PVE.trade serves prediction markets, not option chains. Bullish=YES-side, Bearish=NO-side. “Outcome Pressure” ≈ PCR, “Liquidity/Depth” ≈ OI — honest analogs of real PVE fields.
Top live signals
Smart money — top traders
Volume spikes — entry/exit signals
Hourly activity
Fast Signal Scanner
All ranked signals
Outcomes are measured from live-captured prices after each signal fires (no future data used at signal time). Fills in as time passes and your history grows.
Performance by score bucket
Walk-forward stability (1h)
metric across sequential time foldsEvaluated signals
Which features — and which combinations — actually precede favourable moves in your captured history. A feature "fires" when it is present, strong, and (if directional) aligned with the signal. Read against the baseline; small samples mean nothing.
Per-feature forward performance
Feature combinations (does stacking add value?)
Recent API calls
Raw / normalized
select an endpoint
Signal component weights
Engine
How the score works. Each component is computed only from PVE.trade data and returns a signed direction × strength.
Net = weighted average of available directional components. Signal Score = |Net| × data-coverage × 100 — high scores need
multiple real signals to agree and enough data present. It is a ranking/confidence score, not a probability of profit,
and nothing here is financial advice. Prediction-market analogs (Outcome Pressure ≈ PCR, Liquidity/Depth ≈ OI) are labelled honestly;
no option chain is invented.