Bookbreaker for arbitrage bettors

You already know how to find an arb. The problem is placing it and keeping the account.

Finding the arb is the easy part. Two things decide whether arbitrage is actually profitable, and neither is the margin.

Whether the second leg lands

An arb is only an arb if both legs get on. On a feed running 5.4 seconds behind, a quote showing as 10 seconds old is really 15.4, and its chance of still being there is 84% rather than 89%. Miss the second leg and you hold a one-sided position on a game you had no opinion about.

So every arb here is ranked by margin times the chance of getting on, not by margin. That reorders the screen, and the reorder is the point.

Whether the account survives

The exact stakes for a 3.73% arb are 493.98 and 506.02. That precision is the clearest signal a risk desk reads. Rounding afterwards breaks the lock because the legs are not symmetric — solving over round stakes gives 490 and 500, still guaranteed, for 2.35 of the 37.35.

Commission is netted before anything is called an arb, which matters more than it sounds. +101 on both sides looks like a 0.50% arbitrage; against an exchange taking 2% of winnings it is -0.01%. A screen that skips the netting surfaces that every time, and it costs money on every one.

The full explanation →  · The limiting model →