Why cost per session beats return rate

Ask a search engine about online casino UK play and the answers are wall-to-wall return rates. Ninety-six percent here, ninety-seven there, and a suggestion that the higher figure is somehow the better game. A return rate is a property of a game; it is not a property of your evening. The number that decides whether you enjoy the night, and whether you can afford the next one, is how fast a given bankroll burns at the stake you actually play, over the hours you actually sit down for.

The distinction is easiest to see with a worked case. A game returning 97 percent keeps 3 percent of everything staked, on average, over a very long run. If you stake £1 a round and play 60 rounds an hour, you push £60 through the game in an hour, and the expected cost of that hour is £1.80. Change the stake to £5 and the same hour costs £9 in expectation, even though the return rate never moved. The return rate told you nothing; the stake and the pace told you everything.

This is the arithmetic the marketing never volunteers, and it is the arithmetic this page exists to do in the open. The approach throughout is the same one used across this site: no accounts opened, no games played, no offers claimed. The model reads published rules, including the technical standards the regulator sets for licensed games, and turns them into pounds per hour at stake sizes a reader recognises. Where a figure rests on an assumption rather than a published source, the assumption sits beside the number so it can be argued with, adjusted, or replaced with your own.

Session cost also reframes the question of which market to play in. The licensed half of the British market and the offshore half differ on return rates by fractions of a percent, which is close to invisible at any ordinary stake. They differ far more on the things that surround the game: whether a stake limit you set is enforced, whether a complaint goes anywhere, and whether the clock and the lights are designed with your welfare in the rules rather than in the slogan. Those differences are costed later on this page, because they are the honest commercial comparison, and they are worth more than any fraction of a percent of return.

One boundary statement before the model begins. This page describes and costs offers; it does not help anyone reach them. If you have self-excluded through the national scheme, or blocked gambling sites with software on your devices, the answer this page gives is to keep those protections in place and treat any operator outside them as out of bounds, whatever its terms look like on a screen. Nothing here is written to defeat a protection, and the cost model that follows is useless to anyone who should not be playing at all this evening.

The session model and its inputs

The model needs five inputs, and every one of them is a number a reader can set from their own habits rather than from a brochure. The inputs are bankroll, stake per round, rounds per hour, the house edge of the chosen game, and the length of the session in hours. Feed those five in and the model returns an expected cost for the session, the probability-flavoured spread around that cost, and the point at which the bankroll is statistically likely to run dry.

The house edge input deserves the most care, because it is the one readers get wrong most often. House edge and return rate are the same number seen from opposite ends: a slot returning 96 percent has a 4 percent edge, and a European roulette wheel returning 97.3 percent on even-money bets has a 2.7 percent edge. The edge is not what you lose per session; it is what you lose per unit staked on average. Multiply the edge by the stake and by the number of rounds, and the expected cost of the session falls out. The variation around that expectation, which is where the fun and the damage both live, is a separate quantity that grows with volatility and shrinks with the number of rounds.

Rounds per hour varies more between games than any other input. Online slots run anywhere from 300 to 800 spins an hour depending on the player and the settings; a live dealer table runs at the pace of the human dealing the cards, which lands around 50 to 70 rounds an hour for roulette and rather fewer for blackjack once decisions are counted. Regulated slots in Britain have a floor on spin speed precisely because pace multiplies cost: at a 2.5 second minimum, roughly 1,440 spins an hour is the mechanical ceiling, and the standards ban features that push the pace above what a deliberate player would choose.

Bankroll sizing falls out of the model rather than into it. A session budget that survives its own game should cover many multiples of the stake, and the model’s spread output shows why: a bankroll of 50 times the stake on a low-edge game endures a long cold stretch, while the same multiple on a high-volatility slot dies early more often than not. None of this is advice on how to win, because there is no way to win in the long run against a negative-expectation game; it is advice on how long the entertainment you budgeted for actually lasts, which is the only promise the maths can honestly make.

