The gentlest negative progression in the casino. Here's how it works, why equal wins and losses produce profit, and where the math eventually catches up.
The D'Alembert is what happens when you take the Martingale's core idea and turn the volume way down.
Instead of doubling after a loss, you add one unit. Instead of halving after a win, you subtract one unit. The progression goes 1, 2, 3, 4, 5 instead of 1, 2, 4, 8, 16. Linear growth instead of exponential.
The system is named after Jean le Rond d'Alembert (1717–1783), a French mathematician, physicist, and philosopher who co-edited the famous Encyclopédie with Denis Diderot. D'Alembert was brilliant in many fields. He solved the wave equation. He made lasting contributions to mechanics and fluid dynamics.
But his contribution to gambling came from a mistake.
In his 1754 essay "Croix ou Pile" (Heads or Tails), d'Alembert argued that if a coin lands on heads, the next toss is more likely to land on tails. His reasoning: nature tends toward equilibrium, so outcomes should balance themselves out.
This is wrong. It is the gambler's fallacy, stated plainly. Each coin toss is independent. The coin has no memory. A streak of ten heads does not make tails more likely on flip eleven.
Here's the irony worth sitting with for a second: one of the most popular betting systems in casino history is built on a probability error made by one of the 18th century's greatest mathematicians. D'Alembert nailed the wave equation and got probability wrong. The system persists not because his equilibrium theory was correct, but despite the fact that it wasn't.
One more thing before we get into the rules. As with every system in this series: D'Alembert does not change the house edge. Not by a fraction. It structures your bet sizing. The casino's mathematical advantage stays identical on every single wager, whether you're betting 1 unit or 15.
The rules fit on an index card.
That's it. No doubling. No complicated sequences to memorize. Just +1 or -1.
Let's walk through 14 rounds with a realistic mix of results. Nothing cherry-picked.
Result: 7 wins, 7 losses. Perfectly balanced. And you're up $70.
Notice how the system works. The three consecutive losses in rounds 1–3 pushed the bet up to 4 units. When the wins came, they landed at those higher stakes. The wins in rounds 4 and 5 were worth $40 and $30 respectively, while the early losses cost $10 and $20. Higher-staked wins recovered the lower-staked losses and then some.
This is where D'Alembert gets interesting. Let's run a session where you lose more often than you win.
Result: 8 wins, 10 losses. You lost two more bets than you won. And you're still up 5 units.
That shouldn't feel intuitive. But it's the D'Alembert's signature trick, and it's worth understanding exactly why it happens.
Most articles about the D'Alembert mention this fact in passing. "If you win as many as you lose, you profit." Almost none explain why.
Here's the mechanic.
In the D'Alembert, your bets are higher after you've been losing and lower after you've been winning. So your winning bets tend to land at higher stakes than your losing bets. If you end a session with the same number of wins and losses, the higher-staked wins outweigh the lower-staked losses.
Go back to the first example. The wins came at bet sizes of 4, 3, 3, 2, 3, 3, and 2 units (total: 20 units won). The losses came at 1, 2, 3, 2, 1, 2, and 2 units (total: 13 units lost). Same number of outcomes. Different stakes. That asymmetry is where the profit lives.
The Wizard of Odds expressed this relationship mathematically. If you have W wins and L losses in a D'Alembert session, and the losses exceed wins by D (where D = L - W), the net result is:
Net = W - D × (D + 1) / 2
What this means in practice:
Here's a concrete example straight from the Wizard's analysis. Imagine a session with 22 wins and 28 losses. D = 6. That's a rough session. You lost six more bets than you won.
Net = 22 - (6 × 7 / 2) = 22 - 21 = 1 unit of profit.
One unit. Despite losing six more bets than winning. The system pulled that off because those 22 wins were placed at higher average stakes than the 28 losses.
On a European roulette wheel, even-money bets win 48.65% of the time. Not 50%. That gap is small, but it's persistent.
Over enough rounds, losses will outnumber wins. The question is how far D grows beyond the threshold where the formula flips negative. In short sessions, the gap is usually small enough for the system to produce profits. Over long sessions, the house edge widens it, and the math stops being kind.
This is where D'Alembert separates itself from the pack.
Total wagered after 10 consecutive losses:
After 10 losses, D'Alembert has you betting 10 units. Fibonacci has you at 55. Martingale wants 512. The difference in bankroll pressure isn't subtle. It's a different planet.
The trade-off is recovery speed. Each D'Alembert win only drops the bet by 1 unit. Climbing out of a 10-loss hole takes many more wins than either Fibonacci or Martingale. Martingale erases the entire streak with a single win. Fibonacci needs two wins to step back. D'Alembert just inches back, one unit at a time.
Safety and speed pull in opposite directions. D'Alembert picks safety every time.
The Wizard of Odds ran massive simulations testing the D'Alembert on both baccarat (73 billion+ sessions) and roulette (25 billion+ sessions). The parameters: start at 1 unit, winning goal of 1 unit, tested across bankrolls of 10, 25, 50, 100, and 250 units.
The results followed a familiar pattern. Most sessions ended in a small win. A minority of sessions ended in catastrophic losses that wiped out the bankroll. The wins were frequent and small. The losses were rare and large.
The Wizard's own summary: "Like most betting systems, it usually results in a small win, at the cost of huge losses sometimes. Like all betting systems, not only can't it overcome the house edge, it can't even dent it."
