Limbo is the simplest game Duel makes: you pick a multiplier, the game draws one, and you win if the draw lands at or above yours. The tile says 100% RTP. The game’s own two boxes say 99.90%. Both are true, and the difference is worth thirty seconds of your time — because it is the whole economics of the game. Below: the formula, checked at four multipliers, and a simulator that runs the real numbers so you can watch what a 0.10% edge does over a thousand rounds.

A round is one number against one number.
You type your target — say 2.00×. The moment you do, the game tells you your chance: 49.95%. You press Bet. The game draws a random multiplier and shows it in the middle of the screen. If the draw lands at or above your target, you are paid your stake times the target. If it lands below, the stake is gone.
There is nothing else. No cards, no reels, no decision after the button. Limbo is the one game here where you set your own odds: the multiplier you type is the probability, and the payout is only the other side of the same number. Type 2× and you are flipping a coin. Type 100× and you are buying a lottery ticket that pays a hundred times.
Duel publishes no rules page for Limbo. It does not need to: the two boxes under the game give the whole thing away. Enter any multiplier and the win chance updates instantly, and the relationship never moves:
Checked on the live game, 6 September 2026. Four points, one formula, no drift.
So the rule is:
Which settles the RTP in one line. At 2.00× you win 49.95% of the time and get paid double: 0.4995 × 2 = 0.999. Ninety-nine point nine percent comes back. The house keeps 0.10%.
That is what the 100% RTP tile means — 99.9 rounded up to a round number. It is not a lie, and it is not 100 either. For scale: the same game at most crypto casinos runs a 1% edge, so Duel’s is ten times smaller. On this one, the badge is closer to true than most.
Duel has no demo mode for Limbo — you cannot try it without depositing. So here is the game running on the exact formula above, with fake money. Pick a target, press Bet, and watch where the number lands. Then run a thousand rounds and see what the 0.10% actually does.
Run the simulator at 2× for a thousand rounds and you will notice something: the edge is almost invisible. Stake $10 a round and you turn over $10,000; the house expectation on that is about $10. Everything else you see on the chart — the swings, the run that put you $400 up, the run that took it back — is variance, not the edge.
That is the honest read on Limbo, and it cuts both ways. The edge will not ruin you. The distribution might. Type 10× and you lose nine rounds out of ten by design; the maths works out over thousands of rounds, but you will sit through losing streaks of fifteen and twenty that feel nothing like 99.9% RTP. Type 100× and the median session is simply a straight line down, because the one round that pays for the rest arrives, on average, once every hundred bets.
The number to watch in the simulator is not the balance. It is the longest losing streak, because that is the thing that actually ends sessions.
The tile says 1000×, and the field accepts exactly that: type 1000.00 and Duel shows a win chance of 0.0999% — roughly one round in a thousand. Whether the game will take a target above that, we have not tested and will not guess.
Worth knowing what a 1000× target really means: at one round per second, non-stop, you would expect to hit it about once every seventeen minutes. Most people do not play a thousand rounds of anything.
Limbo carries a Provably Fair badge, and unlike a lot of casino marketing this one is a real mechanism rather than a slogan. Each round’s result comes from three ingredients: a server seed the casino commits to in advance (you see only its hash), a client seed you control and can change, and a nonce that counts up with every bet.
Because the server seed is committed before you bet and revealed afterwards, you can recompute any past round and confirm the result was fixed before you pressed the button. That is a genuine guarantee, and it is more than most operators offer.
What it does not prove is that the odds are good. Provably fair says the deal was not changed after the fact; it says nothing about how the deck was built. The 0.10% edge is a separate claim, and the way to check that one is the arithmetic above — which is why we did it.
Not on Duel. Limbo has no free-play mode: to place a round you need a funded account, which means depositing first. That is the single most-searched thing about this game and the answer is an honest no.
The simulator above is the closest thing that exists — same formula, same win chances, same 1000× ceiling, fake money. It will not teach you anything about Duel’s random number generator, because nothing outside their provably-fair system can. But it will show you, in about ten seconds, what your chosen multiplier does over a thousand rounds — which is the part most people get wrong before they deposit, not after.