Source traceability
Identify origin, author or institution, publication date, exact document version, relevant conditions and primary source whenever possible. Distinguish first-party marketing from independent evaluation.
METHODS / 02
Extraordinary claims.
Ordinary traceability.
A public protocol for evidence, independent testing, transparent forecasts, corrections and conflicts of interest.
Evidence labels
Observed: directly documented. Measured: produced by a disclosed method. Interpreted: inference subject to challenge. Hypothesized: proposed but unconfirmed. Forecast: a dated, resolvable future proposition.
Forecast registration
Before resolution, record a unique ID, exact proposition, probability if quantitative, deadline, resolution source, baseline, known alternatives and publication timestamp. Archive updates rather than overwriting them.
Experimental accountability
Record model/tool versions, test data, evaluation scripts, environment, reviewer process, failure cases and whether the dataset could overlap training material.
Uncertainty disclosure
State sample and selection limits, evidence conflicts, confounders and measurement weaknesses. Avoid treating lack of evidence as proof of impossibility.
Corrections and independence
Readers may submit sourced challenges. Material changes receive a dated revision note. Editorial content does not confer endorsement; conflicts or sponsorship must be disclosed.
There are no scored forecasts in this launch edition.
Open research questions appear in the Futures Observatory, but they are not marketed as predictions. When forecasts are later added, their original publication record, changes and outcomes will remain visible.
See the observatory questions ↗