VWMA Crossover Signals Filtered by MFI, ADX, and kNN

Summary

This experimental strategy combines a fast and slow volume-weighted moving average (VWMA) crossover with a k-nearest-neighbors classifier. A fast-line cross above the slow line proposes a long entry, while a cross below proposes a short entry. The classifier uses money flow index (MFI) and average directional index (ADX) values to compare current conditions with stored historical observations, then uses the neighbors’ past price direction to filter the crossover signal.

The document gives default indicator and neighbor-count settings and describes a one-month BTC/USDT futures backtest configuration, but it reports no performance results. It therefore provides a strategy outline rather than evidence that the method is profitable. The author identifies lag in VWMA, MFI, and ADX, sensitivity to the choice of k, and possible poor live performance as limitations. Validation through backtesting and paper trading is proposed, alongside parameter tuning and testing alternative features or classifiers. The described approach does not specify safeguards against data leakage or explain how historical examples are selected over time, so those details would need careful review before drawing conclusions.

Key ideas

• VWMA fast and slow line crossovers provide the initial long and short signals.
• kNN uses MFI and ADX observations to classify the direction associated with similar historical conditions.
• The classifier’s vote filters crossover signals before the strategy enters a position.
• Indicator lag and sensitivity to the neighbor count may affect the signals.
• The document supplies backtest settings but no performance results, so the strategy remains unvalidated.

Tags

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