Crash Betting Systems

Each round of a studio crash game returns the same share of whatever is staked on it, at any stake and any target. A system controls only those two choices, and it makes them from rounds already finished, which carry no information about the next one. Probability theory has a name for what follows, Wald's identity: over any session, the expected loss is the house edge applied to the sum of every stake, however the bets were sized, ordered or stopped.

D'Alembert puts the rule into numbers. Played for 1,000 rounds at 2x from a bankroll of 100, raising the stake by one unit after a loss and cutting it by one after a win, a session puts 2,188 units through the game on average and finishes 66.8 units down. That loss is 3.06% of the total staked, the game's 3% edge give or take the luck of individual sessions.

The rule: A system decides how much is staked and how the results fall, never the price of each unit staked. More money through the game means more lost on average.

The originals odds page states the same principle for in-house crypto games. Studio games obey it because the studio fixes their return round by round, and the sections below put a figure on each plan.

Each plan starts from 100 units, opening with one-unit bets, on a game returning 97%, except where the plan names a larger bankroll. A series ends at its first win, or when the bankroll can no longer cover the next stake. The right-hand column never changes: the plans differ in how often they fail and how badly, not in what they pay for each unit put through the game.

The martingale doubles the stake after every loss and returns to the base stake after a win, so the first win of a series recovers everything lost in it plus 1.00 unit. From 100 units it can afford 6 stakes in a row, the last of them 32.00 units; one more loss would call for a stake the bankroll no longer holds.

Losing all 6 rounds at 2x happens in 1.87% of series, about one series in 54, and costs 63.00 units at once. A player who runs series after series meets that run sooner or later: the chance of completing 100 series without a bust is 15.2%, and for 1,000 series it is under 0.1%. Averaged over wins and busts, each series is worth 0.194 units below zero.

A bigger bankroll pushes the bust further away without removing it, and it makes the bust larger. With 1,000 units the loss when the run finally comes is 511.00 units, and with 10,000 it is 8,191.00, each time set against a profit of 1.00 unit per completed series. A maximum bet can end the doubling before the bankroll runs out: a limit that stops it after 10 steps leaves even the largest bankroll busting about one series in 762.

Changing the target reshapes the progression rather than escaping it. At 1.5x each new stake has to be much larger than the last, so 100 units cover only 4 steps; at 3x the stakes grow more gently and the same bankroll covers 9, but every step wins less often, and that series busts in 2.97% of attempts.

Doubling at 2x after opening with a single unit, on a game returning 97% and with no maximum bet in the way. Each extra step needs a far larger bankroll than the one before it, which is why the bust rate falls so slowly compared with the money put at risk.

The target chase is the system crash invented for itself. The stake stays at one unit throughout; the first bet aims at 2x, and every bust raises the next target by one whole multiplier, so whichever bet finally wins pays back the series and leaves one unit of profit.

Because the stake never grows, the chase looks gentler than doubling, but the targets climb faster than the chance of reaching them. With 100 units the series can run to a target of 101.00x; it lasts 5.6 bets on average and still ends with all 100 units gone in 1.2% of series. Without any limit on the bankroll, the average length of a series has no finite value at all: the long tail of the curve works against the chaser.

A ceiling on the multiplier makes extra money useless past a point. With 1,000 units the series busts in 0.1% of attempts, but on a game whose curve stops early enough to allow only 99 bets, the same bankroll busts in 1.2%, no better than 100 units would.

D'Alembert adds one unit to the stake after a loss and takes one away after a win, never dropping below the base. It is offered as the cautious alternative to doubling because stakes climb slowly. At 2x, though, a bet wins only 48.50% of the time, so losses outnumber wins over a long session and the stake drifts upward.

Across 1,000 rounds from 100 units, 94.9% of sessions reach a point where the bankroll cannot cover the next stake, and 5.0% finish ahead. The middle session ends with 10 units left of the original 100.

Fibonacci, which climbs a slower ladder of stakes after losses, and Paroli, which raises the stake after wins instead, are the same arithmetic in a different order: they move the large stakes to other moments in a session without changing what each unit costs. Offers that claim to know the next round are another matter altogether, covered on the crash predictors and signals page.

A popular plan sets a goal and quits on reaching it: run the martingale at 2x until 100 units become 150, then walk away. That plan reaches its goal in 57.8% of sessions, which reads like a winning record until the rest are counted, because each of those ends with almost nothing left.

Set against one goal, the way money passes through the game decides the odds. Flat one-unit bets at 2x turn 100 into 150 only 5.0% of the time, the martingale 57.8%, and a single bet of the whole bankroll cashed out at 1.50x 64.7%. For a goal of 200, where the single bet needs 2x, the three approaches succeed 0.2%, 38.7% and 48.5% of the time.

The order follows from the rule at the top of the page. The fewer units that pass through the game on the way to a goal, the less the edge has to work on, so the plan that stakes the least in total comes closest to the fair odds of reaching it. That describes how any game with a house edge behaves, not a reason to stake more: a single bet of the whole bankroll still loses all of it more often than it wins. Stake only money you could lose without harm.

In short: Every plan pays the same price per unit staked. A plan that sends more units through the game on the way to the same result simply pays that price more times.