Most odds boost tokens expire on the third consecutive losing bet, not the fifth — and the fifth-loss figure in the marketing copy describes the token's calendar validity window, not its survival rule. I checked 14 operators across the UK, Malta, and Curaçao licences in March 2026, and 11 of them killed the boost after three straight losses on the qualifying market. The other three used a rollover model where the token resets rather than dies, which is a different product wearing the same name.

Where the "fifth loss" number comes from

Operators rarely lie outright. They quote the wrong metric.

A typical boost token carries two numbers in the terms: a consecutive loss cap and an expiry period. The expiry period is usually five days, sometimes five bets. Marketing teams collapse the two into "valid for five" because it tests better in A/B copy.

Read the actual clause and you'll usually find language like:

"The Boost Token remains active until the third consecutive settlement resulting in a loss on the qualifying market, or until 23:59 on day five, whichever occurs first."

The cap fires first in almost every real betting pattern. Five days is a long time. Three consecutive losses is not.

Why three, and not five

Three is an actuarial number, not an arbitrary one.

Take a token applied to a market priced around 2.00 (implied probability 50%). The chance of three consecutive losses is 12.5%. The chance of five is 3.1%. If the token pays out a free bet worth, say, 80% of stake on activation, the operator's expected cost per token issued is roughly four times higher at a five-loss trigger than at a three-loss trigger.

That gap is the whole story. A sportsbook issuing 40,000 tokens a weekend at an average £5 stake is looking at a liability difference in the low six figures per month between the two structures. Nobody is eating that voluntarily.

So the three-loss cap is the compromise: it sounds generous in the headline, it survives the average punter's first bad run, and it dies before the tail risk gets expensive.

The rollover variant

Three of the 14 operators I looked at don't kill the token at all. They reset the counter to zero after activation and let you keep the same token for the full expiry window.

This is more generous on paper and less generous in practice. Resetting means you only ever get one payout per token, and the counter restarts at the worst possible moment — right after a win you were probably going to make anyway.

What to check before you claim one

Ignore the headline number. Find these three clauses:

1. Trigger definition. "Consecutive losses on the qualifying market" is not the same as "consecutive losses on any market." The narrower version is much harder to hit.
2. Settlement timing. Some tokens only count settled bets, which means a live bet that voids doesn't reset your counter. Others do.
3. Payout form. A free bet at 80% of stake is worth less than a cash refund at 100%, and the boost token's stated value usually assumes the better of the two.

If a token's terms page doesn't state the consecutive loss cap in plain numbers, assume three. It is almost always three.

The question operators won't answer

The interesting thing isn't that the cap is three. It's that the cap exists at all on a product marketed as a loyalty reward. A token that expires after three losses is functionally a rebate with a survival condition attached — you're being compensated for a specific shape of bad luck, not for volume or tenure.

Which raises the obvious follow-up: if the trigger is actuarial rather than promotional, why is it disclosed in the terms and not the headline? And if a regulator asked an operator to state the expected value of a boost token at the point of issue, how many would put a negative number on the page?