right stakes and you lock in the difference as profit, whatever happens on the pitch.

The test is simple. For a two-way market, compute 1/odds_A + 1/odds_B. If the sum is below 1, the arb exists and the guaranteed margin is 1 / sum − 1. The whole discipline is finding these sums fast and executing before they close.

A worked 2-way example

Suppose a tennis match where Betclic offers player A at 2.10 and PMU offers player B at 2.05. The implied sum is 1/2.10 + 1/2.05 = 0.4762 + 0.4878 = 0.9640, below 1, so this is an arb with a margin of 1/0.9640 − 1 = +3.7%.

Staking is proportional to implied probability. With €1,000 total: stake on A = 1000 × 0.4762 / 0.9640 = €494, stake on B = 1000 × 0.4878 / 0.9640 = €506. If A wins, the payout is 494 × 2.10 = €1,037; if B wins, 506 × 2.05 = €1,037. Either way you collect about €1,037 on €1,000 staked, a locked profit of €37.

Notice the shape of the trade: a 3.7% arb is generous, and most real arbs run 0.5% to 2%. The math is trivial; everything difficult about arbitrage is operational.

Why arbs appear between the 8 French books

The French market has eight ANJ-licensed online sports bookmakers, Winamax, Betclic, Unibet, PMU, Bwin, Vbet, DAZN Bet, and Betsson, each running its own trading desk, risk models, and update cadence. They do not move in lockstep: one book reacts to a lineup announcement in seconds, another lags by minutes, and in that window their prices straddle the fair line from opposite sides.

Structural factors widen the gaps. French books carry relatively high margins, which paradoxically creates arbs when two books shade the same market in opposite directions, one boosting a favourite to attract recreational money while another has already cut it. Promotional boosts, stale live prices after a goal, and slow secondary markets, cards, corners, player props, are the classic sources.

Arbs are most frequent exactly when prices move most: the hours before kickoff as team news lands, and during live play. That timing detail drives everything about execution.

Execution speed is the actual product

An arb is a pricing error, and bookmakers correct errors quickly, typically within seconds to a couple of minutes for meaningful gaps, and within seconds during live play. Detection latency therefore eats directly into your hit rate: if your odds data is 60 seconds old, a large share of the arbs you see are already gone, and worse, you will sometimes get one leg on and watch the second leg vanish.

That half-executed state is the real risk of slow data. One leg alone is just a naked bet, often at a bad price, and covering it late can lock in a loss instead of a profit. This is why arbitrage operations run on push-based feeds; a stream like OddStream's WebSocket, with sub-200ms median latency across all eight French books, means the sum-of-inverses check runs against prices that are current to well under a second, and its one-click betslip links cut the time from detection to a filled slip on both legs.

The limits: stakes, restrictions, and friction

Bookmakers are legally entitled to limit stakes, and they do. An account flagged as an arber, and the pattern of always betting the wrong-priced side is easy to detect, will see maximum stakes cut, sometimes to a few euros, within weeks or months. Your effective ceiling is not the arb margin but how much volume your accounts can absorb before being throttled.

Friction costs erode thin margins further. Rounding stakes to unsuspicious amounts costs a few tenths of a percent; a price moving between your calculation and your click can halve the margin; and capital is locked across multiple accounts, so a 1% arb settling in three days is a poor return on tied-up funds if opportunities are scarce. On a 0.8% arb, there is very little room for any of this to go wrong.

An honest risk assessment

Arbitrage is often marketed as risk-free, and the math of a completed arb is. But the practice carries real risks: a voided leg, a palped price cancelled by one bookmaker after the match started, leaves you fully exposed on the other side. Void rules differ between books, so a postponed match can return your stake at one bookmaker while the other keeps the bet live. Human error, wrong market, wrong line, wrong stake, is a constant tax at speed.

The realistic summary: arbitrage on the French market is a genuine but bounded opportunity, capable of grinding low-risk returns for a while, with an expiry date imposed by account restrictions. Many practitioners treat it as a phase, using the same real-time data infrastructure to graduate to value betting, where the same speed advantage is applied to +EV prices that do not require betting both sides.

OddStream API