A result appears here only when it has been measured.

There are no published results yet. This page will not be filled with anything that has not been observed first-hand — a plausible-looking number is worse than a blank, because a blank can still be filled correctly.

Published results

No public results published yet.

No experiment has produced a measured outcome that meets the publication standard. Until one does, this section stays empty by design.

What a published result will contain

When a result is published it will be a complete record, not a highlight. Each one will carry:

The hypothesis as it was stated

Written before the test started, so it cannot be adjusted afterwards to fit the outcome.

What was actually observed

First-party measurements with the period they cover and the method used to collect them.

The decision and its reasoning

Continue, hold or stop — recorded with the date and the reasoning that justified it.

What was learned either way

A stopped experiment is published with the same rigour as a successful one.

The publication standard

Evidence before narrative

A public claim must be traceable to evidence, and evidence has to be observed rather than inferred. Nothing is published here to make the page look complete, and no figure is estimated, rounded up or carried over from another experiment. Where something is unknown it stays unknown.

See the method for the lifecycle states, the decision vocabulary and how evidence is ranked.