Anna Kim

If everyone blindly relies on KS tests without examining the context, then interpreting results becomes a risky strategy.
The realization of the methodology through test execution is conditional, not absolute, as incentives are often misaligned.
The gain for a researcher is to publish similarities, but the market can incur a high cost if the underlying assumptions are shaky.
Imagine a fund using KS tests to prove the similarity of its returns to an index, while its empirical data come from an illiquid market.
Full transparency of data and methods is required to avoid false security and maintain trust.

It is clear that the validity of any conclusion first depends on the integrity of raw data, not on the sophistication of the tool. The incentive to publish "significant" results can lead to threshold selection that optimizes apparent similarity, thus creating a false similarity when unmeasured external factors, such as an unexpected regulatory intervention, distort the distribution. The risk of a sub-optimal balance is that decisions based on these tests lead to ineffective or costly actions.

Where does it fail? The opposition between gating networks and basic methods is not a matter of inherent logic or simplicity versus complexity; it is conditional on input data quality and transparency of the weighting mechanism.
The weak point is the presumption that complexity is inherently unnecessary; a gating network can incorporate weak signals and non-quantifiable textual justifications, offering a depth that clear basic method assessments cannot reach.
The real problem is not complexity, but lack of transparency, for example, if the weighting mechanism is a black box.
If the weighting is auditable and the textual justifications add quantifiable value, it can surpass the raw simplicity of basic methods in volatile markets, such as when OPEC decisions or monetary policy announcements drive price movements.