About this question
Medium · backtesting · Quant Researcher interview question · backtesting, signal-processing, cusum, event-driven, afml
The symmetric CUSUM filter is a technique for sampling financial data based on significant cumulative movements. It works by tracking cumulative log returns and triggering an 'event' when a predefined threshold is breached, either upwards or downwards. In quantitative finance, this method is used to create event-based bars for model training, reducing noise and avoiding oversampling during periods of low volatility. Task Implement the function solution(prices: listfloat, threshold: float) -> li