Overconfidence
Estimates that are too precise and abilities that are rated too highly. It shows up in trading as too much activity and too little diversification.
MadStockAlerts Research · Updated August 28, 2026
What to take away
- Confidence intervals people give are far too narrow.
- Most people rate their own ability above the median.
- It is documented to increase after a run of good outcomes.
- In trading it produces higher turnover, which costs money.
- Studies find the most active accounts perform worst after costs.
MAD Academy Training Video · 0:45
Skill and Luck Look Identical Early
A run of wins feels like evidence of skill, and it is exactly the moment position size tends to increase.
This lesson is part of a Stock Alerts + Tools plan.
The two forms
| Form | What it looks like |
|---|---|
| Miscalibration | Ranges that are too narrow. A 90 percent confidence interval that contains the answer half the time |
| Better than average | Most people rating themselves above the median at most things |
| Illusion of control | Believing an outcome is influenced by effort when it is not |
| Hindsight | Believing an outcome was more predictable than it was |
The first is the one with the most direct application. A forecast stated as a range that is systematically too narrow produces position sizes that are systematically too large.
The documented cost
Research on brokerage account data has repeatedly found that the most active accounts underperform the least active ones after costs, and that the underperformance is driven by turnover rather than by selection.
Scroll the chart sideways to see all of it.
Why it grows after success
A run of good outcomes is attributed to skill and a run of bad ones to circumstances, which is self-attribution bias. The consequence is that confidence rises after good periods and does not fall symmetrically after bad ones.
This produces the specific pattern of increasing size after a winning run, which is when the run is most likely to have been variance and when a reversion is most costly.
What reduces it
- Recording forecasts with explicit ranges and checking how often the range contained the outcome.
- Counting trades and comparing the record against a do-nothing benchmark.
- Separating whether the rules were followed from whether the trade made money.
- Treating a winning run as a sample rather than as evidence, until the sample is large.
The first item is the only one that addresses miscalibration directly, and it is the most effective single exercise in this pillar. Stating a range in advance and scoring it is feedback that the market does not otherwise provide.
The calibration exercise
Miscalibration is the one form of overconfidence with a direct and cheap corrective, because it can be scored.
- 1State a range, not a pointFor any forecast: a range you are 80 percent confident contains the outcome.
- 2Write it down with the dateBefore the outcome, and where it can be found later.
- 3Score it afterwardsHow often did the range contain the answer, across many forecasts.
- 4Widen accordinglyIf 80 percent ranges contain the answer half the time, the ranges are too narrow.
Almost everyone who performs this exercise discovers their ranges are far too narrow, and almost everyone improves with feedback. It is one of the few biases in this pillar that responds to practice.
The application to position sizing is direct. A sizing decision rests implicitly on a range of outcomes, and a systematically narrow range produces a systematically oversized position.