The mechanisms underlying strategic interaction between algorithmic and human traders in modern financial markets remain inadequately characterized within a unified framework. To uncover the behavioral regularities and strategic dynamics of heterogeneous agents

trading behavior is conceptualized as a quantifiable "trading language

" with both trader types systematically characterized across six dimensions: order submission frequency

cancellation rate

order size

price aggressiveness

and others. Drawing on high-frequency trading data from a major equity market spanning 2022–2024

an evolutionary game model is constructed to capture the strategic co-evolution of human and algorithmic traders

from which evolutionary equilibria are derived under varying market volatility regimes. Empirical results reveal significant differences across all six behavioral dimensions: algorithmic traders exhibit substantially higher order frequency and cancellation rates

whereas human traders display greater order-size variability and intraday temporal entropy. Across low-

medium-

and high-volatility regimes

empirical strategy distributions align closely with theoretical evolutionary equilibria. Further analysis indicates that significant excess returns emerge only when the market deviates from equilibrium

with fundamentalist strategies among humans and market-making strategies among algorithms yielding the most pronounced returns. An identification test exploiting an exogenous regulatory shock in March 2023 confirms that the strategy distribution reconverges to evolutionary equilibrium within approximately 20 trading days following the shock. The findings demonstrate that the strategy distribution of heterogeneous traders does not follow a random walk but constitutes a payoff-driven process converging toward an evolutionarily stable equilibrium. Integrating the trading language framework with evolutionary game theory provides a testable theoretical foundation for understanding market dynamics in the algorithmic era

offering implications for market regulation and algorithmic risk management.