Intro
Menu
Published: 20/07/2026 | | Last updated: 21/07/2026
A Cross-Casino Study · Crypto Originals

Dice: the naked casino

Dice has no board, no pegs, no mines. Just a slider. You pick your win chance, the house shaves its cut off the top, and the rest is pure volatility. It's the plainest game in the lobby, which makes it the clearest lesson. Doubling your money is a coin flip you lose more than half the time.

The whole game, slide to set your win chance, and the payout moves the opposite way
99–100%
RTP across the six casinos
6
Casinos compared
49.5%
Win chance that doubles your money
50.5%
Of the time, that flip loses
9,900×
The far end of the dial (0.01% chance)
The premise

The whole casino, in one slider

Dice strips a casino to its skeleton. You drag a slider to set your win chance, rolling under (or over) a number from 0 to 99.99, and the house pays you a multiplier of exactly RTP ÷ your win chance. Ask for a bigger multiplier and your win chance shrinks to match. That is the entire game.

Because your win chance sits on the bottom of that fraction, it cancels out of the expected value. Every setting of the slider hands back the same slice of your stake, the RTP. There's no board to read, no lucky spot to stop. Whatever you pick, the house keeps its cut; all you're really choosing is how wild the ride is. It's the same law running under Mines, Plinko and Chicken Road. Dice just shows it with the mask off.

Fig. 01 · Field data

A week of the real thing

Dice is no relic. It's one of the most-played games in crypto. In the seven days to 7 July 2026, twenty-three casinos took 3.92 million Dice bets between them, about 560,000 a day, with the daily count barely moving all week.

And the play is astonishingly lopsided. Roobet alone took 46% of every Dice bet placed anywhere, and the top five casinos accounted for four-fifths of them. Dice has a clear home, and it's roobet.

We hold the RTP for five of these casinos (marked in cyan below). These are the dials we compare through the rest of this page.

3.92M
Dice bets in the week
~560k
Bets per day (steady all week)
23
Casinos offering it
46%
Of all bets taken by roobet alone
79%
Taken by the top five casinos
Total Dice bets per casino over the seven days (top 15 of 23). Casinos in cyan are ones whose RTP we compare below. Hover any bar for the exact count and share.
Total Dice bets per day across all casinos, holding in a tight band, 520k to 590k a day.
Fig. 02 · The dial

Every payout is the same trade

The bigger the multiplier you want, the smaller your chance of hitting it, and the two are locked together exactly. Ask for double the payout and you roughly halve your odds. Want a near-sure thing? It pays about 1.01×, barely more than your stake. Want 100×? You'll land it under 1% of the time. Slide anywhere you like along that line; it makes no difference to the house, because every point on it pays back the same 99% on average. The dial changes how the ride feels. It never changes how much you keep.

The dice dial at a 99% table: pick a win chance along the bottom, read the multiplier it pays. The ringed point is the 2× bet, a 49.5% win chance. Hover any marker for the exact pairing.
Fig. 03 · One law

How wild the ride is comes down to one thing

The one idea that ties this whole site together. How wild your ride is depends only on the multiplier you chase. Aim low, a 1.1×, and it's gentle. Small, frequent wins, shallow swings. Aim for 100× and it's savage. Long dead runs, then the rare jolt. Nothing else moves it, not the casino, not the game. A 2× on Dice swings exactly as hard as a 2× in Mines or Plinko. Dice just lets you pick the number on a slider instead of with mines or pegs.

How wildly one bet swings (per $1 staked) against the multiplier you target. The ringed 2× bet swings about ±$1 for every $1 you stake. The exact rule, √(RTP × (multiplier − RTP)), has no game-specific term, which is why this same curve fits Mines, Plinko and Chicken Road.
Fig. 04 · The price of doubling

A coin flip you lose more than half the time

Doubling your money sounds modest. It's a coin flip weighted against you. At a 99% table you need a 49.5% win chance to be paid 2×, so you're behind on 50.5% of flips. (Worse table, worse coin. 49% at BC.Game's 98%; and a rebate-backed 100% at Duel or Gamdom is the only near-even flip, at 49.95% / 50%.)

Flat-bet that coin over and over and the edge grinds you down. But how fast is set entirely by the RTP. Track the share of players still in profit as the bets pile up. At Duel's rebate-backed 100% it barely dips; at Stake's 99% it sinks to a quarter by 5,000 bets; at BC.Game's 98% to under a tenth. Same coin, same luck. Only the dial moved.

One thing this curve can't show you is that it's pure theory, a single smooth line, the exact odds averaged over a vast crowd of players. Real gambling is nothing like that smooth. The next figure takes these same odds and plays them out as actual, messy sessions, the variance the average hides.

