FinEvo: From Isolated Backtests to Ecological Market Games for Multi-Agent Financial Strategy Evolution

Conventional financial strategy evaluation relies on isolated backtests in static environments. Such evaluations assess each policy independently, overlook correlations and interactions, and fail to explain why strategies ultimately persist or vanish in evolving markets. We shift to an ecological perspective, where trading strategies are modeled as adaptive agents that interact and learn within a shared market. Instead of proposing a new strategy, we present FinEvo, an ecological game formalism for studying the evolutionary dynamics of multi-agent financial strategies. At the individual level, heterogeneous ML-based traders-rule-based, deep learning, reinforcement learning, and large language model (LLM) agents-adapt using signals such as historical prices and external news. At the population level, strategy distributions evolve through three designed mechanisms-selection, innovation, and environmental perturbation-capturing the dynamic forces of real markets. Together, these two layers of adaptation link evolutionary game theory with modern learning dynamics, providing a principled environment for studying strategic behavior. Experiments with external shocks and real-world news streams show that FinEvo is both stable for reproducibility and expressive in revealing context-dependent outcomes. Strategies may dominate, collapse, or form coalitions depending on their competitors-patterns invisible to static backtests. By reframing strategy evaluation as an ecological game formalism, FinEvo provides a unified, mechanism-level protocol for analyzing robustness, adaptation, and emergent dynamics in multi-agent financial markets, and may offer a means to explore the potential impact of macroeconomic policies and financial regulations on price evolution and equilibrium.

💡 Research Summary

FinEvo introduces an ecological game formalism for evaluating financial trading strategies, moving beyond the conventional practice of isolated back‑tests in static market environments. The authors argue that traditional evaluation fails to capture the interdependencies among strategies that shape real‑world market outcomes. To address this, FinEvo models a population of heterogeneous agents—rule‑based, deep‑learning, reinforcement‑learning (RL), and large‑language‑model (LLM) traders—interacting within a shared, continuously clearing double‑auction market.

At the individual level, each agent i at time t selects an action aᵢ,ₜ from a policy πₖ conditioned on recent price history and exogenous signals (news, macro indicators, etc.). The market aggregates all actions and stochastic shocks ξₜ to produce the next price Pₜ₊₁ via a deterministic matching function F. Agents update cash and asset holdings accordingly, and the realized payoff for a strategy k is computed as the average wealth change of its adopters.

Each strategy maintains an internal parameter state Ψₖ,ₜ that evolves through a generic adaptation operator A(·), which can represent policy‑gradient updates for RL, back‑propagation for deep nets, prompt‑tuning for LLMs, or heuristic adjustments for rule‑based agents. In addition to short‑term rewards, the authors attach a forward‑looking value process Vₖ(t) modeled as an Ornstein‑Uhlenbeck (OU) process: dVₖ = λₖ