Multi-Timeframe Adaptive KDJ Signals with Trend Filtering
Summary
This strategy combines KDJ readings from three configurable timeframes, defaulting to 1, 5, and 15 minutes. It computes a weighted average of each timeframe’s K, D, and J values, smooths them with exponential moving averages, and uses the smoothed total to classify the broader trend. The swing length can be selected from volatility-based, inverse-volatility, or fixed-length methods, although the source code assigns a constant length for the inverse-volatility option. Signals use K/D crosses and J thresholds that vary with the trend classification; anticipated signals use the recent direction and proximity of K and D. Entries also require extreme values of the smoothed total, and exits combine an ATR-based trailing stop with a fixed stop relative to the open.
The document describes the rules and lists a short published BTC/USDT futures backtest window, but gives no performance results or evidence that the signals are profitable. The strategy may be sensitive to its thresholds, weighting, and parameter choices; its own discussion also flags overfitting and weaker reliability in extreme conditions. The anticipated signals should be treated cautiously, since the code plots them with a prior-bar offset while entry conditions use the current signal.
Key ideas
- The strategy blends KDJ values from three timeframes using configurable weights and EMA smoothing.
- Swing length can be selected by volatility, inverse-volatility, or fixed-length settings, though the source sets inverse-volatility length to a constant.
- Trend classification changes the J thresholds used with K/D crossovers to form buy and sell signals.
- Entries require additional extreme-value conditions, while exits combine an ATR trailing stop with a fixed price stop.
- The published backtest settings contain no performance results, so they do not establish strategy effectiveness.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.