Forget what we have invested in these statistical models.
Simply executing Kolmogorov-Smirnov tests is just a mechanical step; it does not realize the validity of an interpretation.
The testing process is separate from the relevance of its results, which fundamentally depend on data quality and assumptions.
For example, KS tests on gold prices in India, without considering demand spikes related to festivals or import restrictions, would give a misleading statistical similarity.
You cannot just apply a tool without auditing its application conditions with fresh eyes.