The D’Alembert is the progression for players who find the Martingale terrifying and the flat bet dull. Named after the 18th-century mathematician, the D’Alembert system raises and lowers the stake one unit at a time, producing a calm, slow climb that is easy to follow and easy on a bankroll.
How the D’Alembert system works
The rule set is a single sentence: add one unit to your stake after a loss, subtract one unit after a win. You play an even-money bet and never drop below your base unit. The idea rests on an old, flawed intuition — that over a session wins and losses roughly even out, so stepping down after wins and up after losses should leave you slightly ahead. It feels reasonable, and in a short, balanced session it often delivers a small profit.
A worked example
Here is how the D’Alembert system escalates the stake across a losing streak from a $10 base — notice how much flatter the climb is than doubling.
| Loss # | Next stake | Total staked |
|---|---|---|
| 1 | $10 | $10 |
| 2 | $20 | $30 |
| 3 | $30 | $60 |
| 4 | $40 | $100 |
| 5 | $50 | $150 |
| 6 | $60 | $210 |
| 7 | $70 | $280 |
| 8 | $80 | $360 |
The running total staked climbs to $360 after 8 losses in a row. The house edge is unchanged the whole way down.
Eight losses in a row takes the stake to $80, not the $1,280 the Martingale would demand. That linear growth is the whole point: the D’Alembert trades the Martingale’s fast, fragile recovery for a slow, survivable one.
Why it still loses
The assumption at the heart of the D’Alembert — that wins and losses balance out — is not true over any fixed session. Each spin is independent, the wheel has no memory, and the 2.70% house edge is deducted on every bet no matter how you size it. A run of losses leaves you with a raised stake and no guarantee the wins will come to unwind it. The D’Alembert is a pleasant, low-stress way to structure a session, and an honest account ends there: it manages variance, it does not beat the wheel.
A full session, spin by spin
The escalator above only shows a losing streak. Real sessions mix wins and losses, so here is a ten-spin run on red at a $10 base unit, tracking the stake, the result, and the running total. The stake climbs one unit ($10) after every loss and drops one unit after every win, never going below the $10 floor.
| Spin | Stake | Result | Running total |
|---|---|---|---|
| 1 | $10 | Loss | −$10 |
| 2 | $20 | Loss | −$30 |
| 3 | $30 | Win | $0 |
| 4 | $20 | Loss | −$20 |
| 5 | $30 | Loss | −$50 |
| 6 | $40 | Win | −$10 |
| 7 | $30 | Win | +$20 |
| 8 | $20 | Loss | $0 |
| 9 | $30 | Win | +$30 |
| 10 | $20 | Win | +$50 |
Five wins and five losses, and the session ends $50 ahead. That is the pattern people fall for. When the count of wins matches the count of losses, the D’Alembert leaves you up by roughly one base unit per win, because each win lands at a higher stake than the loss it cancels. It looks like the maths is on your side. It is not. The profit only appears because this particular sample split evenly and finished on a downswing in stake. Shuffle the same ten results into a front-loaded losing run and you would be staring at an $80 stake with a deep deficit and no guarantee of the wins arriving to claw it back. The sequence, not the system, decided the outcome.
Where it helps and where it hurts
The genuine strength of the D’Alembert system is bankroll survival. Because the stake grows by addition rather than multiplication, a bad patch bleeds you slowly instead of wiping you out. Compare that to the Martingale, where eight straight losses force a $1,280 bet off a $10 base and slam into the table maximum. The gentle climb keeps you in your seat longer, and for a player who mostly wants a structured, unhurried session, that is worth something.
The weakness is the flip side of the same coin. Slow escalation means slow recovery. If you hit a stretch with more losses than wins — which happens constantly, because the results are random — the system carries a raised stake for a long time and never fully unwinds it. You are not protected from losing, only from losing fast. And the whole thing rests on wins and losses evening out, which they are under no obligation to do over any session you will actually sit through. The house edge keeps skimming its cut the entire time, so even a perfectly balanced run of wins and losses is quietly taxed.
That makes the D’Alembert system a reasonable fit for a cautious player with a modest bankroll who treats roulette as entertainment and wants a rule to follow rather than a way to win. It is a poor fit for anyone hoping to grind out a reliable profit, or anyone who cannot comfortably absorb a stake that keeps ratcheting up through a cold streak.
Mistakes that quietly sink the system
The most common error is chasing a losing streak past the point of comfort. The linear climb feels safe, so players convince themselves that one more raised bet will square the books, and the stake drifts far above where they meant to stop. A raised stake is still a raised stake, and the wheel owes you nothing.
The second mistake is picking a base unit that is too large for the bankroll. The D’Alembert only survives a rough patch if the base unit is small enough that a dozen escalating bets barely dent your funds. Start at 5% of your bankroll and a single unlucky run ends the session. A base unit around 1% leaves room for the progression to breathe.
A third trap is believing a long run of one colour makes the other “due.” It does not. Each spin is independent, and a system that adjusts stakes off past results is reading a pattern that is not there. The D’Alembert can organise your betting, but it cannot forecast the next number.
European versus American tables
Where you play changes the maths more than which progression you run on top of it. A single-zero European wheel carries a 2.70% house edge on even-money bets. The double-zero American wheel adds a second green pocket and pushes that edge to 5.26% — nearly double the drag on every stake. Run the D’Alembert system on an American table and the same progression loses ground faster, because the built-in cost of each bet is higher and the extra zero tips a few more spins against you over time.
French tables sweeten the even-money bets further. Under the la partage rule, an even-money bet only loses half its stake when the ball lands on zero, which cuts the effective edge on red/black, odd/even and high/low to about 1.35%. If you are committed to a progression like the D’Alembert, the single-zero European table is the sensible floor and a French table with la partage is better still. None of this turns the system into a winner — the edge is smaller, not gone — but a smaller edge means the wheel takes its cut more slowly, which is the only lever a player actually controls.
See how it compares to the Fibonacci and the rest of the roulette betting systems.
Frequently asked questions
How does the D'Alembert system work?
You raise your stake by one unit after a loss and lower it by one after a win. It is a negative progression, but a linear one, so it climbs far more slowly than the Martingale's doubling.
Is the D'Alembert safer than the Martingale?
Yes. Adding a single unit at a time keeps the stakes and the risk much lower during a losing streak. The trade-off is that recovery is slower and less certain.
Does the D'Alembert win over time?
No. It assumes wins and losses roughly balance, but each spin is independent and the 2.70% edge is charged throughout. Long streaks break the assumption and produce a net loss.
What is the Contra-D'Alembert?
The reverse: raise after a win, lower after a loss. It behaves like a milder positive progression, similar in spirit to the Reverse Martingale.
What base unit should I use for the D'Alembert system?
Keep it small — around 1% of your session bankroll. The progression only survives a losing streak if the stake can climb a dozen steps without draining your funds, so a base that is too large ends the session on the first cold run.
Does the D'Alembert work better on a European or American wheel?
European. A single-zero wheel charges a 2.70% edge on even-money bets, while the double-zero American wheel charges 5.26%. The higher edge drags the progression down faster, so the European table (or a French table with la partage at 1.35%) loses ground more slowly.