Empirical Mode Decomposition for Medium-Term Trading Signals

Summary

This document presents a medium- to long-term trading method based on filtering a price series to extract oscillations, then comparing a smoothed mean with smoothed peak and trough sequences. The included implementation uses a bandpass-filter recurrence on the midpoint of each bar, tracks local peaks and valleys, and scales their averages by a fraction. The resulting comparisons determine a long or short position, with an option to reverse the direction.

The article describes the approach as a way to focus on movements above a threshold and reduce false breakouts, but it provides no performance evidence. It lists overfitting from parameter selection, slow signal formation, and difficulty handling sharp price moves as limitations. The published configuration specifies a Bitcoin futures test period without reporting outcomes. The method's behavior therefore remains unvalidated here, and its parameter settings and risk controls would need independent assessment before practical use.

Key ideas

• A bandpass filter is applied to price data to isolate oscillatory behavior.
• Moving averages of tracked local peaks and troughs provide amplitude references for the signal.
• The mean is compared with scaled peak and trough averages to set long or short exposure.
• The approach is intended for medium- and long-term holding rather than high-frequency trading.
• The article warns of parameter overfitting, slow signals, and poor handling of sharp market moves.

Tags

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