Blog

• 2026-03-24

  CUSUM Filtering and Structural Breaks: What Retail Traders Miss

  CUSUM filtering solves the overlooked problem of when to sample markets — not what to trade — by triggering observations only at statistically meaningful structural breaks, making it the essential first step in the AFML pipeline before any ML prediction begins.

• 2026-03-22

  Dollar Bars: Why Volume-Based Sampling Beats OHLCV

  Raw OHLCV data is sampled at fixed time intervals, which conflates calendar noise with market activity. Dollar bars fix this by sampling on traded value.