Claim-verification results

Every number on this page is read from the last published run at load time. Nothing here is typed by hand, so when a run has not been published this page says so instead of showing you an old number.

Loading the published results.

This run

Reach and latency both depend on where a run was launched from, so the vantage point ships with the numbers.

Pass rate by suite

Bar: pooled pass rate. Bracket: Wilson interval. A dormant sense is reported, not scored.

Stale tasks, probe observations, and dormant senses

A stale task is dropped from its denominator and named here, so a denominator never shrinks silently. A dormant sense is an optional extra that was not installed on the machine that ran the claim-verification suite, which is a capability fact and not a failed task.

Stale tasks

    Probe observations not classified as stale

    These outcomes stayed in the scored denominator because they did not show that the upstream document was stale.

      Dormant suites

        Reachability baseline

        A task enters a receipt-bearing suite only after a recorded query failed to surface its ground truth on the first page of a plain web search.

        OmniSeek · GitHub · Published data · Apache-2.0