1. Home
  2. Roulette strategy
  3. D'Alembert

Betting systems

D'Alembert strategy

One unit up after a loss, one unit down after a win. The D'Alembert is the calmest of the classic roulette systems. Here is how it works, where it goes wrong and what 10,000 simulated sessions show.

How the D'Alembert works

Pick a unit and an even-money bet. Start with one unit. After a loss, add one unit to your next stake. After a win, take one unit off, never going below one unit. That is the whole system.

An example with a $10 unit on red:

D'Alembert example, $10 unit on red
SpinStakeResultThis spinRunning total
1$10Lose-$10-$10
2$20Lose-$20-$30
3$30Win+$30$0
4$20Lose-$20-$20
5$30Win+$30+$10
6$20Win+$20+$30

Three wins, three losses, and you are $30 up: one unit for every win. That is the appeal. Whenever wins and losses even out, the D'Alembert is ahead, because each win is staked one unit higher than the loss before it.

The idea behind it, and the flaw

The system is named after Jean le Rond d'Alembert, an 18th century French mathematician who argued that after a loss a win becomes more likely, so it makes sense to bet a bit more. The roulette wheel disagrees. Every spin is independent, and red wins 18 times in 37 on a European wheel whatever came before.

So wins and losses do not even out. On average, losses run ahead of wins by about one spin in every 37 (the zero), and random swings push the gap much further for long stretches. Each extra loss lifts your stake by one unit, so the deficit grows in a slow, grinding way rather than a sudden one.

A losing streak with a $10 unit
Losses in a rowLost so farNext betMartingale would have lost
4$100$50$150
6$210$70$630
8$360$90$2,550
10$550$110$10,230
12$780$130$40,950

A straight losing streak is cheap compared with Martingale. The danger is the zig-zag: lose two, win one, lose two, win one. Every cycle of that pattern raises the stake by a unit and leaves you further behind, and a win at a high stake only takes one unit off.

What our simulation shows

The tester at the top starts from a seeded run: European wheel, $10 on red, $1,000 bankroll, $1,000 table maximum, stop at +$250, up to 1,000 spins, 10,000 sessions, seed 2026.

  • 69.8%sessions reached +$250
  • 30.1%sessions went bust
  • -$95.87average session
  • $190largest single bet

Sessions were short, 61 spins on average, and every one ended at the target or bust. Each session staked $3,378 on average, so the house edge predicted a loss of $91.29; the simulation lost $95.87 (2.84% of what was staked). The best session finished $350 up, because the last winning bet can overshoot the target.

Remove the target and the slow drift takes over. In our nine-system comparison (same bankroll and unit, no take-profit, up to 1,000 spins), D'Alembert finished in profit in just 5.8% of sessions and went bust in 94.1%, and it staked more per session ($22,690) than any system except James Bond.

Reverse D'Alembert

The reverse or contre D'Alembert adds a unit after a win and takes one off after a loss. It bets more while you are winning and less while you are losing, so losing runs cost little and a winning run can build a profit quickly, until the first loss starts taking it back one unit at a time. Its average loss per amount staked is the same as every other system's.

Risk profile

  • How it feels: relaxed. Stakes move by small steps and there are no dramatic doublings.
  • Where it hurts: long choppy stretches that push the stake up a unit at a time, then a lot of staking at those higher levels.
  • Table limits: almost never reached in normal sessions. In our preset run the largest bet was $190 against a $1,000 maximum.
  • Expected loss: house edge times total staked. Because it keeps playing at raised stakes, total staked climbs fast in long sessions.

D'Alembert myths

"Wins and losses must balance in the long run"

The proportions settle down over many spins, but the gap in raw numbers can grow, and on a wheel with a zero, losses always lead on average. The system needs the count to balance, which is exactly what probability does not promise.

"It is low risk, so it is a good way to win"

Low volatility is not the same as positive expectation. It loses the house edge on every unit, just more gently.

"A loss makes the next win more likely"

That was d'Alembert's own mistake. The ball does not know what happened on the last spin.

D'Alembert FAQ

Does the D'Alembert strategy work?

It cannot beat the house edge. In our seeded test of 10,000 sessions with a $1,000 bankroll, a $10 unit and a +$250 target, 69.8% of sessions reached the target and 30.1% went bust. The average session lost $95.87, 2.84% of everything staked.

Why is D'Alembert considered a low-risk system?

Its stakes grow by one unit per loss instead of doubling, so a losing streak is much cheaper than with Martingale: ten losses in a row at a $10 unit cost $550, not $10,230. The risk comes instead from long stretches where losses slightly outnumber wins, which slowly push the stake and the deficit up.

What is the theory behind D'Alembert?

It is named after the 18th century mathematician Jean le Rond d'Alembert, who wrongly believed that after a loss a win becomes more likely. If wins and losses came in equal numbers, the system would finish one unit up per win. They do not: red wins 18 times in 37 on a European wheel, and each spin is independent.

What is the reverse D'Alembert?

Also called the contre D'Alembert, it adds a unit after a win and removes one after a loss. It presses during winning runs and shrinks during losing ones. It has the same house edge as the standard version.

What is the best table for D'Alembert?

A French table with la partage, where the edge on even-money bets is 1.35% instead of 2.70% on a European wheel and 5.26% on an American wheel. The system does not change, but each unit staked costs half as much on average.

Feel the rhythm of the D'Alembert on the free roulette table, with play money and a live chart of your bankroll against the house edge. 18+.

Related