CGANs for Responsive Synthetic Market Simulations

Summary

The paper addresses a limitation of historical replay backtests: they do not model how the market might respond to a strategy’s actions. It describes multi-agent simulation as a way to represent interactions among traders with different strategies, allowing experimental trading agents to be tested in a simulated market. A practical obstacle is that detailed historical data about individual agents is generally proprietary or unavailable, making realistic calibration difficult.

The proposed approach trains Conditional Generative Adversarial Networks on aggregate historical data to create a synthetic “world” agent that generates orders in response to an experimental agent. The generator is integrated with the ABIDES financial market simulator. The authors report extensive simulations in which their approach outperforms prior work on stylized facts intended to capture market responsiveness and realism. The description does not identify the specific facts, datasets, markets, or performance measures, and simulation results alone do not establish that a generated environment predicts live market behavior.

Key ideas

• Historical replay backtests do not model market responses to a strategy’s orders.
• Multi-agent simulation can represent interactions among traders with differing strategies.
• The method uses conditional generative adversarial networks trained on aggregate historical data.
• A synthetic world agent generates orders in response to an experimental trading agent.
• The system is integrated into ABIDES and is reported to improve realism-related stylized facts.

Tags

# Towards Realistic Market Simulations: a Generative Adversarial Networks Approach


# Towards Realistic Market Simulations: a Generative Adversarial Networks Approach









Simulated environments are increasingly used by trading firms and investment banks to evaluate trading strategies before approaching real markets. Backtesting, a widely used approach, consists of simulating experimental strategies while replaying historical market scenarios. Unfortunately, this approach does not capture the market response to the experimental agents' actions. In contrast, multi-agent simulation presents a natural bottom-up approach to emulating agent interaction in financial markets. It allows to set up pools of traders with diverse strategies to mimic the financial market trader population, and test the performance of new experimental strategies. Since individual agent-level historical data is typically proprietary and not available for public use, it is difficult to calibrate multiple market agents to obtain the realism required for testing trading strategies. To addresses this challenge we propose a synthetic market generator based on Conditional Generative Adversarial Networks (CGANs) trained on real aggregate-level historical data. A CGAN-based "world" agent can generate meaningful orders in response to an experimental agent. We integrate our synthetic market generator into ABIDES, an open source simulator of financial markets. By means of extensive simulations we show that our proposal outperforms previous work in terms of stylized facts reflecting market responsiveness and realism.

Shown in full with attribution under the source's licence. Licence: abstract CC0

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