Simpler to read: Performing Kolmogorov-Smirnov tests does not make the concept of statistical tests inherently concrete if the data are already biased.
Parsimony dictates that if inputs are falsified, outputs are not a tangible reality.
For example, tests on data from our mortgage-backed securities market could show a “normal” distribution post-crisis, but this is only an artifact of massive interventions by the Fed, not a natural market dynamic.
The test is a tool, not proof of truth if the foundations are altered.