I’m not sure Sense is trying to find the “big fish” first. My view is that the issue with many of the small transitions is their lack of ubiquity - too many are similar and undistinguished in their “features” that Sense uses to identify unique fingerprints. Just to put a “face” on it, here’s a still from a Sense video back in 2018 that explained a little about their identification / categorization. It shows one slice of multi dimensional clustering of “features” used as input to identify devices. In this case, it’s a slice of Feature1 (power) and Feature17(p0 -phase). That implies two things:
- There are at least 17 features of a transition that Sense uses to identify/classify it.
- Overall power is one of those “features”. But my guess is that the population of small power transitions is far greater than the population of larger transitions, based on looking at my home’s “transition” stream. That means that the low wattage region of the cluster graph is saturated with device transitions, presumably with only small differences in some of the other features.
So maybe not Sense trying to pay attention to the “big fish”, but the “little fish” too numerous to separate. Picture trying to identify different sardines in a school of thousands, vs. different 2-3 different sharks…
The video is here: