Martingale is a betting scheme older than financial markets: after every loss, increase the stake so that the next win repays the whole streak plus a small profit. Applied to trading it produces something genuinely seductive — an equity curve that climbs in a smooth staircase for months, because every drawdown is bought back by escalating size before it ever appears on the chart. On paper the system cannot lose. With infinite money, it truly can't. You do not have infinite money.
The danger is not that martingale loses often. It is that it loses rarely, invisibly, and completely — and that until the day it does, its track record looks better than almost anything honest.
The math of doubling
Start at 0.01 lots and double after each loss: 0.02, 0.04, 0.08, 0.16, 0.32, 0.64, 1.28. Seven consecutive losses — a streak every real strategy experiences — and the eighth trade must be 128 times the original size, with roughly 255 units of cumulative exposure committed to recovering what the first 0.01-lot trade lost. The account is now one adverse move from a margin call, in order to win back one small stake.
Grid systems and "recovery zones" are the same engine with better marketing: positions accumulate against the trend, average price improves, and everything resolves profitably — until the one trend that doesn't come back.
How to spot it in a track record
- A win rate above 85% combined with a smooth, almost unnaturally steady equity curve.
- Average losses many times larger than average wins — or, more suspiciously, almost no visible losses at all.
- Floating drawdown: open losing positions held for days or weeks while closed results stay green.
- Lot sizes in the history that grow in bursts — 0.01, 0.02, 0.04 — rather than staying proportional to the account.
- A track record shorter than a year, or one that conspicuously restarts after a gap.
The one question every seller should answer
There is nothing dishonest about escalation mechanics in themselves — some professional strategies use controlled, capped scaling with full disclosure. The dishonesty is in hiding the tail risk behind the pretty curve. So put one question to every system vendor, including us: what is the worst losing sequence this strategy can produce, and what happens to the account when it arrives? A serious answer names sizes, drawdown, and a stop condition. Anything else — "the algorithm adapts," "risk is managed dynamically" — is the sound of a bomb ticking politely.