Adaptive Zero-Lag EMA Crossovers with Error Filtering

Summary

This trend-following strategy compares a conventional EMA with an adaptive zero-lag EMA. The adaptive period can be estimated with cosine or I-Q instantaneous-frequency methods, or an average of both. A crossover of the adaptive line above the conventional EMA signals a long entry, while a cross below signals a short entry. The source also calculates an error measure and applies a threshold condition to crossover signals, then uses fixed-point stop and target settings with trailing exit parameters.

The document argues that adaptation may reduce indicator lag and that threshold filtering may remove some false signals, but it reports no evidence demonstrating these benefits. It flags unstable adaptive parameters and fixed exits that may not fit changing market conditions; its own discussion also notes that losses are not effectively bounded. The published example is a brief BTC-USDT futures backtest on a two-hour period. The source contains sizing inputs and risk calculations, but the strategy entries do not visibly apply the calculated lot amount, so effective position sizing is uncertain.

Key ideas

• The strategy compares a regular EMA with an adaptive zero-lag EMA whose period is estimated from signal behavior.
• It uses upward and downward line crossovers for long and short entries, subject to an error threshold.
• Fixed stop and target settings, along with trailing exit parameters, manage positions.
• The document warns that adaptive parameters can be unstable and fixed exits may not suit changing markets.
• No performance results are provided, and the source's calculated position size is not visibly passed to entries.

Tags

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