Cryptocurrency markets as an evolving complex system: stylized facts, scaling, and complexity

Cryptocurrency markets trade around the clock and are fragmented across many independent exchanges. Heterogeneous participants, leverage, and algorithmic order execution interact in these markets, while much of the settlement process is recorded transparently on public blockchains. Cryptocurrency returns are heavy-tailed and display the stylized facts familiar from conventional financial markets: weak linear autocorrelation, volatility clustering, long memory, and multifractal scaling. These properties, however, should not be treated as fixed, universal constants. Evidence from Bitcoin and other major cryptocurrencies shows that scaling exponents, entropy measures, and market-efficiency indicators evolve with institutionalization and market regime and depend on sampling conventions. This Perspective argues that cryptocurrency econophysics should move beyond cataloguing patterns in price time series toward testing the mechanisms that generate them and assessing their predictive value. We propose five research priorities: (i) separating genuine multifractality from finite-size and distributional artifacts; (ii) validating crash early-warning indicators in real time and out of sample; (iii) identifying the endogenous clock of a market that never closes; (iv) measuring systemic risk with dynamic, multilayer, and higher-order networks; and (v) using time irreversibility as a signature of non-equilibrium dynamics. An integrated design combining price, order-book, derivatives, and on-chain data can turn stylized facts from post hoc descriptions into reproducible measures of market regime, fragility, and resilience. Because cryptocurrency markets are unusually observable, they also offer a natural laboratory for testing statistical-physics concepts such as universality, finite-size scaling, critical transitions, and entropy production in a real socio-economic system.