If position size jumps after losses, the smooth curve is borrowed from one future bad streak. Check sizes, not profits.
In 2019 I paid $299 for a gold EA with a 96% win rate. It made 31% in four months. It lost 74% of the account on a Thursday afternoon when gold trended for six hours straight. The seller's page still shows the four good months.
That EA was a martingale. Most "95% win rate" bots are. It's the oldest trick in automated trading and it still sells better than anything else.
#The idea in one sentence
After every loss, the bot makes the next position bigger, usually double, so that a single win recovers everything lost before it.
On paper, that sounds unbeatable. You only need one winner. In practice, it turns a strategy with no edge into a strategy that wins small, often, for months, and then loses huge, once.
#Why the curve looks so good
Most of the time, markets do come back. A bot that doubles down on a losing trade usually gets bailed out by the next swing. So the statement shows:
- dozens of small green trades in a row
- a win rate of 85% to 95%
- a drawdown chart that barely moves
That is exactly what a buyer wants to see, which is why martingale and grid systems dominate bot marketplaces and Telegram "VIP" rooms.
The maths that matters: with doubling, seven losses in a row means your eighth position is 128 times the size of the first. Most accounts cannot fund that, so the run ends at the worst possible moment.
#How to spot it in a trade history
You do not need the source code. Look at position sizes, not profits.
- Sort trades by open time and compare each size with the one before. If size jumps after a loss and resets after a win, that is martingale.
- Look for baskets. Several positions open in the same direction at once, each one larger than the last, closed together at a small profit.
- Check the worst streak. Find the longest run of losses. Now ask what the next position would have been if the streak had lasted two trades longer.
Our free Bot X-ray does all three automatically from an MT4 statement, a CSV, or a Polymarket or Hyperliquid wallet.
#The hidden drawdown problem
Statements only show closed trades. A martingale basket can sit 40% underwater for a week, then close at a tiny profit, and the history will show a green trade. The real risk never appears on the chart until the day it closes red.
So when you read "max drawdown 6%" on a martingale bot, read it as "at least 6%".
#Is martingale ever fine?
Used deliberately, with a hard cap on the number of steps and money you can genuinely afford to lose, it is a known gamble with known odds. The problem is when it is sold as "AI", "low risk" or "consistent monthly returns". It is none of those.
Before you buy a high win-rate bot
- High win rate plus smooth curve is a reason to check sizing, not a reason to buy.
- Size that grows after losses is the fingerprint. Profits are not.
- Closed-trade drawdown understates the real risk for any averaging or martingale system.
- Run the history through a sizing check before you put real money behind it.
Changes to this article
Oct 14, 2026: reviewed and updated. First published Oct 12, 2026. Spotted a mistake? Tell us, and we will fix it and note it here.
This is education and personal experience, not financial advice. Crypto and leveraged trading can lose you money fast. We may earn from sponsors shown on this page; they never edit our conclusions. How we make money.