Duel Casino RTP — What 1,716 Games Actually Return

Duel gave us API access to its own game catalogue. What follows is computed from that file — 1,716 slot titles and 528 live tables, each carrying its return-to-player figure, house edge, volatility, hit ratio, maximum win and bet limits, pulled on 9 August 2026.

That matters because RTP is normally the hardest number to get honestly. Casinos publish it selectively, affiliates copy whatever the provider’s press release said, and the figure that actually runs on a given site is often a different build of the same game. These are the numbers Duel serves to its own front end.

One thing to hold onto before the tables: RTP is a long-run average, not a session promise. A 96% slot does not hand back 96 pence of every pound you stake this evening. It describes what happens across millions of spins, and the variance around it is the whole reason the games are entertaining.

Across all 1,716 slots the median game returns 96.36%, and the mean sits almost exactly on top of it at 96.35%. The spread is remarkably tight:

Ninety-two per cent of the catalogue lives in a single percentage point. That is not Duel being unusual; it is what the modern slot market looks like, because studios build to the band operators will accept. The practical consequence for you is that picking games on RTP alone is mostly noise — the difference between a typical 96.2% title and a typical 96.6% title is far smaller than the difference volatility will make to your evening.

The tails are where the decisions are.

Fifty games clear 97%. These are the ten highest in the catalogue:

GO Craps is the outlier worth understanding. A 100% RTP means the game carries no house edge at all, which is possible because it is a dice game played to a fixed paytable rather than a slot with a bonus economy. Its maximum win is x12, so it is a grind, not a jackpot. Relax Gaming takes six of the remaining nine places, which is consistent with its catalogue average further down this page.

Now the other end, and this is where the catalogue says something the marketing does not.

7 Monkeys returns 92.71%, which is nearly four points below the catalogue median. Stake a pound a spin at 700 spins an hour and the expected cost is about £51 an hour, against £25 on a median title. Rakeback softens that — 50% of the edge comes back, so call it £25 against £13 — but it is still twice the price for the same hour.

The pattern underneath is the more interesting part. Eleven of the twelve lowest-RTP slots in the catalogue are Thunderkick titles. Thunderkick also holds the second-lowest studio average on the site.

Here is the honest limit of what we can tell you: many studios ship the same game in several RTP builds, and the operator picks which one to run. The catalogue tells us what Duel serves; it does not tell us whether Duel selected low builds or whether Thunderkick supplies them that way. We have asked. Until there is an answer, treat Thunderkick titles here as the ones to check before you spin rather than as evidence of anything.

Averaged across every studio with at least thirty games on the site:

The whole league table spans 0.73 of a percentage point, so nobody here is being fleeced by their choice of studio. Still, Red Tiger and Thunderkick sit at the bottom and Relax Gaming at the top, and Pragmatic Play — which supplies a third of the entire catalogue at 577 titles — lands almost exactly on the median. If you play a lot of volume, favouring Relax over Red Tiger is worth roughly 73 pence per £100 staked before rakeback.

Everything above measures Duel against itself. This section measures it against everybody else, and it turns up the one thing in this whole exercise that genuinely surprised us.

Studios ship the same slot in several RTP builds, and the casino chooses which one to run. Hacksaw’s Wanted Dead or a Wild exists at 96.38%, 94.55%, 92.58% and 92.33%. Same artwork, same features, same name in the lobby. The only difference is how much the game keeps.

Across 16 titles that appear at ten or more of the 38 casinos in our comparison set, here is what one game looks like:

The gap between the top and bottom row is 4.05 percentage points. On £1,000 of turnover that is £40 of expected value, on the identical game, decided entirely by which lobby you opened it in.

It is the same three names every time

One low build could be an accident of licensing. Repeat it across a catalogue and it is a commercial policy. Counting how often each casino serves a build below the standard one for that title:

Whale.io and CoinCasino did not serve the standard build once. Not on a single shared title. SolCasino managed it once in eight. If you play any of the three, you are paying roughly two to three percentage points more than the same game costs elsewhere, and nothing in the interface tells you.

Below that line the numbers collapse into noise. Several casinos show a single “cut” of 0.01 to 0.03 points, which is rounding between data sources rather than a different build.

