Sample roll-out
12 rounds at 50/50 threshold · verification cycleThreshold index
What is Dice? A uniform 0 to 100 random number is rolled; the player sets a threshold (e.g., roll over 50, or roll under 30). Win-probability is the threshold range; payout multiplier is the inverse of that probability minus the house edge. Lower-probability threshold picks pay higher multipliers but win rarely. The RTP figure captures the long-run average across whatever threshold strategy the player runs.
The Stake dice build is the canonical uniform 0-100 threshold-pick mechanic at Stake that defined the 99 percent RTP standard for crypto-originals. Stake launched in 2017 under Medium Rare N.V. operating entity and Curaçao eGaming licence; the dice implementation is one of the most-replicated original mechanics in operator coverage, with cross-brand variants at every cluster operator and Stake-alumni-team Shuffle running identical math. Verified per-game RTP is 99 percent (1 percent house edge), audited 2026; the win-chance is configurable from 0.01 to 98 percent with payout multipliers scaling reciprocally.
If you have read the cross-brand dice math at see the dossier, the win-chance versus payout-multiplier framework is familiar; this page is the Stake-specific reproduction.
What is tested, what is pending at the threshold pick
We open with the honest data state. Verification is firm on the published 99 percent cluster baseline.
| Fact | Status | Source |
|---|---|---|
| Fairness method = HMAC-SHA256 | Verified | Stake published implementation docs |
| Server-seed commit-reveal workflow | Verified | First-hand reproduction at Stake |
| Game type = uniform 0-100 threshold pick | Verified | Stake brand-published game info |
| Casino licence (Curaçao eGaming) | Verified | Stake site footer plus Curaçao registry |
| Brand (Medium Rare N.V., Curaçao-incorporated, since 2017, audited 2026) | Verified | Stake terms plus cross-reference |
| Stake dice rtp = 99 percent (1 percent house edge) | Verified | Stake publishes per-game RTP |
| Win-chance configurable 0.01 to 98 | Verified standard | Stake game info screen |
| Roll-over / roll-under toggle | Verified | Stake UI |
| Auto-bet with stop-on-win / stop-on-loss conditions | Verified | Stake UI |
| Instant crypto withdrawals (~3 minutes) | Verified | Stake cashier docs plus first-hand observation |
| Threshold-based KYC | Verified | Stake terms |
| VIP Stake Reload (scales with tier) | Verified | Stake VIP screen |
| No welcome bonus (deliberate brand stance) | Verified | Stake terms |
| brand-side maximum bet at lottery-shape targets | Pending verification | Standard catalogue-wide limits expected |
The Stake Dice fairness side is verified plus the 99 percent RTP is verified as cluster baseline.
Why dice is the reference implementation
The uniform threshold-pick mechanic is the simplest auto-bet-friendly Original in the crypto-casino category. Stake codified the implementation:
- 2017 launch: Stake brought structured publicly verifiable dice to the crypto-casino market alongside Crash and Plinko. Every cross-brand cluster Brand (Shuffle, Yeet, Winna, Gamdom) launched dice with reference to Stake's published implementation.
- Identical math cross-brand: the uniform 0-100 threshold-pick math is structurally identical across operators. What varies is the house-edge factor in the payout multiplier formula.
- Stake-alumni team operators carry the implementation: Shuffle Dice uses the same HMAC scheme directly. Duel Dice at verified 100 percent (within allowance) shares the algorithm with a tighter house-edge factor.
- Auto-bet workflow conventions: stop-on-win, stop-on-loss, max-bet, max-loss configurations on Stake Dice are the convention competitors replicate.
- Real-time hash display: Stake's fairness UI displays the SHA-256 server-seed commitment before each roll. Cross-brand standard.
- What this means for players: Stake dice is structurally equivalent to Shuffle dice and the broader 99 percent cluster. Brand differentiators are operational track record, VIP depth, withdrawal speed, and presence/absence of token rakeback.
Uniform threshold-pick math at the checked 99% baseline
The Stake Dice build runs the canonical uniform 0-100 threshold-pick mechanic at the verified 99 percent RTP target:
- The outcome is a uniform float between 0 and 100, derived from HMAC-SHA256 byte stream.
- Player picks a threshold target T (between 0.01 and 98); the roll wins if outcome falls under T (roll-under mode) or over 100-T (roll-over mode).
- Win chance equals T percent. Payout multiplier equals 99 divided by T (the 1 percent house-edge factor encodes here).
- At T = 50: win chance 50 percent, payout 1.98x.
- At T = 5: win chance 5 percent, payout 19.8x.
- At T = 1: win chance 1 percent, payout 99x.
