§ 09Errata

Things I got wrong.

Learning isn't always about getting the right answer. Sometimes the most valuable lesson comes from discovering that an idea was wrong.

Erratum 01Behaviour2026-04

I thought loss aversion was something that affected other investors, not me.

After researching: I noticed I check a falling number far more often than a rising one, and that I had been describing exactly this behaviour as irrational when other people did it.
What changed: I stopped treating behavioural biases as a topic to write about and started treating them as a description of my own defaults. Writing a decision down in advance is now the only defence I actually trust.

Erratum 02Probability2026-03

After researching: I ran it as a blind test on 60 charts and scored 33 — statistically indistinguishable from guessing. My confidence was not correlated with being correct.
What changed: I no longer accept 'it looks like a pattern' as evidence, including from myself. The experiment is written up in the Research Notebook.

Erratum 03Mathematics2026-02

After researching: Working through the formula properly, time sits in the exponent and everything else sits in the base. Extending the horizon from 20 to 30 years does more than raising the rate by two points.
What changed: I reordered how I think about the whole subject: horizon first, cost second, rate third, starting amount last.

Erratum 04Programming2026-05

After researching: My compound interest function passed its first test because I had chosen a 0% rate, which is the one case where the bug cancels out.
What changed: I now write the test case where the answer is different for the right and wrong implementations, and I check at least one value by hand.

Curiosity, humility and critical thinking are easier to claim than to evidence. This page is the evidence.