User behavior on mobile devices shows a distinct "High-Affinity" signal: The Volume Spike. When a user discovers a track they love, their first physical interaction is often a significant increase in volume (30%–50%+ increase). Conversely, a decrease in volume or a "mute" during a track is a subtle negative signal that current skip-logic misses.
The Proposed Feature: "Volume-Weighted Affinity"
I suggest integrating Volume Delta (the change in volume level) as a third-layer feedback signal to refine the "Taste Profile."
Positive Reinforcement: A volume increase of >X% within the first 60 seconds of a track correlates to high emotional resonance. The algorithm should weight the metadata of this track more heavily for immediate next-track selection.
The "Environmental" Filter: To avoid noise (e.g., turning up volume because the bus got loud), the signal can be normalized using the device's microphone (Ambient Noise Floor) or Accelerometer (detecting movement). If the environment is stable but the volume goes up, the signal is "Active Interest."
Negative Refinement: A "Fade Out" (slowly lowering volume) during a track provides a more nuanced negative signal than a "Hard Skip," suggesting the track doesn't fit the current mood even if it fits the genre.