- At T = 0.01: win chance 0.01 percent, payout 9,900x.
- Cross-brand math is identical; the 99 percent factor places Stake at cluster baseline.
- The 99 percent figure is verified per Stake's published documentation and reproduced in our HMAC replay sample.
The Stake threshold-pick mechanic is the reference. The 99 percent factor encodes the cluster baseline.
The fairness reproduction walk-through for Stake Dice
The per-roll fairness verification works the standard way at Stake. The seven-step routine applies:
- Open the Stake Dice fairness panel and copy the published server-seed hash before placing a roll. Stake exposes the SHA-256 commitment in real time before each roll.
- Place a sample of 50-200 dice rolls at consistent target configurations (for instance T=50 for half). Record per-roll inputs: client seed, nonce, recorded outcome float, recorded win or loss.
- After the sample: rotate the server seed in the Stake account settings. Stake reveals the raw seed.
- Run SHA-256 locally on the revealed seed. Result must match the captured commitment.
- For each roll in the sample: run HMAC-SHA256 over (revealed seed, client seed, nonce, cursor). Apply Stake's published byte-derivation mapping into a uniform 0-100 float.
- Confirm the reproduced outcome matches the recorded outcome on every roll.
In our cycle reproduction on Stake, the HMAC-replay flow worked correctly on every sampled Dice roll across multiple target configurations.
Stake dice rtp at documented 99%: cluster baseline reference
- Verified 99 percent RTP (1 percent house edge) per Stake documentation. Cluster baseline reference.
- Duel dice at verified 100 percent (within allowance) (coverage leader on raw RTP, 0.9 percent above Stake).
- Rollbit Dice at 99.5 percent (0.5 percent above Stake), plus RLB token rakeback overlay.
- BetFury Dice at 99.28 percent (0.28 percent above Stake), plus BFG dividend yield.
- Standard 99 percent cluster: Stake, Shuffle, Yeet dice, Winna dice, Gamdom dice (all matched on raw RTP to Stake baseline).
- Fairspin TFS dice: 97 percent expected.
- Roobet dice: expected 96-98 percent (undisclosed precisely; Crash anchor signal).
- Stake has no token rakeback; relies on VIP Stake Reload for tier-scaled compensation.
For raw-RTP optimisation Duel leads; for raw-RTP plus token-rakeback overlay Rollbit/BetFury/Shuffle compete; for reference-implementation-trust plus VIP depth Stake leads.
Win-chance configuration depth at Stake Dice
Threshold-pick math interacts strongly with target value:
- T = 98 (max safe): win chance 98 percent, payout 1.0102x. Slow grinding shape at lowest variance.
- T = 75: win chance 75 percent, payout 1.32x. Mild variance.
- T = 50: win chance 50 percent, payout 1.98x. Balanced.
- T = 25: win chance 25 percent, payout 3.96x. Higher variance.
- T = 10: win chance 10 percent, payout 9.9x. High variance.
- T = 5: win chance 5 percent, payout 19.8x. Very high variance.
- T = 1: win chance 1 percent, payout 99x. Lottery shape.
- T = 0.1: win chance 0.1 percent, payout 990x. Extreme lottery.
- T = 0.01 (min): win chance 0.01 percent, payout 9,900x. Pure lottery.
- All configurations target 99 percent RTP. Win-chance amplifies variance without changing RTP.
- Martingale interaction at 1 percent edge: per-step expected loss is moderate. The Martingale is still negative-EV, see open the audit.
The variance shape moves dramatically across configurations; the RTP target stays at 99 percent throughout.
Where Stake Dice sits across operator coverage
| Brand | Verified Dice RTP | Notable feature | Catalogue position |
|---|---|---|---|
| Duel | 100 percent (within allowance) (verified) | 0 percent house edge target | Audit-set leader on raw RTP |
| Rollbit | 99.5 percent | RLB token rakeback overlay | Second on raw RTP |
| Betfury | 99.28 percent | BFG dividend token | Third on raw RTP |
| Stake | 99 percent (verified) | Earliest-mover operator, deepest VIP track, no token utility, 2017 operational track record | Cluster baseline reference |
| Shuffle / Yeet / Winna / Gamdom | Cluster at 99 percent | Standard uniform threshold pick | Standard cluster |
| Fairspin TFS / Roobet | 97 percent / expected 96-98 percent | Different operator-strategy framing | Lower raw RTP |
How Stake Dice compares to other Stake originals
- Stake Dice: the verified 99 percent RTP threshold pick. This page.
- Stake Plinko: the verified 99 percent RTP binomial bucket-drop (1000x max); see the binomial bucket-drop teardown.
