Martingale Trading: A Risk-of-Ruin Machine, Explained
The martingale trading strategy doubles your position size after every loss, so that the first win recovers everything lost plus one unit of profit. Borrowed from 18th-century casino betting, it exploits nothing about markets — its "edge" is purely arithmetic: if you can always double again, you eventually win back the sequence. That "if" is the whole story. This page exists as education and a warning, because martingale is the most seductive account-destroyer in trading, and every generation of traders rediscovers it right before finding out why it has a 300-year-old reputation.
How it works (the seductive part)
- Base unit: risk $100 on a trade
- After a loss: double it — $200, then $400, $800, $1,600...
- After any win: the sequence resets, and you're up exactly one base unit ($100) on the whole cycle
- The pitch: you don't need to be right often — you only need to be right once per sequence, so the strategy appears to win almost every cycle
And it does — almost every cycle. Martingale produces long, smooth streaks of small wins. The equity curve looks fantastic right up until the moment it doesn't exist.
The math of ruin
Start with a $10,000 account and a $100 base risk — just 1% per trade, apparently conservative. The doubling ladder: $100, $200, $400, $800, $1,600, $3,200, $6,400. After six straight losses you've lost $6,300 and the seventh trade demands $6,400 — more than you have left. Seven losses in a row ends the account.
How rare is that? With a 50% win rate, any given streak of 7 losses is about 0.8% — but you're not running the sequence once. Over 1,000 trades, the probability of hitting at least one 7-loss streak is effectively certain (better than 97%). The expected value never improved either: each trade still has the same edge (or lack of one) — martingale only rearranges when the losses land, trading many small wins for one total loss. And unlike a coin, markets serve correlated losing streaks on purpose: trending moves against a mean-reversion entry produce exactly the conditions that chain 7+ losses. Grid and "DCA" bots that add fixed size into losers are soft martingales and fail the same way, just slower — the strategy commonly pitched for range markets dies on the breakout that ends the range.
Practice this on real historical charts
Replay real price action candle by candle and test your entries with a paper account. Free, in your browser, no sign-up.
Start replaying charts — freeWhat to do instead
The legitimate answer to "how do I size positions" is fixed fractional sizing: risk a constant percentage — commonly 1% — of current equity per trade, with a stop-loss defining that risk. Losing streaks then shrink your bet size instead of exploding it: ten straight 1% losses leaves you at ~90.4% of the account, fully alive and able to keep trading. Pair that with a strategy that has an actual per-trade edge — verified over 50+ trades using a proper backtesting method — and let expectancy compound. If you're drawn to martingale because your entries keep losing, the fix is a better entry model (a trend-following approach like swing trading or pullback trading), not a bigger next bet.
Common mistakes (for those who try it anyway)
- "I'll cap it at 3 doublings" — a capped martingale is just a strategy that regularly loses 7 units; the cap removes the ruin and the illusion of always winning.
- Trusting a smooth backtest — a martingale test that never hit the fatal streak is a sample-size failure, not a validation.
- Confusing it with averaging into a planned position — scaling in with a fixed total risk and a stop is position building; doubling because you're losing, with no stop, is martingale.
- Running it on leverage — leverage shortens the ladder; on 10× futures the fatal streak is 3–4 trades, not 7.
FAQ
Does the martingale strategy work in trading?
It "works" in the sense that it wins most sequences — while guaranteeing that a sufficiently long losing streak, which becomes near-certain over enough trades, takes the entire account. Its expected value is never better than the underlying entry's, and its risk of ruin approaches 100% with time. That combination is why no risk-managed fund uses it.
Why do martingale bots show such good results?
Because the catastrophic streak hasn't happened yet in the period shown. Martingale returns are a picked-up-pennies profile: months of smooth gains, then one drawdown larger than all of them. Judge any martingale track record by asking what a 7-trade losing streak would have done to it.
What is the safest alternative to martingale?
Fixed fractional sizing — a constant ~1% of equity risked per trade with a hard stop — combined with a positive-expectancy setup you've verified yourself. You can experience the difference risk-free: run the same entries with martingale sizing and 1% sizing in the free chart replay tool and watch which account survives the first bad week.
Put it into practice — risk-free
Reading about trading only gets you so far. Replay real historical charts candle by candle and paper trade your setups — free, in your browser.
Start replaying charts — free