The die, the target value and the chance
Dice game: how a dice round with crypto is decided
One figure between 0 and 99.99 decides the round, and you decide yourself how large the chance should be. Here is how the target value, the chance and the payout hang together, and what long losing runs cost.
A dice game is the simplest house game you will find at a crypto casino. The game draws one figure between 0.00 and 99.99, and before the draw you have chosen whether the figure should land above or below a target value you set yourself. Hit it and you get the stake multiplied by a multiplier that depends on the chance you chose. Miss and the stake is lost.
There is no six-sided die here, even though the name and the animation can give that impression. The space of outcomes is 10,000 possible figures, and that lets you set the chance almost freely, from around 1% to well over 90%. Below we work through the arithmetic, the checking of a round, and why auto-staking with doubling does not make the game profitable. The dice game is one of several house games we work through, and the whole comparison is gathered on the home page of this site.
Dice: the roll and the result
How a dice game works: the figure, the direction and the target value
A round has four choices and one result:
- The stake in the coin your balance is held in.
- The direction: over or under.
- The target value: the figure the result should land above or below.
- Start: the game draws a figure from 0.00 to 99.99.
Choose «under 25.00» and you win if the figure comes out 0.00–24.99. That is 2,500 of 10,000 possible outcomes, so a 25% chance. Choose «over 74.99» and you win on 75.00–99.99, and the chance is also 25%. The direction plays no part in the chance, then; it is the width of the range that counts.
Most versions have a slider that moves the target value, with the chance and the multiplier beside it. Move the slider and both figures change at once. Which limits on the chance and the multiplier the game has is stated in the information window and can vary between operators.
The payout worked out
The multiplier follows a fixed formula: (100 − the house edge) / the chance in per cent. The house edge varies between house games and operators, so the figures below are an example.
A worked example with a 1% house edge:
- chance 90%: 99 / 90 = 1.10×
- chance 49.5%: 99 / 49.5 = 2.00×
- chance 25%: 99 / 25 = 3.96×
- chance 10%: 99 / 10 = 9.90×
- chance 1%: 99 / 1 = 99.00×
Multiply the chance by the multiplier and the answer is 99% every time. Wherever you set the target value, you get back on average 99 of every 100 units staked. With no house edge a 50% chance would have given exactly 2.00×; with the house edge you have to come down to 49.5% to get the same multiplier.
This is also the simplest check you can make yourself. Read off the chance and the multiplier in the game and multiply them together. If you get 0.98, the house edge is 2%; if you get 0.97, it is 3%. A target game with the same logic, but where you choose the multiplier directly instead of the chance, is in the guide to the target multiplier.
The seed and the nonce: how a dice round is checked
House games like this usually use an arrangement of three values. The server seed is made by the operator and kept secret, but the hash of it is shown before you play. The client seed is yours, and you can often change it. The nonce counts your rounds, and the figure goes up by one per game.
The figure from 0.00 to 99.99 is worked out from these three values. Switch to a new server seed and the previous one is shown in the clear. Then you can:
- hash it and check that the answer is the same as the hash you were shown,
- work out the result for each nonce with your client seed,
- compare it with your history.
If all three steps match, no round in the series was tampered with after the hash was published. The check does not show how large the house edge is, and it says nothing about the account rules or the licence. The procedure itself is described step by step in the explanation of the fairness check.
Auto-staking and losing runs in a dice game
Many players use the auto function, often with a rule that doubles the stake after a loss. The idea is that one win should recover all the losses before it. That sounds logical, but the arithmetic shows where the risk ends up.
A worked example. You play with a 49.5% chance and a 2.00× payout, start at 1 USDT and double after every loss. The chance of a loss in one round is 50.5%.
- 5 losses in a row happen in around 3.3% of cases. You have then staked 31 USDT and must put 32 USDT on the next.
- 8 losses in a row happen in around 0.42% of cases. Total staked: 255 USDT, next stake 256 USDT.
- 10 losses in a row happen in around 0.11% of cases. Total staked: 1,023 USDT.
That looks rare, but auto-staking plays many rounds. Over 1,000 rounds the chance of at least one run of 10 losses is around 41%. Each series won gives only 1 USDT of profit, while one long losing run can cost over a thousand. The 1% house edge is still there in every round; the doubling only moves the loss into rare, large blows.
If you want to use auto-staking, set a stop on the number of rounds, the total loss and the total win before you start. Most panels have fields for all three. A stop on the total loss is the most important of them, because it sets an upper limit on what an unlucky series can cost you, however long the function would otherwise have carried on. The relation between the return and the swings over time is explained in the walk-through of the house edge and the return percentage.
A high or a low chance: which suits you?
The average is the same, but the session feels completely different.
- A high chance, for example 90%: you win often, but little. One loss eats the winnings from around nine victories.
- A middling chance, around 49.5%: wins and losses alternate, and the balance swings around the starting point.
- A low chance, for example 10%: the chance of missing 20 times in a row is around 12%, so long dry spells are common.
Choose the chance according to how many rounds your amount should last, not according to how much you hope to win.
A session of 100 rounds, worked out
A worked example. You put 0.50 USDT per round with a 25% chance and a 3.96× payout, and play 100 rounds. The total stake is 50 USDT. On average you hit 25 times and get 1.98 USDT each time, 49.50 USDT in all. The expected loss is therefore 0.50 USDT, or 1%.
But few sessions end exactly there. The standard deviation for such a session is around 8.6 USDT, so a result 8–9 USDT above or below the average is entirely ordinary. Hit 30 times and you come out ahead by almost 9.50 USDT; hit 20 times and you lose around 10.40 USDT. Neither says anything about the next session. The more rounds you play, the closer the result in per cent lands to the expected loss, even though the amount in money can swing more.
A dice game with coins: the account before the stake
The game is fairly similar from one place to another, but the account rules are not. Before you deposit, check three things: what makes the operator ask for ID, at which amount it happens, and which period applies for the documents. Rocketpot gives a document threshold of $2,000–$2,500 in cl. 11.4 and 12.4, and grants you 30 calendar days to deliver (cl. 12.2). Metaspins sets no ceiling, and cl. 12.5 leaves room for larger sums to be paid out in several parts.
The answers for the twelve ranked operators, with the clause number and the date, are gathered in the ranked overview of crypto casinos. Set yourself a loss amount before the first round, too. Hjelpelinjen (hjelpelinjen.no) offers free help if play becomes hard to stop.