PARKED · data tier
Hypothesis
Easley–López-de-Prado VPIN (volume-synchronized probability of informed trading) spikes before volatility events; fading extreme VPIN, or stepping aside, avoids adverse selection.
Math — VPIN
Over $n$ equal-volume buckets, VPIN is the average absolute buy/sell imbalance:
$$ \text{VPIN} = \frac{\sum_{i=1}^{n}\,|V^{B}_i - V^{S}_i|}{\sum_{i=1}^{n}\,(V^{B}_i + V^{S}_i)} $$
Method
Bulk-volume classification (BVC) on aggregate trades, 50-bucket VPIN. Tested as a volatility pre-warning and as a fade trigger. Blocked by classification noise on public tape.
Promising as a risk gate (high VPIN does precede vol), but bulk-volume buy/sell classification on public aggregate trades is noisy without true tick-level signed flow. Parked until a cleaner signed-flow source is sourced; not a standalone fade.
VPIN needs honest buy/sell classification. Inferring trade sign from public bars (Lee–Ready / BVC) injects enough noise to swamp the signal on fast crypto books.