Trend Detection from the Distribution of Bar Closes

Summary

This trend-following method classifies each bar by where its close falls within its high-low range. It distinguishes closes near the high, near the low, and in the middle using thirds of that range. Over a configurable lookback, it computes the share of bars in each category and smooths those shares with exponential moving averages. A crossover between the smoothed high-close and low-close shares triggers long or short entries.

The document gives example parameters of a 34-bar lookback and five-period smoothing, and reports backtest settings for BTC/USDT futures over a limited date range. It provides no performance figures or detailed risk controls; its general assertion of good backtest results therefore cannot be assessed from the supplied evidence. The method may produce false signals in sideways markets, depends on parameter choices, and can lag turning points. The text recommends considering stops, volatility and volume filters, support and resistance, and multiple timeframes, but these are suggestions rather than tested components.

Key ideas

• Bars are grouped according to whether their close lies near the high, near the low, or between them.
• The recent proportions of each bar type are smoothed with exponential moving averages.
• Crossovers between the high-close and low-close proportions generate directional entries.
• The method is trend-following and may generate false signals during choppy conditions.
• The supplied backtest settings do not include performance statistics or establish robustness.

Tags

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