Two wheels, one stake, very different bills

The cleanest demonstration of session cost uses the two roulette wheels, because they are identical in every respect except the one number that matters. The European wheel carries a single zero and a 2.7 percent house edge. The American wheel adds a double zero and the edge rises to 5.26 percent. The games look the same, play at the same pace, and cost nearly twice as much per hour for no additional entertainment whatsoever.

Model assumptions: £1 staked per spin on even-money outside bets, no wins reinvested beyond the running balance, expected decline computed as stake times rounds times house edge, with the two wheels at 2.7 and 5.26 percent. Individual sessions land above or below these lines; the gap between them, however, is structural and does not wash out.

Read the gap column rather than either endpoint. Over an evening of 300 spins, the double zero wheel costs you about £15 more than the European one for the same number of decisions, the same graphics and the same pleasure of watching a ball bounce. At £5 a spin the gap is £75. No payout difference compensates for it, because the payouts are identical; the only difference is an extra pocket on the wheel, and the whole extra edge is the price of that pocket. This is why experienced players treat the American wheel as a puzzle rather than a game: the correct answer to the puzzle is not to play it.

The licensed British market settles this for you more often than not, since operators serving British players overwhelmingly stock the single zero wheel and live studios built for that audience. At the offshore end of the market both wheels appear side by side, along with games whose edge is published nowhere in the lobby. The lesson generalises: before any session, find the edge of the specific game in front of you, usually in its own help screen, and let that number, not the lobby’s presentation, set your expectation for the evening. A reader who does nothing else from this page but check for a double zero before sitting down has already saved more than most bonus offers are worth.

What slots and live tables do to the same bankroll

Take the same £100 bankroll from the wheel comparison and move it to the slots. The edge is worse, at the 4 percent that a 96 percent return implies, but the pace is the real difference. A deliberate table player makes 300 decisions in an evening; a slot player running quick spins makes that many in well over an hour, and an engaged player chasing a feature can triple it. The expected cost of the same bankroll on slots is therefore not a little higher but a multiple higher, because the number of rounds, not the edge, is doing the multiplying.

Regulated slots in Britain are built to keep that multiplication honest. The technical standards the regulator imposes require a minimum spin speed of 2.5 seconds, ban turbo buttons and slam stops, and forbid celebratory animations for losses that dress a losing spin up as a near win. Those rules exist because pace is cost, and they apply to every operator on the Commission’s public register. An offshore lobby carries no such guarantee: the same game may run at whatever speed the player can click, and game versions exist whose published return differs from the British deployment of the same title. Where a lobby offers the same slot at two return figures, the difference is real, and the higher-edge version always sits where the rules are lighter.

Live tables run the other direction. A human dealer sets the pace at 50 to 70 rounds an hour, the edge on the good games is among the lowest in the building, and the result is the cheapest serious entertainment per hour in the casino: a blackjack session at basic strategy and a £2 stake costs roughly £3 an hour in expectation on a 3-to-2 game with the common rule set, while a £2 slot stake over the same hour costs several times that. Baccarat on the banker bet and French roulette with la partage sit close behind blackjack; American wheels and side bets sit far ahead of it in cost, in the wrong direction.

The practical shape of the choice is simple enough to carry in your head. If the evening you want is long, slow and cheap per hour, live tables at the lowest edge and the smallest stake the table allows are the efficient answer. If the evening you want is bright, fast and loud, slots deliver that, and the price of delivery is pace, which the model prices per spin and per hour both. Neither choice is wrong; paying for one while believing you bought the other is the only mistake the model exists to prevent.

The hour-by-hour picture at three stake sizes

Scaling the model across stake sizes makes the cost structure of the market visible in one glance. The table below holds the game constant, at a 3 percent edge typical of a European wheel session, and varies the stake. Rounds per hour are held at 60, which suits a live table. Every figure is an expectation, and short sessions swing around it, but the proportions hold at any swing.

