Overconfidence After A Winning Streak
A string of winning trades tends to feel like proof that a trader's judgment has improved, but a hot streak is frequently just variance running in a favorable direction — and mistaking one for the other is how streaks end badly.
Five, six, seven winning trades in a row produces a very specific and very natural feeling: that something has changed, that the trader is reading the market better than before, that size can reasonably go up. Sometimes that's true — genuine skill improvement happens. Far more often, a streak of that length is simply what variance looks like when it runs in a favorable direction for a while, which it will do periodically for any trader with a real edge, and just as periodically for one with no edge at all. The two situations feel identical from the inside. Only the size of the sample and a level head can tell them apart.
Why the streak itself is weak evidence
A short run of wins is weak statistical evidence of a skill change because random processes with no underlying shift produce clusters of wins routinely — a trader making the exact same decisions with the exact same edge will still see stretches where results run well above average, simply because outcomes are noisy around any true average. Treating a six-trade winning streak as proof that judgment has sharpened is drawing a strong conclusion from a sample far too small to support it, and the natural next step — increasing size because "this is working" — is precisely the moment sizing decisions are least connected to any actual change in the underlying edge.
A trader who had six consecutive winning trades on a mean-reversion setup increased position size by 50% on the seventh, reasoning that the recent run showed the setup was working particularly well right now. The seventh and eighth trades both lost, each larger than any of the prior wins, because the increased size wasn't grounded in any real change to the strategy's win rate — it was a reaction to a streak that random variation alone could fully explain. The trader's actual edge on the underlying setup, measured honestly, hadn't moved at all across the whole sequence; only the trader's confidence had.
The corrective isn't to distrust every good run — it's to ask what evidence, beyond the streak's existence, would actually distinguish "the edge improved" from "variance ran hot," and to size according to that evidence rather than the streak's emotional pull.
A short winning streak is weak evidence of improved skill because random variation around a stable edge produces clusters of wins routinely, with no underlying change required. Increasing size purely because of a recent hot streak, rather than because of real evidence the edge changed, is a common way a good run turns into a bad one.
Further reading
- Kahneman, Thinking, Fast and Slow