guide
Crypto Dice Explained: Over/Under, Provably Fair, and the House Edge
How crypto dice works — the 0–99.99 target mechanic, provably fair verification, the 1% house edge behind the payout formula, and the 400,000-session simulation showing why no betting system moves it. Updated August 2026.
Crypto dice is a bet on whether a random number between 0 and 99.99 lands above or below a target you choose. Most crypto dice games run at a 1% house edge — a 99% RTP — as of August 2026, and the edge is stated in the payout formula before every bet, which makes dice the most mathematically transparent game at any crypto casino. No betting system changes that number: in our 400,000-session simulation, every progression tested lost the house edge exactly.
This page contains affiliate links. Commissions we may earn do not influence casino ratings or placement.
How Crypto Dice Works: Target Number, Roll Over, Roll Under
One slider sets everything: the target number determines your win chance, and the win chance determines your payout multiplier — the house edge is the fixed gap between the two. You drag the slider to set your target number, then choose Roll Over or Roll Under. The game shows three derived values before you confirm the bet:
- Win chance — the percentage of the 0–99.99 range on your side of the target
- Payout multiplier — how much you receive (including your stake) on a win
- House edge — the built-in margin, usually displayed as the RTP complement (1% house edge = 99% RTP)
A sample configuration at a 1% edge: a win chance of 49.50% pays a 2.00× multiplier, because 0.495 × 2.00 = 0.99 — one percent short of break-even. Some interfaces present the same edge the other way around, showing a 50.00% win chance paying 1.98×; the product is 0.99 either way. That missing 1% is the house’s share, and it is identical at every slider position.
The roll itself is generated by a pseudorandom number algorithm seeded by cryptographic inputs. The result lands anywhere in the 0–99.99 range with equal probability. There is no memory between rolls; past outcomes tell you nothing about the next.
How the dice edge compares with our measured data
Dice belongs to the crypto-originals family, where the house edge is set in the payout formula rather than hidden in game design. Casino Aurum measured house edges of 0.47–2.01% across 24 million simulated rounds of Plinko and Wheel, against a stated ~1% target, in a study published July 2026 — the full ranked table is in house edge statistics. The same invariance holds for dice’s win-chance slider as for crash’s cash-out target: in 475 million simulated crash rounds, every target cost the same 1% — the slider you choose sets variance, not price.
We did not separately simulate dice, and do not need to: the dice edge is computable directly from the payout formula, which is the entire point of the format. The figure to check is the multiplier a site pays at a given win chance, not a published RTP.
How Provably Fair Verification Works in Crypto Dice
A properly implemented crypto dice game lets you verify every roll yourself: the server commits to a hashed seed before the roll and reveals the full seed after it, so you can re-derive the outcome independently. This is what distinguishes crypto dice from an opaque RNG slot.
Before each roll, the server commits to a server seed by publishing its cryptographic hash (a fixed-length fingerprint that cannot be reverse-engineered). You supply a client seed — either auto-generated or one you set manually — and a nonce that increments with each bet. The roll outcome is derived by combining all three inputs through a deterministic hash function, typically HMAC-SHA256.
After the roll (or when you change your seed), the server reveals the plain-text server seed. You can hash it yourself and confirm it matches what was published before the round, then re-derive the outcome to verify the casino generated the same number you saw. Our provably fair guide walks through the verification steps in detail.
Two honest caveats: first, most players never actually run the verification — the system’s integrity depends on you doing it, not just trusting that you could. Second, the hashing implementation needs to be audited; a casino could publish a valid hash but use a different algorithm for the live game. Check whether the site’s code has been independently reviewed. For a comparison of provably fair versus traditional RNG, see provably fair vs RNG.
What Is the House Edge and RTP of Crypto Dice?
Most crypto dice games run at a 1% house edge, meaning a 99% RTP, as of August 2026. A 1% house edge means an average loss of one unit per hundred wagered over a long series of bets, at every target setting. Session variance can be dramatic — dice is high-frequency, and a favourable streak feels like evidence of skill or pattern — but the math is indifferent to streaks.
The payout formula is:
Multiplier = (100 − house edge %) ÷ win chance %
At a 1% house edge and a 49.50% win chance: 99 ÷ 49.50 = 2.00×. Push the slider to a 10% win chance (a risky long-shot bet): 99 ÷ 10 = 9.9×. The expected value of every bet is identical: −1% of the stake. Adjusting the target shifts variance — more extreme targets mean bigger swings — but the long-run expectation does not move.
