debug-pnl

Diagnosing market maker PnL losses requires a structured workflow that evaluates drag from spread, adverse selection, inventory costs, and fees, then routes step-by-step to the relevant component. It assesses attribution, edge metrics, and calibration metrics to guide fixes.

Adverse selection and inventory drag in market making PnL are caused by miscalibrated models and regime-specific failures. Identifying these loss drivers involves checking calibration metrics like Brier score and information ratio to determine if model miscalibration is at fault.

Fixing negative PnL from spread and fees involves systematically identifying the drag sources and routing to the relevant diagnostic components. A reproducible validation workflow checks attribution and edge metrics to guide the necessary fixes.

Use Brier score and information ratio to check model calibration for PnL when diagnosing regime-specific failures or model miscalibration. These calibration metrics determine if miscalibrated models are the underlying fault driving negative PnL results.

Attribution and edge metrics validate market maker PnL fixes by providing a reproducible workflow to confirm loss drivers are resolved. Evaluating these analytics modules alongside Sharpe ratios ensures the applied fixes effectively resolve spread, adverse selection, and inventory drag.

The best way to identify loss drivers in market maker PnL data is running a diagnostic workflow that evaluates spread, adverse selection, inventory costs, and fees. It assesses attribution and edge metrics to route step-by-step to the relevant component for fixes.