Dice simulator

How crypto dice payouts are calculated

The multiplier is 99 divided by your win chance. Roll under 50 wins 50% of the time and pays 1.98×. A fair game would pay 2×, and the missing 0.02× is the 1% house edge.

Every target has the same edge. A 2% chance pays 49.5×, a 98% chance pays 1.0102×. Changing the target changes how bumpy the ride is, not how much you lose on average.

Why rakeback matters more than the target

At a 1% edge you lose about $1 for every $100 wagered over time. Rakeback pays part of that edge back. At 50% rakeback (Duel's rate) the effective edge drops to 0.5%. That is the only lever that changes your long-run result.

Win chance and payout for common targets

Every dice target sits on the same curve: multiplier = 99 ÷ win chance. Here are common settings at the default 1% edge. Set Edge % to 0 in the simulator and the multipliers become the fair 100 ÷ win chance, which is what the house edge takes away.

Every one of these settings returns 99%. What changes from the safest to the riskiest is how often you win and how long you wait between wins.

• 98% win chance pays 1.0102× (roll under 98.00 or over 2.00)
• 90% win chance pays 1.10× (roll under 90.00 or over 10.00)
• 75% win chance pays 1.32× (roll under 75.00 or over 25.00)
• 50% win chance pays 1.98× (roll under 50.00 or over 50.00)
• 25% win chance pays 3.96× (roll under 25.00 or over 75.00)
• 10% win chance pays 9.90× (roll under 10.00 or over 90.00)
• 2% win chance pays 49.50× (roll under 2.00 or over 98.00)
• 1% win chance pays 99.00× (roll under 1.00 or over 99.00)

Variance: how long losing streaks last

At a 49.5% win chance you lose 50.5% of rolls. Ten losses in a row happen about once in 1,000 tries, and over a 1,000-roll batch the longest losing streak is typically around nine. At a 10% chance the same batch usually contains a losing streak of around 40.

The swing in your result is much larger than the edge. On 1,000 rolls of $1 at 49.5%, the expected loss is $10, but one standard deviation is about $31. Finishing $20 up or $40 down is ordinary. At a 10% chance the spread roughly triples, to about $94.

This is why a betting system can look brilliant for a few hundred rolls. Run several ×1,000 batches at the same setting and the observed RTP moves around 99% from batch to batch, while the long-run average stays where the formula puts it.

Dice and provably fair results

At a crypto casino, each roll comes from a server seed, your client seed and a nonce rather than your browser. Stake documents its dice formula, and our provably fair verifier follows it, so you can paste the seeds from a real bet and recompute the roll yourself once the server seed is revealed.

Verification proves the roll was not changed after you bet. It does not change the 1% edge, and it cannot predict the next roll. To lower the cost of the rolls you make, compare rates on the rakeback calculator.