Credit card casino: what the phrase means after the ban

A credit card casino used to be an unremarkable entry in the British market: a licensed operator whose till took your plastic on credit and let the statement arrive at the end of the month. April 2020 closed that door. Credit card gambling has been banned for operators licensed by the Gambling Commission since that month, so the phrase now lives in two places only. One is the past tense, a description of a regulated market that no longer exists. The other is the offshore tier, where tills holding no UK licence still print the rail precisely because the rule that removed it cannot reach them. Both readings can be costed, and both are costed here.

The legal frame is precise and often garbled. Providing facilities for gambling in Great Britain without an operating licence is an offence under section 33 of the Gambling Act 2005, and inviting or advertising gambling the Act prohibits is a separate offence under section 327. Both offences sit with operators and advertisers; an individual player commits no offence by playing. The Gambling Commission is the licensing authority for Great Britain and keeps a public register of every licensee it has approved, searchable by anyone in under a minute. That register, not any site’s own footer, settles what a site may lawfully do at a British till.

Two markets therefore sit side by side, and they are not selling the same product. The licensed market funds its tills from money you actually hold, wrapped in protections that exist in law rather than in goodwill. The offshore tier holds no UK licence, answers to no British regulator, and in places still accepts the credit rail the licensed market lost in 2020. Describing that tier honestly means pricing what a reader gives up by funding play there, not mapping a way in. This page does the first and refuses the second, because a protection described as an obstacle has already been misdescribed.

The method is costing, and it runs on a short list of inputs. Borrow a stake on a card, play it, and carry whatever remains unpaid. Two meters then run at once. The interest meter adds three percent of the unpaid balance a month on this page’s assumptions, simple and non-compounding. The play meter takes its cut through a blended three percent house edge, with the stake played through twice. Each assumption is printed where it is used, and each can be swapped for your own numbers without breaking the arithmetic. The sections below run seven borrowed stakes through the model, draw the interest curve, and then widen the lens to what an offshore rail costs beyond the interest line.

Who is the costing for? For the reader who has seen an offshore till advertise a credit option and wondered what the catch is priced at, and for the reader deciding whether the protections attached to a licence are worth the smaller headline that usually travels with them. It is not written for anyone hunting a workaround. If you have excluded yourself through GAMSTOP, the offshore tier sits outside that exclusion, which is a reason to close the tab rather than read on, and the care section near the end of this page says so plainly.

Why borrowing to gamble costs more than it looks

Start with what the ban actually changed, because the mechanics matter. From April 2020, operators licensed by the Gambling Commission have been barred from accepting credit cards for gambling. A licensed till is funded from money you hold: a debit card, a bank transfer, a prepaid voucher, an e-wallet fed from your own balance. The credit version of the same plastic, the one that turns a loss into a bill arriving next month, cannot be used there at all.

The reasoning behind the ban is worth restating, because it is the reasoning this whole page tests. Spending to gamble and borrowing to gamble are different acts wearing the same clothes. A debit transaction can only lose what you have. A credit transaction can lose more than you have, silently, and convert a bad evening into a debt that grows at the card’s rate while you sleep. The ban is a consumer protection built on that difference, and it sits alongside the rest of the protective architecture: deposit limits, reality checks and account history that licensees must make available, affordability checks set out in the 2023 gambling White Paper, and online slot stakes capped at £5 per spin for adults and £2 per spin for 18 to 24 year olds since September 2024.

What the ban left outside is the structural gap this page is about. The rule binds licences, not players, and not foreign tills. An operator with no UK licence may still accept credit cards, and some do. Nothing about that arrangement is a clever route around the rules; it is the absence of the protection, presented as a feature. A reader who steps through it does not outsmart the system, they exit it, and every safeguard the system carried exits with them.

Price the exit before taking it. The interest arithmetic comes first and lands in the next section: three percent a month on whatever remains unpaid, which is £10.80 on a £60 stake over six months before the play itself has cost anything. The second price is harder to see and larger in consequence: no self-exclusion reach, no complaints route, no stake caps, no regulator. Those are not abstractions. They are the difference between a dispute you can escalate and a dispute you can only endure, and the sections below cost them in full.

The cost of a borrowed stake, modelled

Take the smallest realistic case and cost it in full. You borrow £60 on a card to fund an evening, play it, and the balance is not cleared when the statement lands. From that point the two meters run together. The interest meter adds three percent of the unpaid balance a month on this page’s assumptions, simple and non-compounding, with no fees folded in. The play meter has already taken its cut: at a blended three percent house edge, each pass of the stake through the games costs about £1.80 in expectation.

We carry two times through the model, as a conservative count of how often a borrowed stake is replayed before the night ends: a stake played once is the floor, a stake turned over twice before the evening closes is realistic, and the expected loss column below is built on the second. The three percent a month is likewise an assumption, not a quote of any card, and the bands in the table span the range a carried balance might plausibly attract. Your own statement’s rate is the number that belongs in the arithmetic, and the model accepts it without complaint.