Nineteen casinos ran the full build every time

Including Duel — and this is the part an affiliate site is supposed to overstate, so let us be precise instead. Duel is not special here. It is one of nineteen casinos in the set that served the standard RTP on every shared title:

500 Casino, BetFury, Betstrike, Chips.gg, Cloudbet, Duel, Duelbits, Gamba, Gamdom, Goated, Jackpot Bet, Rainbet, Roobet, Shuffle, Stake, Thrill, Toshibet, Winna, Yeet.

Running the full build is the industry default, not an achievement. What the table shows is that the default is not universal, and that the exceptions are consistent enough to name. If you take one habit from this page, take this: check the RTP in the game info panel before you spin, on any casino, including this one. It is two clicks and it is the only number that follows you through the whole session.

Where these numbers come from, and one thing we did not do

The comparison uses a third-party review dataset covering 91 crypto casinos, dated 8 August 2026. Sixteen titles appear at ten or more casinos, which is what makes a like-for-like comparison possible; the per-casino counts in the table above are how many of those sixteen each one carries.

Duel’s figures in this section are the third party’s, not Duel’s own. That matters. Duel’s catalogue API reports slightly higher RTP for itself than the third-party set does — 96.43% against 96.38% on Wanted Dead or a Wild, and nine such gaps across twelve overlapping titles, every one in Duel’s favour. The differences are small, but taking an operator’s self-reported numbers and setting them against a third party’s numbers for its competitors would tilt the table before anyone read it. So the whole comparison runs on one source, and Duel is measured by the same ruler as everyone else.

Two limits worth stating. The sample is sixteen titles, not a full catalogue audit. And RTP builds change — a casino serving the standard version in August can switch in September, which is exactly why the habit of checking the info panel beats trusting any table, this one included.

None of the figures above apply to Duel’s own games, and they come from a different place. The Originals are not in the catalogue API — that file covers the third-party library only. The numbers below are the ones Duel publishes on its own game pages, which we check by hand.

Ten of the eleven run at a true 100%, and Beef is the one exception at 99.2%. This is the Zero Edge model, and it is the single thing that makes Duel structurally different from the casinos it competes with: a 0% edge means the game itself takes nothing, and Duel makes its money elsewhere. Note that Zero Edge runs inside a daily allowance rather than without limit, which the Zero Edge page covers in full.

Two details are worth a second look. Castle Roulette caps at x48, which matches the wheel exactly — its six bet sizes divide 48 segments into halves, quarters, eighths, sixteenths, twenty-fourths and forty-eighths, so every bet on the wheel carries identical expected value. And Beef reaches x183,278, by a wide margin the largest multiplier anywhere on the site — which is presumably why it is also the one Original that keeps a sliver of edge.

Duel’s own live blackjack table pays 60% instant rakeback against 50% on every third-party game. We have reported that since 6 August, and it is genuinely the best headline rate on the site.

The catalogue lets us check what that rate is actually worth, because it also carries the RTP of every live table. Net edge is what you pay after rakeback comes back:

The 60% table finishes third. Pragmatic’s ONE Blackjack starts from a better RTP, and the extra ten points of rakeback do not close a gap of 0.15 percentage points of house edge. On £10,000 of blackjack turnover the difference is about £2 — small, but it runs the opposite way to what the headline implies.

Two caveats we are not going to bury. Blackjack RTP figures assume optimal basic strategy; play by instinct and you will give back several times this difference before the rakeback rate matters at all. And Duel Blackjack is a physical table with a human dealer, so it sits outside Zero Edge and cannot be provably fair — you cannot run SHA-256 over a shoe of cards. Some players will take the real dealer over two pounds, and that is a perfectly good reason.

Live blackjack is where the site’s depth actually is, incidentally: 326 of the 528 live tables are blackjack, against 72 baccarat, 42 roulette and 28 game shows.

The catalogue figures sit alongside the rest of what we have measured directly:

• Zero Edge explained — how a 0% house edge game makes money
• On-chain data — what Duel’s public wallets move
• Slots at Duel — the library itself
• Live casino and game shows
• Rakeback rates, game by game