Bet selection

In the game of dice, you chose a number and bet whether you will roll over or under that number. What you are really deciding is what probability you want to give your bet of winning. Betting you will roll over 66.66 is the same as betting that you will roll under 33.33, both have a 33.33% chance of winning and payout 3 m฿ for every 1 m฿ that you bet (2.97 m฿ after 1% house edge).
These are the most common numbers/probabilities to bet with.

  • 25
  • 33.33
  • 50
  • 66.66
  • 75

The number you make your betting strategy around depends on what probability is suited to your staking plan. The staking plan is more important than the number because the 1% house edge is always the same no matter what number you choose.

Patterns in previous rolls

The most common way people choose selections is by identifying patterns in previous rolls and betting on those patterns to break. The classic example is when you see your last 10 rolls have been under 50 and then start betting over 50 because it is due to happen.

Another possibility is if you observe that the last 10 rolls have been alternating over 50, under 50, over 50, under 50 and so on, you could start betting on the rolls to stop alternating. To do this you would bet over 50 after the last roll was over 50 and under 50 after the last roll was under 50 in the hope that over or under would come up twice in a row.

These concepts can be expanded to patterns of any complexity. You might notice the dice have been rolling over 75 twice in a row then under 60 3 times for 20 rolls, which is 4 cycles. Any pattern you find can be bet against, or indeed you could bet on the pattern to continue but given the rolls are random we intuitively expect the pattern to break.

Past random events have no bearing on future random events. However given that the numbers are random, patterns cannot continue indefinitely and they are fun and interesting to find and bet against.

Staking plan

Because all dice bets have the same -1% expected value, dice are all about the staking plan. The simplest staking plan is to always bet the same amount, if you do that for long enough your results should be in line with the house edge. Let’s look at some more fun and potentially profitable possibilities.


This staking plan is so popular it has its own Wikipedia page. In a martingale system, you progressively increase your bets during a losing streak. When you do win a bet the profit will cover all previous losses accumulated during the streak and give a profit of the original amount you were going for. This is the classic roulette system of doubling down on red until a red comes up.

Here is an example with dice. Bet 1 m฿ to roll over 50.
If that loses bet 2 m฿ to roll over 50.
If that loses bet 4 m฿ to roll over 50.
If that loses bet 8 m฿ to roll over 50.
and so on, if any of these bets win it will recover all previous losses plus the 1 m฿ profit you were going for with the first bet.

Where martingales come unstuck is when you have a very long losing streak and either run out of funds or reach the operator’s maximum bet. In this case, you cannot keep increasing the bets to an amount that would recover the losses. To mitigate this you can use a stop loss where once you have lost a certain amount in a streak you stop betting or change the strategy.

Variation: Martingail to break even

Instead of starting off your bet progression staking 1, 2, 4, 8 etcetera make it 1, 1, 2, 4, 8 etcetera. This means if your first bet wins you win 1, if any subsequent bet wins you break even but your stakes increase slower and you can withstand a losing streak 1 loss longer.

Variation: Inverse martingale

Most staking plans also make sense if you do the exact opposite. In the case of the inverse martingale that means increasing your bet size when you are on a winning streak rather than when you are on a losing streak.

The martingale sees you making the biggest bets on your worst losing streaks, wouldn’t you rather make your biggest bets when you are on a winning streak? Set the % of the streak accumulated winnings you wish to stake on each bet.

For example, for each bet, you will risk 50% of the streak winnings. The below table shows your stakes during a 10-roll winning streak on a 50/50 bet.

Roll Stake Amount won
1 1 1
2 .5 1.5
3 .75 2.25
4 1.125 3.375
5 1.688 5.063
6 2.531 7.594
7 3.797 11.391
8 5.695 17.086
9 8.543 25.629
10 12.081 38.443

Because you only bet 50% of your accumulated winnings, even after you lose you keep the other 50%.

Whatever amount your losing bet turns out to be is the same as the amount you win from the streak after that losing bet. So if you lost on bet 6 which had a stake of 2.531 you have kept 2.531 in winnings from the previous winning bets during that streak.

Auto-betting bots do the calculations for you

All of these staking plans can be programmed into the site’s auto-betting feature, the best of which is at Betcoin and Nitrobetting.

Generally, no series of bets with negative expectations can have a positive expectation overall. We have found 1 way to play dice with positive expectations. When Nitrogen’s jackpot is over 3650 m฿ it becomes mathematically profitable to bet because the jackpot is more than you expect to lose by betting enough that you expect to win the jackpot.

Our list of dice reviews:

18+ only. Gambling can be addictive, please play responsibly.

    Will Wood

    Verified icon for Will Wood


    A pro gambling writer since 2015, immersed in the world of crypto since 2016. I've built up a wealth of knowledge and experience in both crypto gambling and crypto betting, making me one of the most prominent voices in the industry.

    More by Will Wood Read more arrow


    Will Wood


    Updated on 16/04/2024