Featured test case · GW150914 · LIGO

One famous event · two different claims

The same 2015 event can support claims of different strength. The physical-signal detection and the categorical source attribution therefore receive separate, time-indexed records.

Why this pair matters

Detecting a gravitational-wave signal is not the same inferential task as establishing exactly what produced it. The pair was selected before the verdicts were known.

11 February 2016

GW-D · Detection

PASS

Direct source claimScope · exact GW150914 detection claim

Question one: was there a physical gravitational-wave signal rather than an accidental coincidence from detector noise or a local instrumental/environmental artifact?

What claim was tested?

On 14 September 2015 at 09:50:45 UTC, the two LIGO detectors observed a transient gravitational-wave signal; the reported false-alarm rate was less than 1 event per 203,000 years, corresponding to >5.1σ.

What evidence state applied?

11 February 2016 · coordinated public GW150914 release. Only evidence publicly available by this date enters the run.

Comparator: accidental coincidence from detector noise; separately, local instrumental or environmental artifacts.

What was the result?

PASS. No applicability-relevant residual remains for this exact frozen detection claim.

Why?

Independent searches, empirical time-shift background, calibrated two-detector coincidence, model-minimal reconstruction and instrument/environment checks leave no applicability-relevant residual for this exact detection claim.

What does this not mean?

PASS does not certify that LIGO or general relativity is true, nor that every later criticism is wrong. It means only that this frozen detection claim is applicability-closed at the stated evidence date.

What would trigger a new run?

Materially changed evidence or practice, or a claim-constitutive reformulation. Later evidence creates a new run; it does not rewrite this one.

11 February 2016

GW-A · Source attribution

OPEN

Direct source claimScope · categorical source-class attribution

Question two: what licenses the stronger step from the detected signal to the categorical claim that the source was the coalescence of two black holes?

What claim was tested?

GW150914 was produced by the coalescence of two black holes.

What evidence state applied?

11 February 2016 · the same coordinated public release and the same event. Only evidence public by this date enters the run.

Comparator: ordinary compact-object alternatives assessed in the release plus the source-acknowledged exotic compact-object branch, for example boson stars.

What was the result?

OPEN. One localized contrast deficit remains for the exact categorical source claim.

Why?

The frozen source set strongly supports the black-hole-merger reading against the modeled alternatives, but it also acknowledges exotic compact-object alternatives without a GW150914-specific failure-sensitive discriminator for that branch.

What does this not mean?

OPEN does not mean that the black-hole-merger interpretation is false, doubtful, or retrospectively refuted. It records one localized contrast deficit for the exact categorical source claim at 11 February 2016.

What would trigger a new run?

A claim-preserving, target-specific discrimination of the acknowledged exotic branch when such a procedure is materially available. A narrower source-class claim would instead be a new token and must be run separately.

Later debate · audit context only

Critiques and reanalyses published after 11 February 2016 may be linked for transparency, but they did not enter either verdict above. They are post-t_R material. Any formal use of later evidence requires a new run.

Expert extension · GW-P parameter inference

Verdict: OPEN.

The expert-only third token concerns the exact published mass and distance posterior summaries. Its localized issue is prior-sensitivity; it is not a generic objection to Bayesian inference and does not say that the published parameter values are wrong.

Detection source ↗Properties analysis · exotic caveat (R04) ↗Evidence & source seal →

← Back to all examples