Decades of judgment research show most people rate themselves above average at most skills — and feedback loops in trading make it worse, because in a rising market or a lucky run, low-skill decisions still get paid. The brain attributes wins to skill and losses to bad luck (self-attribution), so every streak writes a flattering story: I’ve figured it out. Risk management is the first thing that story deletes.
- Post-streak sizing. The position that finally hurts you is usually taken right after your best week.
- Skipped homework. “I can read this market by feel now” — the checklist that built the streak gets abandoned mid-streak.
- Prediction inflation. Confidence in directional calls rises far faster than accuracy; conviction language (“100% going up”) replaces probability language.
- Leverage creep. In F&O, overconfidence has a built-in amplifier: the same “certainty” that loosens criteria also adds lots.
Fixed size schedule. Size changes only by written rule after sustained clean execution — never inside a streak.
Probability language, enforced. The journal records “60% setup” not “sure shot” — and tracks calibration: of your “80% confident” trades, how many actually worked?
Streak protocol. After N consecutive winners, the plan mandates a review session — the exact moment confidence peaks is the moment process gets re-checked, not relaxed.
Attribution honesty. Every win in review answers one question: how much of this was my process, and how much was the market’s direction? Wins outside the plan are labelled luck — publicly, in the journal.
Overconfidence works hand-in-hand with recency bias — recency supplies the evidence, overconfidence writes the story. The containment system for both is the same: rules that do not move with mood.