Detecting Live Trading Edge Decay with Bootstrap-Calibrated CUSUM

Summary

The Edge Drift Detector compares closing-deal outcomes from a live Expert Advisor with a backtest baseline to assess whether average per-deal performance has weakened. Its central method is a cumulative-sum alarm calibrated by resampling baseline deals, with a user-selected false-alarm probability over a planned monitoring horizon. It also estimates bootstrap tail probabilities over short and long recent-deal windows and decomposes expectancy changes into win rate, win size, and loss size effects. A health grade and operational recommendation summarize the diagnostics.

The tool can read tester reports, CSV files, or account history, with filters for magic number, symbol, and date; it can also normalize results per lot and warn about differences in symbol mix. Resampling is controlled by a seed for repeatability. The document describes software features and input choices, but supplies no validation results, calibration study, or evidence that its grades predict future performance. Conclusions depend on baseline comparability, trade selection, horizon and threshold settings, and the representativeness of observed deals.

Key ideas

• The detector compares live deal outcomes with a backtest baseline to monitor possible expectancy decay.
• A bootstrap-calibrated CUSUM test raises alarms over a chosen monitoring horizon.
• Short- and long-window tail checks complement a decomposition of expectancy changes.
• Inputs support deal filtering, per-lot normalization, and symbol-mix warnings.
• The description provides no evidence that its alarms or grades predict future strategy performance.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.