KNN Trend Following with Indicator Features and ATR Risk Controls

Summary

This strategy classifies market direction by comparing current indicator features with historical feature vectors, then voting across the nearest labeled observations. The description mentions RSI, ROC, CCI, and volume as possible inputs, normalized to a common range, and specifies Euclidean distance for neighbor selection. Predicted direction drives long or short entries, with a moving average filter, optional volatility filter, date controls, and ATR-based stop distances and position sizing tied to a stated equity risk percentage.

The document provides defaults and a BTC/USDT futures backtest configuration covering one month in late 2023, but reports no performance results. The source explicitly warns that signals repaint, a substantial limitation for evaluating historical or live performance. The accompanying text recommends tuning the neighbor count and features, trying other distance measures, and using rolling train/test separation; these are suggestions rather than validated improvements. It also notes that parameter choices can cause errors or overtrading.

Key ideas

• The classifier uses distances between current and historical indicator feature vectors to infer bullish or bearish direction from nearby labels.
• The described inputs include RSI, ROC, CCI, and volume, with a moving average filter and optional volatility filter.
• ATR-based stops and position sizing are intended to scale risk with market volatility.
• The source warns that signals repaint, and the short published backtest configuration has no reported results.
• Rolling train/test evaluation and alternative features or distance measures are proposed but not validated.

Tags

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