Skip to content
ML Academy

Introduced on day 1

Decision time

The moment a decision is made and acted on. Only information that exists at that moment may feed the prediction, the baseline or the rule that supports the decision.

In detail

An overbooking limit for a Friday flight may be set three days before departure. At that moment you know bookings so far, the route, the weekday and recent no-show history - not the final check-in count, and not the fare changes made the day before departure.

Fixing decision time first prevents two mistakes: using a feature that only appears later (leakage), and building a model whose prediction arrives after the decision has already been taken.

Day 2 formalizes this as the decision cutoff, with an observation window before it and a label window after it, and checks each feature's availability against it.