Blog

Evidence Beats Assumptions

2026-07-17

A confident trader with years of pattern-recognition experience and a real, checkable result can arrive at the same conclusion about a strategy. Only one of them is evidence. The other is a feeling — often a well-earned, reasonably-informed feeling, but a feeling all the same, and feelings don't survive contact with a market the way a checked result does.

Why a strong feeling isn't a substitute

Confidence scales with experience and pattern exposure, not with whether a specific pattern is real this time. A trader can be genuinely skilled, genuinely experienced, and still be wrong about a specific idea, because skill and experience don't provide a way to check any one claim — they provide a prior, a reasonable starting point, not a verified answer. The entire point of everything covered in this part of the loop — determinism, a written specification, exhaustive behavioral testing, the ability to replay a result and see exactly what happened — is to give a specific idea something more than a prior. Evidence, in this sense, isn't a stronger feeling. It's a checkable one.

What "checkable" actually rules out

A result resting on assumptions can't be interrogated the same way — there's no specific claim to verify, just a general sense that it seems right. A result resting on evidence can be challenged specifically: run it again and see if it reproduces, check the fill against the specification, replay the trade that looks suspicious, validate it against data it hasn't seen. Each of those is a real test with a real answer, not a matter of how convinced the person presenting it sounds.

Why this is worth insisting on

None of this is a claim that intuition is worthless — intuition is usually where an idea comes from in the first place. It's a claim that intuition is a starting point, not a stopping point, and that the distance between the two is exactly what determinism, specification, testing, and replay exist to cross. A trading idea's worth was never going to be settled by how confident it feels or how well it's described. It's settled by whether it survives being checked — and evidence, not conviction, is the only thing that actually survives that.


Full reference: docs · The workflow this maps onto: reamerlabs.com