Classic over/under dice - mostly accountless, mostly ~1% house edge, all provably fair where noted.
operators on this page: 1 · playable today (excl. watch list): 1 · proven ZEC support: 100% · hold balance in ZEC natively: 100%
quick picks
A long-running dice site that documents native ZEC play - but the Anjouan license number claimed in its own terms does not appear in the official register (checked 2026-10-24). Watch-grade: small deposits only, test withdrawal first.
Dice is the purest house-edge product there is. The scheme: you pick a win chance (say 49.5%), the site rolls 0-9999, and the payout multiplier is set so the expected return is the complement of the edge:
Two readings of that table matter in practice:
ZEC dice sites inherit a pattern older than Zcash: SatoshiDice ran on Bitcoin in 2012-13 and at peak produced a large share of all on-chain transactions - the Federal Reserve's Bitcoin technical analysis documents the scale [3]. The design moved to shielded-capable chains largely intact: commit-reveal seed pairs, on-chain-verifiable outcomes, and house edges of ~1%. What changed with Zcash is the ledger privacy - bets at a shielded-native site reveal no amounts or addresses, a property Bitcoin dice never had.
Every site in this category claims "provably fair." The claim is checkable, and the check takes thirty seconds:
serverseed:clientseed:nonce → map to 0-9999.Our provably-fair verifier does the recomputation for you. If a site's "verify" page never shows the unhashed seed, or the recomputed numbers disagree with history - close the tab and blacklist the operator. The provably fair guide covers the schemes in detail.
Three ZEC-specific points decide more than the edge:
At a fixed 1% edge, session survival is a bankroll and bet-sizing question: flat bets of 1% of bankroll survive long downswings; anything Martingale-shaped dies on a streak the table's limit conveniently truncates [2]. The expected value never turns positive - no bet pattern beats a negative-edge game, and any tool claiming otherwise is selling something. The honest play is treating the 1% as entertainment cost, sizing stakes to lose it slowly, and verifying fairness rather than trusting it.
Around 1%. The classic scheme (payout = 99/chance) has run since SatoshiDice in 2012-13, when dice bets famously consumed up to half of Bitcoin's transaction volume. Above 2% is poor value for the genre; the exact number belongs in each review's house_edge field.
Dice schemes publish server seed (hashed before the bet), client seed and nonce. Hash server seed + client seed + nonce, map the result to 0-9999, compare against your win chance. Our verifier recomputes it - if a site will not show you its seed pair, that is the answer.
Yes - the operator's confirmation table lists native ZEC at 1 confirmation (0.168 ZEC minimum), against 10-12 for most token networks. One confirmation on ZEC's 75-second blocks means deposits are playable in roughly a minute or two.
No. Doubling after losses needs bankroll exponential in the losing streak, and table limits cut the progression at 8-10 steps. The expected loss per round is fixed at the edge; Martingale only reshapes the loss distribution into rare catastrophic hits.
High-chance/low-multiplier bets feel safe but pay the same 1% edge per roll, and auto-bet turns variance into a smooth grind toward it. The edge scales with total wagered, not with bet size - 1,000 rolls at 0.01 ZEC costs the same expected 1% as one 10 ZEC roll.
If the seed scheme is honest, no - each roll is independent and the hash function has no exploitable bias. Any product claiming to predict dice outcomes is selling the predictor, not the edge (see the crash hub for the same scam pattern).