Predict whether the next roll will land over or under a target you choose. Payout is proportional to the probability, bet on a 5% chance and win 19.8×, bet on a 95% chance and win 1.04×. 99% RTP across every target. Variance entirely up to you.
Slide the target between 0.01 and 99.99. Pick over or under. The probability of winning equals the slider position; the payout multiplier is 99 / probability (the 1% gap is the house edge). Roll. The result is a number 0.00–99.99; if it falls on the predicted side, multiplier × stake pays out.
All target selections share the same 99% RTP, pick what matches your variance tolerance, not what feels lucky. 95%-win/1.04× is a smooth grind; 50%-win/1.98× is balanced; 5%-win/19.8× is high-variance hunting. Martingale on Dice is a classic bankroll killer: doubling after each loss only works until you hit the table cap or your wallet limit, both of which arrive faster than people expect.
Each roll is derived from HMAC-SHA-256 of the seed pair + nonce; the first 4 bytes are mapped to a number in [0, 99.9999]. Recompute the digest, apply the same modulo, get the same roll the casino published.
Read the full fairness guideAt 0.01% win chance, payout is 99 / 0.0001 = 990,000× theoretical, capped at 9,900× by table policy. The cap is published in the rules and reflected in the displayed multiplier.
No. It survives long losing streaks until it doesn't, at which point you hit the table cap or your bankroll. Long-run EV is unchanged at 99%.
Each roll is independent. Streaks happen because of variance, not because of memory. Past results don't change the next roll's probability.
Yes. Auto mode supports number of rolls, target/over-under settings, increase-on-win, increase-on-loss, stop-on-profit and stop-on-loss conditions.