Blockchain Metrics and Indicators in Cryptocurrency Trading
Juan C. King, Roberto Dale, José M. Amigó
TL;DR
This study develops blockchain-based trading ribbons by extending the Hash Ribbon to 21 blockchain metrics derived from Bitcoin’s network data and evaluates their ability to predict price directions and support profitable trading. It combines nonlinear functional-dependence analysis using the Chatterjee coefficient $\xi_n$, regression-based adjustments (e.g., AdCPTRA), and derivative-based enhancements to create a family of blockchain ribbons. Through backtests and predictive-model experiments (Random Forest and LSTM with ribbon-derived features), the authors show that several ribbons deliver a statistical advantage for long trades and that certain adjusted indicators, especially AdCPTRA, markedly improve performance and predictive usefulness. The findings suggest that public blockchain metrics can yield actionable information in the highly volatile cryptocurrency market, supporting an AMH perspective when combined with traditional price-based signals and machine-learning approaches. $RMSE$ and $MASE$ results indicate that nonlinear, texture-based features enable better generalization, with LSTM performing best when percentage-based inputs are used.
Abstract
The objective of this paper is the construction of new indicators that can be useful to operate in the cryptocurrency market. These indicators are based on public data obtained from the blockchain network, specifically from the nodes that make up Bitcoin mining. Therefore, our analysis is unique to that network. The results obtained with numerical simulations of algorithmic trading and prediction via statistical models and Machine Learning demonstrate the importance of variables such as the hash rate, the difficulty of mining or the cost per transaction when it comes to trade Bitcoin assets or predict the direction of price. Variables obtained from the blockchain network will be called here blockchain metrics. The corresponding indicators (inspired by the "Hash Ribbon") perform well in locating buy signals. From our results, we conclude that such blockchain indicators allow obtaining information with a statistical advantage in the highly volatile cryptocurrency market.