Digital Asset Claims Research Methodology
Tooling produces candidates. Methodology decides which candidates survive. The same three tests are applied to every output regardless of which system generated it.
Three Tests Applied To Every Output
Reproduction
A second analyst reproduces the result from the same source records without reference to the first analyst's working.
Second-source testing
The result is tested against an independent source of a different type wherever one exists.
Adversarial reading
The strongest alternative explanation is written down and tested, and survives in the report where it cannot be excluded.
Recording What Did Not Survive
Discarded hypotheses are retained in the case record with the reason for discarding them, so a reviewer can see what was considered and rejected.
Where two readings both survive, both are reported and the conflict is described rather than resolved by preference.
Limitations
Validation reduces error; it does not eliminate it. Every report states the conditions under which its conclusions would need to be revisited.