Model assumptions: 60 rounds per hour, 3 percent house edge, expected cost computed as stake times rounds times edge, and survival estimated as bankroll divided by expected hourly cost. Variance shortens survival on volatile games and lengthens it on lucky runs; the expectation is the honest centre of that spread.

Two readings of that table matter. The first is that cost per hour at low stakes is comparable to other paid entertainment: a pound-fifty an hour at the £1 stake is less than a coffee, and the comparison is fair because the expectation is what you pay for the whole evening’s play, win or lose. The second is that the scale doubles with the stake and never stops. At £10 a spin the same game costs as much per hour as a restaurant meal, and the high roller rooms take that scale to figures where an hour’s expectation exceeds what most readers spend on a hobby in a month. The model does not rank these choices; it prices them, so the choice is made with the price showing.

Bonuses interact with this picture in a way worth stating plainly, because the connection between a welcome package and a session is closer than it looks. A deposit match at this end of the market usually asks around forty times the bonus in turnover before it converts, and the turnover requirement is fed by exactly the per-spin stakes this section prices. A player clearing a £100 bonus at 40x is placing £4,000 of bets, and at a 4 percent slot edge that costs £160 in expectation against a bonus worth at most its capped figure. The bonus is not free money sitting on top of the session; it is a schedule of sessions with a cap on the upside, and costing the two together is the only way to see whether taking it improved the evening.

How named operators compare on the numbers that matter

The table below puts a dozen named operators side by side on the terms a session player meets first: the welcome range, the wagering attached to it, the smallest deposit, the payout window and the rails. Figures are category-typical ranges compiled from operator-published terms at the time of writing, not quotes of any single promotion, and they shift as operators shift them. British-licensed rows are the domestic baseline; offshore rows are shown so the two halves of the market can be read next to each other.

Figures are category-typical ranges compiled from operator-published terms at the time of writing, not quotes of any single promotion, and they change as operators change their terms. UKGC rows are the British baseline shown for comparison; none of the terms in the UKGC rows are claimable at offers outside the British register, and none of the offshore rows carry British licence protections. Where a figure is uncertain, the row says so rather than guessing.

The pattern to carry away from the table is the same one the session model shows from a different angle. The licensed rows compete on cheaper wagering, faster published windows and familiar rails, because the licence obliges them to publish terms that hold. The offshore rows compete on headline size and on rails the licensed market cannot run, such as crypto cashiers and credit cards, and the ranges in their rows are wider because their terms are set operator by operator with no common floor beneath them. A reader using the table should read the wagering column before the headline column in every row, and treat any offshore figure marked as varying as a question to answer on the operator’s own terms page, not a number to plan around.

What the licence changes in pounds

The licence gap between the two halves of the market is usually presented as a legal footnote. This section presents it as what it is for a session player: a list of things the British licence does that follow you into the evening, each with a cost attached when it is absent.

Priced against the session model, the most expensive item on that list is the last-but-one: a dispute over a frozen withdrawal has no referee in the offshore half, and the value at stake in such a dispute is typically the entire account balance, not the evening’s expectation. Everything else on the list is insurance against a bad month rather than a bad evening, but insurance is priced by the tail, and the tail is where a session player with no limits and no dispute route eventually lands. The honest summary is that the offshore half sells a slightly bigger headline and charges for it with the safety net, and a reader who has read this far can see both sides of the trade on the same page.

The gap also explains a quiet pattern in the comparison table: licensed operators rarely run the loudest offers, because their marketing budget pays for the obligations instead. That is not weakness. A 100 percent match at 30x from a brand on the Commission’s public register is, for a reader who values any item on the list above, a better offer than a 300 percent match at 50x from outside the register, and no arithmetic on this page contradicts that, because the list items are exactly the ones the arithmetic cannot price.

Questions readers ask

Keep the session affordable