Why are your worst decisions clustered at the end of a long session?

Why this matters. This lesson is unusual: the two famous studies you have heard cited for it do not hold up, and the honest version of the answer is thinner and more useful than the popular one.

Most people recognise the pattern in themselves. Late in a session, decisions get faster, looser, and less like the ones made that morning.

The two explanations you have been given, and why we do not use them

Removed from this curriculum, not softened: ego depletion — the idea that self-control is a finite resource drawn down by use — was for years the standard explanation. A preregistered replication across 23 laboratories returned an effect of d = 0.04 with a confidence interval spanning zero. We do not teach it as a mechanism.

Also removed: the "hungry judges" study, in which parole decisions appeared to depend on time since the last break. It is the most quoted finding in this area and the magnitude has been seriously challenged, with later work arguing much of it could be a statistical artifact of case ordering. It is not a trading lesson. It is a lesson in how a seductive result travels.

If you have read about decision fatigue in a trading context, you have almost certainly been reading one of those two.

What survives

Less, and enough. You can measure it in yourself. Whether your late-session decisions differ from your early ones is a fact about you, and it is available without any theory at all — it requires only that decisions carry a timestamp.

This is the domain's method in miniature: where the general claim is not supported, the personal version is often still measurable, and the personal version is the one you can act on.

The artefact

A timestamp on every decision, and session start time recorded once a day. Nothing else — no rating, no mood, no theory.

After a few weeks you can sort by hours-into-session and look. You may find nothing. Finding nothing is a real result, and it is more than the literature can currently give you.

Try it now

Add the time to your last decision record, and write down when that session started. Two numbers, and the measurement has begun.