Linear Feedback Control for Filtering Moving Average Breakout Trades

Summary

This article adapts closed-loop control theory to a moving average channel strategy. The baseline buys when price breaks above a channel formed from moving averages of highs and lows, and sells on a break below. A controller first observes 90 days of market prices, account balance and equity, indicator readings, and position types. It fits a linear model to successive snapshots, then predicts whether a prospective trade will be profitable and pauses trading when it forecasts a loss.

In a five-year backtest on daily EUR/USD data, the article reports higher profit, fewer positions, lower gross losses, a higher Sharpe ratio, a larger share of profitable trades, and improved expected payoff after adding the controller. It presents these results as evidence that system identification can help a strategy respond to its own performance. The observation window was chosen arbitrarily, however, and the reported backtest does not establish that the model will generalize to live markets. The author identifies nonlinear system identification as a possible next step.

Key ideas

• A moving average channel strategy enters long or short positions when price breaks above or below the channel.
• The controller records system snapshots during an initial 90-day observation period before intervening.
• A linear model maps successive snapshots to predict whether the next trade will be profitable.
• The controller blocks trades it predicts will lose, changing the strategy’s exposure to market conditions.
• The reported EUR/USD backtest improved several performance measures, but does not establish live-market results.

Tags

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