If anyone has ever sold you a trading system, the first thing they showed you was the percentage of winning trades. It is the most quoted metric in the industry and one of the easiest to manipulate.
You do not even need to cheat. You just never close a loser.
The case we measured
We studied a commercial reference bot on data identical to ours. The result:
- +$10,129 realised
- 98.69 % winning trades
- Total realised loss: only $1,809
With those three numbers on a slide, anyone signs.
The fourth number: $64,077 of floating loss carried along the way.
That 98.69 % win rate is the hole staying open. It is not a system that wins 98 % of the time. It is a system that refuses to acknowledge a loss and therefore almost never books one.
How the trick works
The mechanism is a grid that averages down. You enter, price goes against you, you add another position lower, price goes further against you, you add another. Your average price keeps dropping. When the market bounces even slightly, the whole set closes green at once.
And it closes green for real: every order books as a winner. The percentage is not lying about what it measures. It is lying about what you think it measures.
What you do not see in that percentage is the stretch where you sat with six open positions and half the account equity gone. If the bounce arrives, it never shows up in the statistics. If it does not, it shows up exactly once and takes everything.
The numbers that actually matter
Our own measured configuration is not exempt. It closes 96.6 % of its orders green. And here is the pair of numbers that percentage hides:
- Average win: $1.03
- Average loss: $12.90
One loser eats twelve winners. At that ratio the system needs to be right more than 92 % of the time just to break even. It is right 96.6 %, so expectancy comes out positive — but the margin is far thinner than "96.6 % winners" suggests.
Always measure these three together:
- Net dollars, computed as closing balance minus opening balance minus deposits. Not the percentage.
- The average-win / average-loss ratio. If it is 0.16, you already know what kind of system you have.
- Equity drawdown, not balance drawdown. Balance has no idea what is floating.
The definitive test
We built a faithful replica of that reference bot and ran it over 2.4 years. It lost 99.99 % of the account.
It was still closing 97 % of its cycles in profit right to the end.
Written while building Cerberus at Tuurt Labs. High-risk instrument: trading leveraged CFDs can cost you your entire capital. This is not investment advice.