Fractal-Matching Forecasts with ATR Trend-Flips and Scaled Exits

Summary

This strategy combines an ATR trailing-stop trend signal with a historical pattern-matching filter. At a trend flip, it compares a recent window of normalized log returns with earlier windows, selects the closest match, and checks the subsequent returns of that historical episode for directional agreement. Trades can be limited to long or short positions, and the agreement filter can be disabled. A related ghost-candle projection visualizes the matched episode but does not drive orders.

Risk and exits adapt to volatility: the strategy defines a unit of risk using ATR and its percentile rank, then places partial profit targets at successive multiples and a stop at a configurable multiple. The document specifies assumed capital, trade size, commission, slippage, and bar-close processing, but supplies no actual backtest results. Historical similarity is not a reliable forecast by itself; outcomes may vary across markets and timeframes, and results depend on realistic transaction-cost assumptions.

Key ideas

• The strategy finds historical windows whose normalized log returns most closely resemble recent price action.
• The direction of the returns following the best match can filter ATR trailing-stop trend flips.
• Ghost candles illustrate the matched episode and do not affect trade decisions.
• ATR and its percentile rank scale the risk unit used for targets and stops.
• The document reports backtest assumptions but no performance evidence, and historical matches may not persist.

Tags

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