Skip to content

All updates

Designing validation to be falsifiable

A protocol fixed in advance, a pass threshold against a weather-only baseline, and a published result either way.

Published: · Written by VoltreeAI editorial

The validation programme is being designed around a protocol fixed in advance with an independent academic partner, a pass threshold against a weather-only baseline, and a published result either way. Formal registration in a public study registry is planned, not yet filed.

A protocol fixed before data collection, agreed with an independent academic partner, sets out what counts as a pass and what counts as a failure. The comparison runs against a weather-only baseline and against established reference methods.

The two-week head start remains a validation target. Formal registration in a public study registry is planned and has not yet been filed, and we will say so on this site until it has.

Related pages