There's a second finding from the Wizard's forum discussions that most D'Alembert articles miss entirely, and it's a critical one:
"Flat betting 1 will last longer than a D'Alembert starting with 1, for a very simple reason: you will expose more money to any house edge with D'Alembert than you would with a flat bet equal to the initial D'Alembert bet."
Read that again. Many players think D'Alembert is safer than flat betting because it feels structured and conservative. But because it increases bet sizes after losses, it puts more total money at risk over the course of a session. More money exposed to the house edge means faster expected losses.
D'Alembert's real advantage isn't mathematical. It's psychological. The bet sizing creates a sense of progression, of controlled response to variance. That feeling has value for a lot of players. But the raw math? A flat $10 bet every round will outlast a D'Alembert that starts at $10.
The natural home for D'Alembert. Even-money bets (red/black, odd/even, high/low) pay 1:1 and create the clean win/loss binary the system needs. European roulette (2.70% house edge) is better than American (5.26%). French roulette with La Partage (1.35%) is the best option available. If you're playing crypto roulette, check which variant the table runs before sitting down.
Player bet pays even money with a 1.24% house edge and no commission. That makes it the lowest-edge even-money bet in the casino. Crypto baccarat tables are a clean fit for D'Alembert. Avoid the Banker bet here. The 5% commission means your wins don't pay true even money, and that breaks the +1/-1 progression.
Pass line (1.41% edge) and don't pass (1.36% edge) both work. Simple even-money payouts on the base bet. The system pairs well with crypto craps if you're sticking to those two wagers and ignoring the rest of the layout.
Poor fit. Blackjack pays 3:2 on naturals, allows doubles and splits, and has variable outcomes that don't map cleanly to the D'Alembert's +1/-1 logic. If you win a hand by doubling down, did you "win one unit" or two? The system doesn't have a clean answer. Whether you're playing crypto blackjack or live tables, you're better off using basic strategy and flat betting.
The Contra D'Alembert flips the direction.
It's a positive progression. You press winning streaks gradually and scale back when you're losing. The idea is to capitalize on hot runs while keeping cold streaks cheap.
Compared to the Reverse Martingale (which doubles after wins), the Contra D'Alembert is far more conservative. You're only adding one unit per win. A streak of five wins takes you from 1 to 5 units. Reverse Martingale would have you at 32.
The structural weakness is the same one every positive progression system shares: a single loss after a winning run gives back the last (and largest) bet. Since that's always your biggest wager of the streak, it stings.
One clarification that gets muddled in a lot of sources: Contra D'Alembert is specifically the +1/-1 mirror of the standard D'Alembert. It is not the Paroli system (which doubles after wins). Different mechanic, different risk profile, different name.
Two things are true at the same time.
The system produces profits in most short sessions. When wins and losses are roughly balanced (which they usually are over 20–50 rounds), the "win at higher stakes" mechanic generates a small profit. Sessions ending at +1 to +5 units are common. They feel earned. They're satisfying.
The system cannot overcome the house edge. Over enough rounds, losses outnumber wins because even-money bets in a casino are not fair coin flips. They hit 48.65% on European roulette, not 50%. That 1.35% gap seems tiny. Over hundreds and thousands of rounds, it's relentless. The gap between wins and losses grows, the formula turns negative, and the accumulated small wins get erased by extended sessions where the math catches up.
The irony is right there. D'Alembert built this system on the belief that wins and losses naturally balance each other out. They don't. The house edge is a thumb on the scale, slight but permanent. A fair coin would make D'Alembert a money printer. Casino games don't offer fair coins.
But let's be fair to the system. Among negative progression strategies, D'Alembert is the gentlest. It will not blow up your bankroll on a single losing streak the way Martingale can. The bets grow by one unit. The losses accumulate slowly. If your goal is to stretch your entertainment budget across the longest possible session while still having a structured approach to bet sizing, D'Alembert is arguably the best negative progression system for that purpose.
It won't make you rich. But it probably won't make you broke in one night either.
Size your base unit at 1–2% of your session bankroll. A $10 unit with a $500–$1,000 bankroll is a sensible starting point. This gives you room to absorb a losing streak without the bets climbing into uncomfortable territory.
Set a profit target. +5 to +10 units per session is realistic and aligns with how the system naturally produces profits. When you reach it, leave. The temptation to keep going is the system's biggest enemy.
Set a loss limit. 15–20 units is reasonable. Without a hard stop, a long losing run will push your bets higher and higher with no exit plan. That's how a gentle system becomes a painful one.
Play European or French roulette if you can. The lower the house edge, the closer to 50/50 your wins and losses stay. That keeps D in the formula small, and that's where the system produces profitable sessions most often.
Write down your current bet level. The +1/-1 increments are simple in theory. After 30 rounds in a noisy casino (or with three browser tabs open), it's easy to lose track. One wrong bet undermines the entire sequence.
Don't combine D'Alembert with other systems mid-session. Switching strategies because you're losing is not a new strategy. It's panic wearing a different hat.
Understand the flat-bet comparison. D'Alembert exposes more money to the house edge than flat betting the same base unit. If your only goal is minimizing expected loss per session, flat betting is mathematically superior. D'Alembert's value is in session structure and the psychological satisfaction of scaled bet sizing. Know what you're trading and why.