See our house edge guide for a fuller treatment of how RTP compounds across session length.
Which Crypto Casinos Offer Dice Games?
Every major crypto casino carries a dice game — the format arrived early, with Satoshi Dice launching in 2012, because it needs minimal infrastructure and maps cleanly onto provably fair verification. The table below orders operators by our honest overall rating, with a documented strength and a documented weakness for each.
| Casino | Rating | Trust | Documented strength | Documented weakness |
|---|---|---|---|---|
| Stake | 4.4 | High | House-built Stake Dice, widely cited as the reference implementation; rakeback welcome offer, no wagering lock | Rakeback is modest in cash terms — about $35 back per $100,000 wagered at a 1% edge |
| Shuffle | 4.3 | High | Publishes its bonus rollover formula in full — rare transparency | Newest of the group (2023), shortest track record; 35x wagering with a $10 max bet on the match bonus |
| BitStarz | 4.2 | High | Hybrid crypto + fiat; accepts fiat deposits | Bonus terms cap bonus-derived payouts at 100 EUR/USD and bets at 5 EUR/USD while a bonus is active |
| Cloudbet | 4.2 | High | Oldest BTC casino (since 2013); rakeback welcome package, so no rollover locks the deposit | Terms bar 20 locations including Germany, the UK, Spain and the US |
| BC.GAME | 4.0 | Medium | Wide coin selection; own dice title | Bonus terms could not be verified at source in August 2026 — the help centre was unreachable throughout our checks |
| Bitcasino | 4.0 | Medium | Early BTC casino, established track record | Trust rated Medium in our assessment; no standout documented policy advantage |
Also carrying dice: Roobet (3.9 — crypto-only, geo-restricted in some regions), Duelbits (3.8 — crypto-focused, sportsbook integration) and Rollbit (3.8 — NFT and trading features alongside the casino).
Ratings reflect our honest assessment across licensing, payout record, and transparency — not commercial relationships. Verify current licence details on each operator’s site directly; licensing status can change. For a deeper comparison of the top picks, see Stake vs BC.GAME vs Shuffle.
Does Playing Dice Clear a Casino Bonus Faster Than Slots?
At Shuffle, no — and the operator’s own published formula proves it. Shuffle’s help centre states, as read on 8 August 2026:
Wager Amount × (Game Edge / 0.04) = Rollover Progress
— Shuffle help centre, 100% first deposit bonus terms
Shuffle’s own worked example credits a $5 dice bet with $1.25 of rollover progress, because dice’s 1% edge is a quarter of the 4% baseline. The consequence: clearing Shuffle’s 35x requirement on a $100 deposit plus $100 bonus carries an expected cost of $280 whether the wagering runs through 4% slots or 1% dice — the low-edge advantage is cancelled by the formula. BitStarz reaches the same outcome differently, weighting table and live games at about 5% of turnover. We worked through five operators’ actual bonus terms here, and in our simulation of 1.3 million bonus claims, only 16.16% of players cleared a 40x requirement at all.
For dice play without a wagering lock: Cloudbet and Stake run rakeback welcome offers with no published rollover multiplier, so no requirement ever attaches to the deposit.
Can Martingale or Any Betting System Beat Crypto Dice?
No. Each roll is independent with a fixed negative expectation, and Casino Aurum’s 400,000-session Monte Carlo study (published August 2026) measured what that means in practice: Martingale, D’Alembert, Fibonacci and flat betting each lost between 2.68% and 2.72% of total turnover on European roulette — the game’s 2.70% house edge, unchanged by any progression. The study ran on roulette even-money bets, but the arithmetic transfers to dice unchanged, because both are sequences of independent fixed-odds trials. The full model and seed are published here.
Because dice offers fine control over win probability, it attracts more explicit strategy claims than almost any other casino game. The measured results:
Why Martingale fails at dice
Martingale doubles the stake after every loss and resets after a win. In the 400,000-session study, Martingale exhausted a $200 bankroll in 56.93% of sessions before 250 rounds were up — even with a $500 table maximum — and still lost 2.72% of everything wagered. The deeper a losing streak goes, the larger each successive bet must be; the expected outcome per roll never improves. You are borrowing variance from the future, not eliminating the edge. See the Wikipedia entry on Martingale for the formal proof.