Share of flat-stake players still in profit after N doubling bets (2×), starting from a single coin-flip bet on the left, where all three sit near 50%. Exact probabilities; log scale on bets.
Fig. 05 · The bent coin

Why "49.5%" is not a coin flip

Fig 04 gave the cold odds, the smooth average. This figure shows the variance those averages hide. The same 2× bet, played out as real, ragged sessions. A 2× feels like a coin flip, you win just under half the time, but "just under" is the whole game. Below are 30 gamblers, each flat-betting that 2× over and over, their running profit traced as a line. Pick a win chance and run it.

At a truly fair 50% the paths scatter symmetrically around the line. Some win big, some lose big, about half finish ahead. Pure noise, no drift. Switch to a casino's 49.5% and the short-run swings look identical, but a slope has appeared. Over a few hundred bets you'd never spot it; over 50,000 nearly every path is drowning. That's the trap. A fair coin can land heads 100 times running, it's just vanishingly unlikely. The casino's coin doesn't need luck. It just needs you to keep playing.

Win chance
Bets each
30 independent players, flat $1 on a 2× bet, cumulative profit/loss. Green paths finished ahead, red behind; the dashed amber line is where the house edge drags the average. Re-run for a fresh 30, the mix changes, the slope doesn't.
Fig. 06 · Try it yourself

Set the dial, then live with it

Pick a casino and drag your win chance. Watch the multiplier, the volatility and your odds move together in real time, then autobet a session at that setting and watch a $500 bankroll live it out, bet by bet. Set it to a near-certain 90% and grind tiny wins; set it to a 2× coin flip; or reach for a long shot and feel the swings.

Casino
Jump to
Bets
A $500 bankroll, flat $10 bets at your chosen win chance. Green above the line, red below; the dashed line is where the house edge pulls you on average. Rebate-backed casinos (Duel, Gamdom) drift at their 99%/99.9% table rate here, the rebate is a separate daily-capped credit.
Fig. 07 · Walking away

The only edge you have is the exit

You can't change the odds, but you can choose when to stop, and that's the whole game. The last figure showed the trap. The longer you play, the more surely you lose. Turn that around and the lesson is simple. Your best chance of being ahead is early. After a single 2× bet you're up 49.5% of the time; that's the high-water mark, and it only falls from there. So if the plan is "double my money," take your shot in as few bets as you can and walk the moment you're green. How green is your call, but every extra spin only feeds the edge.

The opposite instinct, playing it "safe" with a high win chance for small, frequent wins, is a trap dressed as caution. At a 1.1× bet you win 90% of the time, but each win is worth a tenth of a loss, so a single loss wipes out ten wins. You feel busy and lucky while quietly bleeding, and you almost never get meaningfully ahead. Bigger multipliers flip it. You win rarely, but one hit clears a long losing run. Same 99% edge on every one, just a very different ride:

1.1×
The grind
win chance 90%
a win nets +$0.10
10 wins to recover
one $1 loss
Feels safe and constant, but one loss erases a ten-win streak. You rarely climb.
The coin flip
win chance 49.5%
a win nets +$1
1 win cancels
1 loss
Even money, near enough. The cleanest shot at doubling, best taken in one move.
10×
The gambler
win chance 9.9%
a win nets +$9
1 win recovers
9 losses
Long dry runs, then a jolt that clears them. The bankroll swings hard.
100×
The moonshot
win chance 0.99%
a win nets +$99
1 win recovers
99 losses
You'll almost never hit it, but when you do, it pays for a hundred misses.

Every card is the identical 99% RTP. These aren't better or worse odds, just different-shaped roads to the same destination. Pick the ride you can stomach, keep the session short, and leave while you're up.

The verdict

What the slider really offers

1. You cannot beat the edge. This is the proof. Every setting of the slider returns the identical RTP. No game on the site makes it plainer. There is no stopping point, no pattern, no target that changes your expected result.
2. The only real choice is volatility. It's fixed by the multiplier you chase. A near-certain 1.01× is calm, a 100× is brutal, and it's the exact same risk curve as Mines, Plinko and Chicken Road.
3. Doubling is a losing coin flip. 49.5% to win at a 99% table, so you're behind on more than half your flips. It feels fair for a while precisely because the edge is smaller than the noise, but flat-bet it long enough and the outcome is certain.
4. RTP is the one lever you control. Choosing a 100% (rebate-backed) table over a 98% one is a ~50× difference in how long your bankroll survives the very same bet. On Dice, picking the house is the only decision that moves the needle.

Multiplier equals table RTP ÷ win chance, read from Stake, Duel, Roobet, Shuffle, Gamdom and BC.Game. Expected return equals the RTP at every slider setting; volatility is √(RTP × (multiplier − RTP)); "% still in profit" figures are exact binomial probabilities for flat 2× bets. Duel and Gamdom advertise ~100% via a daily-capped rebate over a 99.x% table, modelled here at the table rate, with the rebate noted separately. Field data (Fig. 01) reflects bet volumes supplied by our data partner, Gamblytix, reported directly by 23 crypto casinos over 1–7 July 2026: 3,917,835 bets in total. Model in dice_model.py.