We interviewed Tobias Wrenner, a quantitative analyst who has spent eight years auditing investment decision logs at Canadian pension funds and private equity firms.
The pattern in the logs
Wrenner pulls up a consistent finding. In post-decision audits, investment teams overestimated the accuracy of their return forecasts by an average of 34% when compared against actual outcomes over a 36-month window. The teams were not incompetent. They were systematically overconfident.
This is not a soft observation. It shows up as a hard number when you compare predicted ranges against realised results.
Where overconfidence does the most damage
Wrenner identifies concentration risk as the primary consequence. When a team is highly confident in a thesis, they allocate more heavily than a diversified model would recommend. If the thesis is wrong, the loss is amplified. He tracked 18 cases across three firms where single-position concentration above 22% of portfolio value correlated with losses exceeding the firm's own stated risk tolerance.
A structural fix, not a cultural one
Wrenner is skeptical of training programs designed to make analysts less confident. He sees more value in structural constraints. Pre-set position size limits, mandatory devil-advocate documentation before any major allocation, and third-party review of high-conviction positions all reduce the operational impact of overconfidence without requiring anyone to change how they think.