Why winning most sessions still loses money
Fibonacci finished ahead in 66.86% of simulated sessions with a median result of +$37 — while averaging a loss of $15.62 per session. Both numbers are true at once: progressions convert many small wins into a few large losses, and the cost concentrates in the sessions where the recovery never comes. Judged by a typical night, a progression looks like it works; the arithmetic only shows up in the tail. Reverse Martingale (Paroli), which raises bets after wins instead, caps the downside but equally caps the upside — and the house edge applies to every bet regardless of sequence.
Why high win-chance target hunting does not reduce the edge
Some players set very high win chances (97.5%, for example) to minimise variance. That does reduce the chance of a catastrophic single-round loss, but the house edge is unchanged — at 99% RTP the expected loss is still 1% of each of many small bets, which accumulates.
The fundamental reason no system works: each roll is independent. The outcome distribution resets every round. There is no pattern to exploit, no due roll, and no seed configuration that biases the outcome in your favour.
How Fast Does the House Edge Compound at Dice Speed?
Faster than at any other casino game, because dice is the fastest game on the floor. Automated bet modes run hundreds of rolls per minute, and at a 1% house edge, 10,000 rolls at one unit per roll carry an expected loss of 100 units — regardless of the strategy applied. A 99% RTP sounds generous until it is applied at machine speed.
The session management tools some operators offer — bet limits, loss limits, session timers — are worth using. Our responsible gambling tools guide covers what to look for and what different licences require operators to provide.
Gambling carries real financial risk. 18+ only. Play only where online gambling is legal in your country of residence.
Bottom line: Crypto dice is the most transparent game at most crypto casinos. The payout formula is visible, the provably fair mechanism is auditable, and the house edge is stated explicitly before you bet. That transparency is genuinely valuable — it means you know exactly what you’re paying for. What it does not do is make the edge disappear. No target setting and no betting progression changes the fact that the expected return is negative on every roll. Play within a hard budget, run the provably fair verification at least once so you understand what you’re trusting, and verify licensing before depositing.
18+ only. Gambling involves real financial risk. Play only where it is legal in your jurisdiction. If gambling is causing harm, contact your national helpline or visit NCPG.
FAQ
- What does 'roll over / roll under' mean in crypto dice?
- Before each roll the player picks a target number on the 0–99.99 scale and predicts whether the result lands above it (roll over) or below it (roll under). The chosen range sets both the win probability and the payout multiplier — a narrower winning range pays more but wins less often, and every combination carries the same house edge.
- Is crypto dice provably fair?
- Yes, in properly implemented games. Before the roll the server publishes a hashed server seed; after the roll it reveals the full seed, so the player can independently verify the outcome using the published algorithm. The catch is that 'provably fair' only means something if the verification is actually run — and if the operator's hashing implementation has been independently audited.
- Can Martingale or any other system beat crypto dice?
- No. Casino Aurum simulated 400,000 betting-system sessions (study published August 2026): Martingale, D'Alembert and Fibonacci each lost between 2.68% and 2.72% of total turnover — the house edge, unchanged — and Martingale exhausted a $200 bankroll in 56.93% of sessions. Each roll is independent with a fixed negative expectation; progressions change the shape of the ride, never the price.
- What is the RTP of crypto dice?
- Most crypto dice games run at a 99% RTP — a 1% house edge — as of August 2026. Over a long series of bets that averages a loss of one unit per hundred wagered. Session variance is large because dice is high-frequency, but the long-run expectation does not change with any target setting.
- Does target hunting with high win chances reduce the house edge?
- No. A very high win chance such as 97.5% reduces the chance of a catastrophic single-round loss, but the house edge is unchanged: at 99% RTP the expected loss is still 1% of every stake, spread over many small bets. Adjusting the target shifts variance, not expected value.
- Does playing crypto dice clear a casino bonus faster than slots?
- At Shuffle, no — the operator's published rollover formula scales progress by game edge divided by 0.04, so a $5 dice bet contributes $1.25 of progress and the expected cost of clearing the 35x bonus is $280 on a $100 deposit whichever game is played. BitStarz weights table and live games at about 5% for a similar effect. Cloudbet and Stake publish no rollover multiplier at all, because both run rakeback welcome offers instead of match bonuses.
