Overconfidence shows up in two flavors when you trade, and they hurt in different ways. The first is overprecision, being too sure about a specific number. You say Bitcoin closes between $80,000 and $90,000 when an honest range would run from $40,000 to $120,000. The second is overplacement, quietly believing you're better than the average trader. Almost everyone thinks that about themselves, which of course can't be true for everyone at once. Both push you toward positions that are too big, stops that are too tight, and risk math that pretends the future is narrower than it is.
The way researchers measure this is simple. They ask people for a confidence interval on some estimate, then check how often the real answer lands inside it. If you're well calibrated, your 90 percent intervals should contain the truth about 90 percent of the time. In practice most people's 90 percent intervals catch it 50 to 60 percent of the time. We are wildly overconfident about how precisely we know things.
Why this wrecks your position sizing
Bad calibration flows straight into bad sizing. Say you think a trade has a 70 percent chance of working and the real number is 55 percent. Any sizing framework, Kelly or otherwise, will tell you to bet too much, because you fed it a lie. That gap between felt probability and real probability is enough to turn a strategy that should make money into one that bleeds it out through oversized bets.
The good news is calibration responds to practice and feedback. Here's an exercise I like. At the start of each week, write down predictions for five market variables. Where Bitcoin closes, which way the S&P moves, whether a specific data release beats expectations, that kind of thing. Put a confidence number on each one, 60, 70, or 80 percent. At the end of the week, score yourself. Do it for a couple of months and you'll see whether your 70 percent calls actually come in 70 percent of the time. Most people find their 70s are landing closer to 50 or 55.
What you do about it
The answer isn't to stop predicting. It's to widen your intervals and shrink your positions to match reality. If your 70 percent calls hit 55 percent of the time, either relabel them as 55 and size for that, or sharpen your analysis until the accuracy catches up to the confidence. Both are fine. Pretending the number is 70 is the only bad option.
Watch yourself hardest right after a winning streak. Success feeds confidence, and in trading that extra confidence usually shows up as a position size bump that quietly hands the recent gains back. The market has no memory of how well you just did. The odds on your next trade are independent of the last five. Treating each trade as its own thing, instead of feeling bulletproof after a run, is one of the cheaper forms of insurance you can give yourself.
A decision journal is what makes any of this stick. Record the prediction, the confidence level, and the reasoning, not just the outcome. After six months you can look back at your calls with their confidence tags and see the miscalibration in the data, which is a lot harder to argue with than a vague sense that you're pretty good at this. The journal is the feedback loop.
One last gut check before you enter anything. Ask what odds you'd take if you were betting against your own trade with a friend. If you wouldn't accept 3-to-1 against it, meaning you think it fails more than 25 percent of the time, your confidence is probably reasonable. If you wouldn't even take even money against it, you think failure is nearly impossible, and you're almost certainly overconfident. Framing it as a real bet with a friend on the other side makes you a lot more honest, because now being wrong costs you something.