VPIN — the Volume-Synchronised Probability of Informed Trading — is a metric introduced by David Easley and Marcos López de Prado that estimates the share of trading volume coming from informed (vs. uninformed) participants. The original method buckets equity volume into equal-size blocks, classifies each block as buy or sell flow, and tracks the imbalance over a rolling window.
ChartGEX's VPIN-inspired implementation
ChartGEX does not implement classical VPIN. Instead, it adapts the VPIN logic to options flow: 15-minute aggregated CBOE options volume (calls + puts), classified by Bulk Volume Classification using the underlying spot return as the imbalance signal. The resulting score is sigma-normalised against a 5-day time-matched baseline (same 15-minute bucket on prior sessions) and reported as a CDF percentile — safe, elevated, or toxic.
The blended index VPIN is constructed across SPY (40%), QQQ (30%), IWM (15%), and DIA (15%) to produce a single options-flow toxicity read for the broad US equity market.
VPIN_ChartGEX = σ_norm( BVC( ΔPut/ΔCall, return ) )Why it matters
Elevated and toxic VPIN readings precede many of the largest single-day moves in modern equity indices. The signal is most powerful when combined with dealer-structure detection and a negative-gamma regime — that combination has been the precondition for the majority of >2% daily moves in SPY in recent years.