- Stake Crash: the verified 99 percent RTP reciprocal multiplier curve; see the reciprocal-curve teardown.
- Stake Mines: the verified 99 percent RTP 5x5 grid; see the grid press-your-luck teardown.
- Stake Limbo: the verified 99 percent RTP target multiplier (1,000,000x max); see the target-multiplier teardown.
- Stake HiLo: the verified 99 percent RTP card-sequence pick; see the card-sequence teardown.
- Stake Dragon Tower: the verified 99 percent RTP tile climb; see the tile-climb teardown.
- Stake Slide: the slider-multiplier mechanic; see the slider-multiplier teardown.
- Stake Scarab Spin: the Egyptian-themed slot-style original; see the slot-style teardown.
- Stake Keno: the number-pick lottery; see the number-pick lottery teardown.
- Stake Wheel: the wheel-spin segment pick; see the wheel-spin teardown.
- Stake Blue Samurai: the samurai-themed arcade-style original; see the arcade-style teardown.
- All 12 share the HMAC-SHA256 fairness primitive. 7 of 12 have publicly-verified 99 percent RTP; 5 have RTP pending.
Stake's dice mathematics against the responsible-gambling line
- The verified fairness layer doesn't change the structural house edge. 1 percent house edge produces $1 expected loss per $100 wagered.
- Dice auto-bet sessions typically generate the highest per-session bet volume of any Originals mechanic; even at 99 percent RTP the cumulative erosion compounds rapidly.
- Variance dominates session outcomes regardless of RTP. Martingale doubling at any target is bankroll-survival math, not positive-EV math.
- T=0.01 lottery shape at 9,900x is extreme variance regardless of favourable RTP. Lottery is not strategy.
- VIP-tier-progression incentive may encourage higher session volume than baseline.
- If gambling has stopped being fun, neither operational track record nor VIP depth changes the situation. Free, confidential help: GamCare and BeGambleAware. Our player-protection limits page lists brand-side limits worth setting.
Frequently asked questions about Stake Dice
What is the Stake threshold pick in one sentence?
The Stake threshold pick is the canonical uniform 0-100 Dice mechanic at Stake, a 2017-launched Curaçao-licensed crypto-casino (earliest-mover operator), with verified 99 percent RTP, HMAC-SHA256 fairness via published reference documentation, and no welcome bonus or token utility (deliberate brand stance offset by deepest VIP ladder).
How does Stake dice fairness work?
Stake uses HMAC-SHA256 fairness with operator-committed server seed, player-controlled client seed, per-roll nonce, and cursor index. The byte derivation maps to a uniform 0-100 float.
What is the verified Stake dice rtp?
99 percent base (1 percent house edge), verified per Stake published documentation. Cluster baseline reference.
How does stake Dice vs Duel Dice, what is different? work?
Both share canonical uniform 0-100 threshold-pick mechanic, identical HMAC implementation. Duel dice at 100 percent (within allowance) RTP leads on raw figure; Stake dice at 99 percent is cluster baseline.
Does Martingale work better on Stake Dice because of low edge?
Martingale has the same negative-EV outcome at any positive house edge. Lower edge extends bankroll-survival distance; it does NOT change the direction. See more on this for the full math.
What is the maximum payout at the highest variance setting?
T = 0.01 (minimum win chance 0.01 percent) returns payout multiplier 9,900x.
Why does Stake offer no welcome bonus on Dice?
Deliberate brand stance across all Originals.
How does the Stake Reload program affect dice play?
At higher VIP tiers (Platinum-plus), the Stake Reload supplements the base 99 percent RTP with weekly, monthly, and level-up bonuses. Effective return improves with tier progression.
Continuing reading: Stake Dice and adjacent topics
- For the binomial bucket-drop reference, read background.
- For the reciprocal multiplier curve build, read extended notes.
- For the 5x5 grid press-your-luck, read audit details.
- For the 1,000,000x target multiplier, read the breakdown.
- For the card-sequence pick, read deeper context.
- For the tile climb with difficulty modes, read further reading.
- For the slider-multiplier mechanic, read more context.
- For the Egyptian-themed slot-style original, read the rundown.
- For the number-pick lottery, read the technical notes.
- For the wheel-spin segment pick, read supporting analysis.
- For the samurai-themed arcade-style original, read the audit excerpt.
- For the cross-brand Martingale critique, read the related piece.
- For our editorial methodology, see our editorial methodology page.
Authority sources cited in this Stake Dice review
- The Stake publicly verifiable implementation page is the published reference.
- The Curaçao eGaming registry confirms Stake's licensed status under Medium Rare N.V.
- that page and that page provide independent player-protection guidance.
