Is Matched Betting Still Profitable? The Honest Version

There is a reason this question is hard to get a straight answer to. Almost every site that answers it has something attached to the answer: a subscription, an odds-matching tool, a referral arrangement with an operator, a members’ forum with a monthly fee. None of that makes those sites wrong. It does mean their conclusion was decided before the analysis started.

This page is no exception. Elsewhere on this site there are links to matched betting services that pay a commission if someone subscribes through them, and that is stated wherever those links appear. So the verdict below should not be taken on trust, from us or from anyone else. What can be offered instead is the arithmetic: every figure here is worked through in full, from formulas that can be checked against your own numbers. A conclusion you can verify yourself is worth more than an assurance about who is paying for it.

So: is matched betting still profitable? Yes, in a narrow and shrinking sense. It works less well than it did five years ago, it does not scale, and it is not a substitute for an income. The rest of this piece explains the mechanism so that verdict can be checked rather than believed.

The mechanism in one paragraph

A promotional credit at a bookmaker has a value that can be partially converted into cash by taking the opposite position on a betting exchange. A selection is backed with the bookmaker and the same selection is laid on the exchange — Betfair Exchange being the deepest market for this in most sports. If the sizes are set correctly, the outcome of the event stops mattering: whatever happens, the combined position settles at roughly the same number. The profit, where it exists, does not come from predicting anything. It comes from the promotional credit itself, minus the friction of getting it out.

That friction is where most of the honest discussion lives.

Why a credited stake is worth less than its face value

The critical distinction is whether the credited stake is returned in the winnings. Most promotional credits are not: if a credit is used and the selection wins, the returns are calculated on the profit portion only. That single design choice is what caps the value of the whole exercise.

Take a credit of size F, used at back odds B, laid at exchange odds L, with exchange commission c on net winnings. The lay stake that equalises both outcomes is:

l = F(B − 1) / (L − c)

and the cash retained is l × (1 − c).

Run the numbers at odds of 3.0, laid at 3.1, with 2% commission: the lay stake comes to about 0.649F, and the retained cash is about 0.636F. Just under 64% of the face value.

At odds of 6.0, laid at 6.2, the same arithmetic gives roughly 79%. At odds of 10.0, laid at 10.5, roughly 84%.

Two things fall out of that. First, the face value printed on a promotional credit is not the value of the credit — a substantial slice never becomes cash, because the stake is never returned. Second, retention improves as the odds rise, which is why longer-odds selections are used, and why the sensible ceiling for a stake-not-returned credit sits somewhere in the high seventies to mid eighties as a percentage, before anything else is deducted.

The “anything else” matters. Longer odds mean wider spreads between back and lay prices, thinner exchange liquidity, and larger lay liabilities tying up capital. The theoretical retention and the achieved retention are not the same number.

Where the qualifying loss comes from

Promotional credits are usually conditional on a prior settled bet with the operator’s own money. That bet also gets covered on the exchange, and covering it costs something. The cost is the spread.

Back odds B, lay odds L, commission c, stake s. The equalising lay stake is:

l = sB / (L − c)

At back odds of 3.0 laid at 3.1 with 2% commission, the lay stake is about 0.974s and both outcomes settle at roughly −0.0455s — a loss of about 4.6% of the stake. Not a disaster, but not nothing. That loss is deducted from whatever the credit eventually retains.

So the full arithmetic of a single cycle is: retained value of the credit, minus the qualifying loss, minus commission already accounted for, minus the value of the time spent, minus the cost of having capital immobilised on the exchange and at the bookmaker while positions settle.

When the spread on the qualifying market is tight and the credit can be worked at long odds, the net is positive. When the spread is wide — an illiquid market, a short-notice event, a sport with thin exchange volume — the net can approach zero. Occasionally a market moves between placing the back bet and placing the lay bet, and the result is a genuine loss on that cycle. Anyone who has done this for a while has a list of those.

The structural problem: the supply is controlled by the counterparty

Everything above is mechanics. The mechanics have not changed much. What has changed is the supply.

Matched betting has no independent source of return. It does not extract value from other bettors, or from a mispricing that the market will eventually correct. It extracts value from a marketing budget, and the marketing budget belongs to the operator. The number of promotional credits available, their size, the odds restrictions attached to them, the minimum stake requirements, the rollover conditions, whether the stake is returned, whether cashed-out bets qualify — every one of those parameters is set by the party paying out.

That is a structurally weak position to be in. Operators have spent the last several years narrowing these parameters, for two reasons that have nothing to do with any individual customer: acquisition offers have become a smaller and more tightly targeted part of their marketing, and regulatory pressure across several jurisdictions has restricted how promotions can be constructed and advertised. The direction of travel has been consistent. There is no mechanism by which it reverses.

Account lifetime is finite

The second structural constraint is detection. Operators identify accounts whose betting patterns show no exposure to risk. The signals are not subtle: bets sized to the penny against exchange liquidity, selections clustered at the odds bands that maximise retention, activity that appears only when a promotion is live, no losing streaks, no discretionary betting at all.

The consequence is restriction — reduced stake limits, exclusion from promotions, or closure. It is not a punishment and there is no appeal worth making; the terms permit it. The practical effect is that every account has a lifetime, measured in weeks or months rather than years, and once it ends the promotional supply from that account ends with it.

This produces the shape that people who have done it recognise. The early period is the productive one, when the backlog of available introductory promotions is being worked through. What follows is a much thinner stream of ongoing promotions, from a shrinking set of accounts, at a much lower hourly return. The productive window per person is bounded, and it is getting shorter as detection improves.

Why it does not scale

Three hard limits.

Identity. One person can hold one account per operator. There is a finite number of operators. Attempting to work around this means opening accounts in other people’s names or misrepresenting identity, which is fraud, voids all balances, and can end in a criminal matter. That is the end of that avenue, and any resource that treats it as a grey area is not worth reading.

Capital. Lay positions require liability to be held on the exchange, and stakes sit with bookmakers until events settle. The working capital requirement is real and returns nothing while it waits. Expressing returns as a percentage of turnover flatters them; expressed as a return on capital and hours, the picture is more sober.

Time. Each cycle requires finding the market, checking the spread, sizing both legs, placing both legs before the price moves, and recording it. Errors happen — wrong selection, wrong lay size, a leg that does not get matched. The realistic hourly figure, once the introductory backlog is exhausted, is modest, unglamorous, and does not improve with practice beyond a certain point, because the constraint is the supply of promotions rather than the operator’s skill.

The honest conclusion

It still works. The arithmetic is sound and nobody has repealed it. A person with time, patience, working capital and good record-keeping can convert promotional credits into cash at something like 60–85% of face value, less qualifying losses.

It works less well than it did. Fewer promotions, tighter terms, faster restriction.

It does not scale, for reasons that are structural rather than a matter of effort or technique.

And it is not an income. It is a finite pot of value, distributed across a finite number of accounts, that gets used up. Treating it as a salary substitute means eventually running out of supply and facing the temptation to make up the shortfall with actual speculation — which is a different activity with a negative expected return, and is where a meaningful number of people who started with matched betting end up.

One last point, offered without pressing it: when weighing any answer to this question, including this one, it is worth establishing what the answering party earns if the reader agrees. Subscription sites, tool providers and affiliate publishers all have an interest in the answer being an enthusiastic yes. That interest does not automatically make them dishonest. It does mean the incentive is worth pricing in.

---

This article is informational and analytical in nature. It does not constitute betting advice, a recommendation to gamble, or an invitation to open an account with any operator. Gambling can be addictive and can cause financial harm. Strictly 18+; support services are available in most jurisdictions for anyone concerned about their gambling.