Classifying Optimal Rebalancing Frequency for Crypto Pairs Trading

Summary

The study applies machine learning to estimate an optimal rebalancing frequency for pairs trading. It uses minute-level prices for 50 large cryptocurrencies from Binance during 2022 and 2023. For each asset pair, a simulated pairs-trading algorithm identifies a profit-maximizing rebalancing frequency. Portfolio mean, variance, skewness, kurtosis, value at risk, and correlation serve as model inputs. Pairs are grouped by correlation, and six classifiers predict one of four frequency categories.

Reported results vary by correlation group and horizon: negative-correlation pairs are classified most accurately, while positive-correlation pairs are weakest in the cited validation. The study also reports that its pairs strategy outperformed passive holding in its tested short- and long-term settings. These are historical findings, not evidence of future returns. The summary does not detail transaction-cost assumptions, data-splitting safeguards, or how the frequency categories transfer to other venues and periods, so the reported accuracy and profitability may not generalize.

Key ideas

• The study predicts rebalancing-frequency categories for crypto asset pairs using portfolio statistics and correlation.
• It compares six classification methods and separates pairs into positive, weak, and negative correlation groups.
• Negative-correlation pairs show the strongest reported classification results, while positive-correlation pairs are weakest in validation.
• The study reports that its pairs-trading algorithm outperformed passive holding over the tested horizons.
• The findings come from historical crypto data and may not generalize to other markets or periods.

Tags

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