Wavelet Smoothing for Candlestick Slope Trend Following

Summary

This strategy smooths open, high, low, and close prices with repeated wavelet convolutions, then follows the direction of the filtered close. A rising value opens a long position; a falling value closes it. The described default uses a seven-coefficient Mexican Hat filter over three decomposition levels, with other wavelet choices and smoothing depths available. The method aims to reduce short-term noise while retaining responsiveness to trend changes. The document reports favorable backtest comparisons with moving averages and other wavelets, including claimed improvements in noise filtering, false signals, signal delay, and Sharpe ratio. It does not provide enough details about the test design or underlying results to independently assess those comparisons. It also describes stronger performance in directional markets and weaker results in ranges, where repeated reversals and fees can hurt returns. The stated risks include lag, parameter sensitivity, and losing streaks; position sizing and market context are presented as important safeguards.

Key ideas

• The strategy applies a wavelet filter to price data before generating trend signals.
• A rising filtered close opens a long position, while a falling filtered close closes it.
• The default configuration uses a Mexican Hat filter and three smoothing levels.
• The document describes trend markets as more suitable than sideways markets, where whipsaws may accumulate.
• Reported backtest advantages are not accompanied by sufficient detail to independently verify them.

Tags

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