Skip to content
All library documents

Using Autocorrelation to Filter Moving Average Crossover Trades

Article MQL5 articles
Read the original

Summary

The study examines why a moving average crossover system’s performance changes over time, using sixteen years of M15 EURUSD and GBPUSD data. It compares periods of strategy returns with ATR, Bollinger Band spread, and autocorrelation of bar price changes. The author finds that the volatility measures do not align with profitable and unprofitable periods, while autocorrelation patterns around trades show some separation between winners and losers.

An autocorrelation threshold is added as a filter to the crossover entry trigger. Reported Strategy Tester results show fewer trades and higher total profit and profit per trade for both currency pairs in the tested sample. The analysis also examines how the separation between winning and losing trades varies with the crossover’s slow SMA period. These are historical, parameter-dependent results from two pairs and one test strategy; the article does not establish out-of-sample robustness or account for all live trading costs and risks.

Key ideas

  • The study tests whether volatility indicators or autocorrelation track changing crossover performance.
  • ATR and Bollinger Band spread do not appear to match the strategy’s profitable periods in the examples.
  • An autocorrelation threshold is used to filter moving average crossover entries.
  • The reported historical tests show improved profit measures alongside fewer trades for EURUSD and GBPUSD.
  • The findings are limited to the tested pairs, period, and strategy configuration.

Tags

ForexTechnical indicatorsStatisticsBacktesting

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