Assume the role of a Senior Data Scientist. Define a step-by-step logical algorithm (pseudo-code or detailed description) for detecting subtle anomalies in high-frequency financial trading data without relying on supervised learning models. Your approach must account for seasonality, non-stationary volatility, and concept drift. Specifically, explain how you would implement an adaptive moving average combined with a statistical process control method (like the EWMA chart) to minimize false positives during market open/close volatility.