Falsification
How do you tell in advance what result would prove a trading idea wrong, so the idea can actually be tested?
In plain words
Falsification means writing down, before the trade, the specific outcome that would show your idea is wrong. Instead of asking what would confirm it, you name the result that would force you to abandon it, and you judge the idea by whether that result occurs.
Why it matters
An idea that no outcome can contradict cannot be tested, and a losing position can always be explained away afterwards. Stating the falsifying condition in advance turns a belief into a test, and separates being wrong about a call from being wrong about the reasoning behind it.
In a real week: ETH, week of 26 January 2026
Suppose the idea stated before the week of 26 January 2026 was that ETH's weakness against BTC was temporary, falsified if ETH closed the week below BTC. ETH closed at 2270.15, down 19.41%, giving back 8.22 percentage points to BTC as the ETH-BTC ratio fell 9.25%. The condition was met, and the weekly return sat at the 5.0 percentile of 415 weeks of history.
A common mistake
Writing the falsifying condition so vaguely, or so far away, that almost any outcome can be absorbed. A condition set after the result is known tests nothing.