Position size increased just 1.05× after losses — barely noticeable. But combined with a 207% tilt drop and losers held nearly 3 days versus 24 hours for winners, that small escalation produced a 43% drawdown on a system with a 4.31 profit factor.
Most traders don't notice it. The lot size goes up by a fraction after a loss. Not doubling — that would be obvious martingale. Just a slight increase. Barely visible on a trade log. Completely invisible on a results page.
But when that small escalation compounds with other behavioral patterns, it becomes the first domino in a chain that can take down an account.
We analyzed a signal provider with a track record that most traders would envy:
Growth: 3,278%
Win rate: 81.2%
Profit factor: 4.31
Subscribers: 32 active
Capital following: $101,000
Algo trading: 91%
Trades analyzed: 2,136
A profit factor of 4.31 is rare. An 81.2% win rate is exceptional. This system genuinely makes money. The behavioral scan confirmed a real structural edge with a profitability score of 83/100.
And yet, the maximum drawdown was 43%.
How does a system this profitable produce a drawdown that large? The answer was in the behavioral data.
Loss Response: Position size increases 1.05× after losses. Subtle but confirmed.
Tilt: 100/100 (Critical). Normal trades average +$2.38. After 2+ consecutive losses: -$2.53. A 207% quality drop.
Loss Aversion: Losers held 2.87× longer than winners. Average winner: 1,452 minutes (~24 hours). Average loser: 4,172 minutes (~69 hours, nearly 3 days).
Each of these patterns alone is manageable. Together, they form a feedback loop:
A loss occurs. Position size increases by 5%. The next trade is now slightly larger. If that trade also loses, the system enters tilt — per-trade performance inverts from +$2.38 to -$2.53. Meanwhile, losing trades are being held nearly three days (compared to 24 hours for winners), giving them more time and room to grow.
The lot size increase was only 1.05×. Barely noticeable. But combined with tilt and loss aversion, that small escalation is the first domino in a chain that produced a 43% drawdown on a system with a 4.31 profit factor.
Revenge trading is talked about constantly in trading education. "Don't increase your size after a loss." But nobody measures it. Nobody can tell you whether you're actually doing it, by how much, and what happens when it interacts with your other behavioral tendencies.
A 5% size increase after a loss doesn't look like revenge trading. It looks like noise. But the statistical test confirms it's a pattern, and when mapped against the tilt and loss aversion data, the mechanism becomes clear. The system doesn't blow up because of one bad trade. It blows up because a small sizing escalation triggers a cascade of compounding behavioral responses.
Check your own data. Export your trade history and look at your average position size after a winning trade versus after a losing trade. If the number after losses is even marginally higher, you have a loss response pattern. It doesn't need to be dramatic to be dangerous — it just needs to exist alongside other behavioral tendencies that amplify it.
The question isn't whether you revenge trade. The question is whether your sizing after a loss is statistically different from your sizing after a win. The data will tell you.
The loss response cascade connects directly to tilt susceptibility and loss aversion. For the full picture across 101 accounts, see our aggregate findings.
Upload your trade history. The behavioral scan measures loss response patterns and five other dimensions. Free scan available — no signup required.
Scan Your Trades →