Both systems are negative progressions, but they sit at opposite ends of the aggressiveness scale. Martingale is the steepest; D'Alembert is among the flattest. That difference shapes everything about how they feel to play.
Quick Comparison Table
| Feature | Martingale | D'Alembert |
|---|---|---|
| After a loss | Double the bet | Add one unit |
| After a win | Reset to base | Subtract one unit |
| Growth speed | Exponential | Linear |
| Variance | High | Low |
| House edge | Unchanged | Unchanged |
Martingale Rule
Double after every loss, reset after a win, so one win recovers the whole sequence plus a base unit. The mechanics are covered in the Martingale strategy guide.
D'Alembert Rule
Raise the stake by one unit after each loss and lower it by one unit after each win. The idea is that wins and losses roughly balance over time, so the bet drifts up and down gently rather than exploding.
Progression Speed
Martingale grows exponentially; D'Alembert grows linearly. After six losses Martingale needs 32 units on the next bet, while D'Alembert needs only seven. This makes D'Alembert dramatically slower to escalate.
Recovery Behavior
Martingale recovers a whole losing run with a single win. D'Alembert only recovers gradually and assumes wins and losses will roughly even out, which they do not always do in a short session. Its recovery is partial and slow.
Bankroll Pressure
D'Alembert's linear growth means it uses bankroll far more slowly, so it rarely produces the sudden large drawdowns Martingale can. The trade-off is that it takes many more spins to recover a bad run.
Table-Limit Pressure
Because stakes climb slowly, D'Alembert almost never approaches the table maximum, whereas Martingale can hit it after a handful of losses. The choice of base bet, discussed on the initial bet sizing page, still sets the starting scale for both.
Expected Value Reminder
Slower and gentler does not mean profitable. D'Alembert places the same negatively-priced bets as Martingale, just in different sizes, so the long-run risk of ruin is lower per session but the expected value is still negative. Neither system overcomes the house edge.