Evaluating Technical Trend Filters for a Bollinger Band Breakout-and-Reversal Strategy Across Multiple Timeframes Empirical comparison of seven technical trend filters across five BTCUSD timeframes 15 October 2026
Year: 2026
Type: preprint
Abstract: This study evaluates whether technical trend filters improve a Bollinger Band breakout-and-reversal strategy on BTCUSD across multiple chart timeframes. Seven filters—EMA, EMA200, Supertrend, moving-average crossover, MACD Trend, ADX Trend, and Renko HH-HL—are compared with an unfiltered baseline at five timeframes: 1 minute, 3 minutes, 5 minutes, 15 minutes, and 1 hour. The 40 configurations are evaluated using exported historical backtest metrics, including closed-trade net profit, profit factor, trade count, commission, and maximum intrabar drawdown. The moving-average crossover filter produces the strongest relative improvement over its matched baseline at each tested timeframe and reduces trade frequency by approximately 82–85%. Nevertheless, all filtered configurations remain net-negative at 1, 3, and 5 minutes. At 15 minutes, the moving-average crossover is the only filtered configuration with positive net profit; at 1 hour, the moving-average crossover, EMA, and Renko HH-HL configurations are positive. This pattern is consistent with the possibility that directional confirmation reduces exposure to unsuccessful short-horizon signals, but false breakouts were not separately labelled or directly measured. The study is exploratory and limited to one instrument, one approximately one-year historical period, fixed parameters, and 40 configurations. No untouched out-of-sample evaluation is documented, and the positive results are sensitive to drawdown and outlier-trade concentration. The results therefore indicate historical associations, not guaranteed or generalizable profitability. JEL Classification: C53; G14; G17
Source: Research Square
Language: English
Subfield: Finance
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7237213781