Read the interest column against the expected loss column and the balance of power tells the story. On a stake carried for a single month, the interest is a rounding error beside the play. By six months the ordering has flipped: on the £200 row, £34.90 of interest sits against £12.00 of expected losses, and on the £500 row the interest alone has passed £112, close to a quarter of the original borrowing. The entertainment is a one-night cost; the carry is a monthly subscription to it, billed whether or not you play again.

The same curve drawn as a picture, with the share of the original borrowed stake lost to interest on the vertical axis and the months carried beneath:

The bar worth memorising is the last one. Six months of carry at the modelled rate consumes 18 percent of the original stake in interest alone, and the model is built conservatively: fees a card might add to a gambling transaction are excluded entirely, so the true drag sits at or above the bar rather than below it, whatever your own card’s terms turn out to say. Three months of carry takes 9 percent, one month 3 percent, and the line between them is straight because the model compounds nothing and hides nothing.

The calculator prices the play meter on its own. It takes a bankroll, a stake per round, a house edge and a number of sessions, and it returns the expected cost of that play. It does not forecast an evening, it does not remember the last session, and it does not schedule a recovery; it prices the arithmetic of volume, which is the only part of gambling that behaves predictably.

The prefill sets a £60 bankroll, £1 stakes, a three percent house edge and eight sessions. We carry eight sessions through the model, as a fair month of modest evening play funded this way. At £1 a round and a three percent edge, each pound of turnover costs about three pence in expectation, so the tool and the table stay in separate columns: the calculator prices the play, the table prices the carry, and the total cost of a borrowed stake is their sum. Change the edge input to match the game you would actually play, because slots commonly run a wider edge than three percent and some tables run narrower, and the honest costing is the one with your own numbers in it.

What the model refuses to do matters as much as what it does. It does not project a win that clears the balance, endorse a staking plan, or imply that a loss is owed a reversal; nothing in the arithmetic schedules a turnaround, because nothing in volume play produces one. The expected cost is a floor on the average price of the entertainment, not a ceiling, and the interest meter keeps running on the balance regardless of how any single session lands.

The offshore gap: what the rail costs you beyond interest

An offshore till that prints a credit rail is an operator with no UK licence. That is the whole definition, and everything else follows from it. The offences in the Gambling Act 2005, for providing unlicensed facilities and for advertising prohibited gambling, sit with the operator and the advertiser, never with the player. The protections a licence carries are not offences, they are entitlements, and entitlements simply do not exist where the licence does not.

Line up what disappears, because the list is longer than most readers expect. GAMSTOP, the free national online self-exclusion scheme for Great Britain, covers only operators licensed by the Gambling Commission, so an exclusion you set in good faith cannot reach an offshore account. The approved alternative dispute resolution that every licensee must be registered with, the route that turns a withheld withdrawal into an escalated case, has no offshore equivalent as of right. The UK regulator with the power to investigate, fine and delist has no jurisdiction there. The credit card ban itself, the protection this page is named after, does not apply. And the online slot stake caps, £5 per spin for adults and £2 for 18 to 24 year olds at licensed operators since September 2024, stop at the licence line too.

The tools vanish with the entitlements. Deposit limits, reality checks and account history are duties on a licensee; offshore, they exist only at the operator’s discretion and in whatever form it chooses to offer. A limit you can set is a different instrument from a limit you can request, and the difference shows up at exactly the moments limits exist for.

Then there is the tax position, which changes at the licence line as well. Operators licensed by the Commission pay Remote Gaming Duty on their GB-facing profits; operators holding no UK licence pay no UK gambling tax on their GB turnover. That is a fact about the market’s economics rather than about anyone’s personal bill, but it belongs in any honest accounting of what the two tiers actually are.

Put the rail back into that frame and it stops looking like a convenience. A credit option at a till with no UK licence is the protection being declined, printed as an offer. The reader gives up self-exclusion reach, dispute resolution, a regulator, stake caps and the ban itself, and receives in exchange a way to lose money they do not yet have, at a rate the card then compounds monthly. This page describes that trade in full and helps with none of it, because a gap in the fence is not a gate.

The market on one page: licences, offers and rails

The table gathers a slice of the market onto one page: licence labels, category-typical welcome ranges, wagering bands, minimum deposits, payout windows and the rails each kind of till tends to carry. Read the licence column before any other. The UKGC rows are the British baseline, where the credit rail has been closed since April 2020 and the card slots in the rails column mean debit. The offshore rows sit outside the ban, outside the scheme and outside the regulator, and those absences are the real difference between the two halves of the table.

Figures are category-typical ranges compiled from operator-published terms at the time of writing. They are not quotes of any single promotion. UKGC rows are the British baseline, shown for comparison; none of them are claimable at offshore offers.

Two cautions travel with the table. The ranges describe categories, not promises, and an individual cashier can sit outside its band in either direction at any time. And the rails column is the least decisive thing on the page: payment services including Visa, Mastercard, PayPal, Paysafecard, MuchBetter, Apple Pay, Skrill, Neteller and Trustly serve UK consumers, but what a given operator accepts varies by operator and changes over time, so the column is a sketch of the till, never a quote of it. The licence label is the column that prices the product, and it is the one no bonus percentage can move.