CounterSpoof

Counter-Spoof is the first scanner of its kind to not only detect spoofing, but also generates consistent alpha through countering spoofed orders. By identifying the telltale patterns of fake liquidity (unbalanced quoting, phantom depth, cyclical cancellations), the scanner pinpoints when manipulators are overextended and enters reversal trades positioned to profit from their forced exit. Machine learning classifiers confirm spoofing probability while an LSTM model tracks the manipulation lifecycle, ensuring you're trading with—not against—the inevitable unwind.
The system combines eight behavioural metrics (High Quoting Activity, Unbalanced Quoting, Abnormal Cancellations, Low Execution Probability, Trades Opposing Quotes, Cancels Opposing Trades, and Cyclical Patterns in Depth and Cancellations) into a unified spoofing probability score. Random Forest and Boosted Tree classifiers process these features, then merge via logistic regression to generate a real-time probability between 0 and 1. When this score exceeds 0.95, an LSTM-based RNN trained on court-validated spoofing cases classifies the current manipulation phase (stacking, reaction, or extraction), determining optimal entry timing and trade direction opposite the spoofed side.