Superseding a hypothesis

Sometimes a hypothesis is overtaken by events: the underlying assumption changed, or you've learned enough to write a new bet that replaces it. Rather than resolving the old one and starting from scratch, you can mark the new one as superseding the old one. The two are then linked in your dashboard and library.

Setting Supersedes at creation

On the hypothesis creation form, the Supersedes (optional) dropdown lists every hypothesis you've created. Pick the one this new bet replaces. Leave it as — None (standalone hypothesis) if there's nothing to link.

Editing Supersedes after creation

You can also set or change the supersedes link from the hypothesis detail page. The control is below the header, before the causal chain.

What it means in the system

• The old hypothesis stays as it was. If it was already resolved, its outcome is preserved. If it's still active, it's auto-marked as superseded (it stops nagging you for resolution).
• The new hypothesis carries forward a link back to the old one — visible in the detail view and the Learning Library.
• Calibration is not transferred. The new bet is a fresh prediction with its own confidence and its own outcome.

When to supersede vs. start fresh

Supersede when you'd genuinely call this the next version of the same idea — the segment is the same, the metric is the same, but you've learned enough to retune. Start fresh when the underlying intervention or audience is different. The link is a story-telling tool for the team; if it makes the story confusing, don't force it.

Previous: Extending a resolution date ←    Next: PM Dashboard →