Get real: realism metrics for robust limit order book market simulations

@article{Vyetrenko2020GetRR,
  title={Get real: realism metrics for robust limit order book market simulations},
  author={Svitlana Vyetrenko and David Byrd and Nick Petosa and Mahmoud Mahfouz and Danial Dervovic and Manuela M. Veloso and Tucker Hybinette Balch},
  journal={Proceedings of the First ACM International Conference on AI in Finance},
  year={2020}
}
Market simulation is an increasingly important method for evaluating and training trading strategies and testing "what if" scenarios. The extent to which results from these simulations can be trusted depends on how realistic the environment is for the strategies being tested. As a step towards providing benchmarks for realistic simulated markets, we enumerate measurable stylized facts of limit order book (LOB) markets across multiple asset classes from the literature. We apply these metrics to… 

Figures from this paper

Towards realistic market simulations: a generative adversarial networks approach
TLDR
A synthetic market generator based on Conditional Generative Adversarial Networks (CGANs) trained on real aggregate-level historical data that can generate meaningful orders in response to an experimental agent is proposed and integrated into ABIDES, an open source simulator of financial markets.
Learning who is in the market from time series: market participant discovery through adversarial calibration of multi-agent simulators
TLDR
A novel two-step method is proposed to train a discriminator that is able to distinguish between “real” and “fake” price and volume time series as a part of GAN with self-attention, and then utilize it within an optimization framework to tune parameters of a simulator model with known agent archetypes to represent a market scenario.
CTMSTOU driven markets: simulated environment for regime-awareness in trading policies
TLDR
A novel stochastic process is introduced to model the fundamental value perceived by market participants: ContinuousTime Markov Switching Trending Ornstein-Uhlenbeck (CTMSTOU), which facilitates the study of trading policies in regime switching markets.
Multi-agent reinforcement learning in a realistic limit order book market simulation
TLDR
A model-free approach by training Reinforcement Learning (RL) agents in a realistic market simulation environment with multiple agents for optimal execution in an RL setting in which an intelligent agent can make order execution and placement decisions based on market microstructure trading signals in HFT.
ABIDES: Towards High-Fidelity Multi-Agent Market Simulation
TLDR
The design of the ABIDES simulator is introduced and its use and configuration is illustrated with sample code, validating the environment with example trading scenarios and the utility of ABIDES for financial research is illustrated through experiments to develop a market impact model.
Towards a fully rl-based market simulator
TLDR
A new financial framework where two families of RLbased agents representing the Liquidity Providers and Liquidity Takers learn simultaneously to satisfy their objective is presented, a step towards a fully RLbased market simulator replicating complex market conditions particularly suited to study the dynamics of the financial market.
Bit by bit: how to realistically simulate a crypto-exchange
TLDR
This paper presents a multi-agent simulation capable of effectively representing a cryptocurrency exchange, by populating a high-fidelity market simulator with a set of low intelligence agents to represent distinct types of market participants.
Reinforcement Learning for Quantitative Trading
TLDR
A taxonomy of RL-based QT models is devised, along with a comprehensive summary of the state of the art, and current challenges and proposed future research directions in this exciting field are discussed.
ABIDES-gym: gym environments for multi-agent discrete event simulation and application to financial markets
TLDR
This paper proposes to use the OpenAI Gym framework on discrete event time based Discrete Event Multi-Agent Simulation (DEMAS), and introduces a general technique to wrap a DEMAS simulator into the Gym framework.
Similarity Metrics for Transfer Learning in Financial Markets
TLDR
Probabilistic Policy Reuse is used to balance the exploitation/exploration in the learning of a new financial MDP according to the similarity of the previousfinancial MDPs whose knowledge is reused.
...
...

References

SHOWING 1-10 OF 58 REFERENCES
How to Evaluate Trading Strategies: Single Agent Market Replay or Multiple Agent Interactive Simulation?
We show how a multi-agent simulator can support two important but distinct methods for assessing a trading strategy: Market Replay and Interactive Agent-Based Simulation (IABS). Our solution is
An agent based model of the E-Mini S&P 500 applied to flash crash analysis
TLDR
A zero-intelligence agent-based model of the E-Mini S&P 500 futures market is proposed, which allows for a close examination of the market microstructure and confirms the leading hypothesis for the cause of the May 6th 2010 Flash Crash.
A NASDAQ Market Simulation - Insights on a Major Market from the Science of Complex Adaptive Systems
TLDR
The book describes the agent-based model of the Nasdaq stock market, including strategies used by market-makers and investors, market participants interactions, and impacts of rules and regulations and presents conceptual foundations for modeling markets as complex systems.
A Calibration Procedure for Analyzing Stock Price Dynamics in an Agent-Based Framework
TLDR
A calibration procedure for validating of agent based models and shows how the parameter values of the calibrated model are important in interpreting the trader behavior in the different markets investigated, and how this new approach improves the model ability to predict market prices.
High frequency trading strategies, market fragility and price spikes: an agent based model perspective
TLDR
A novel agent-based simulation for exploring algorithmic trading strategies is proposed and is able to reproduce a number of stylised market properties including: clustered volatility, autocorrelation of returns, long memory in order flow, concave price impact and the presence of extreme price events.
Simulating Security Markets in Dynamic and Equilibrium Modes
An asynchronous discrete-time model run in “dynamic mode” can model the effects on market prices of changes in strategies, leverage, and regulations, or the effects of different return estimation
A re-examination of the “zero is enough” hypothesis in the emergence of financial stylized facts
In recent years, a growing literature has claimed that the market microstructure is sufficient to generate the so-called stylized facts without any reference to the behaviour of market players.
Spoofing the Limit Order Book: An Agent-Based Model
TLDR
An agent-based model of manipulating prices in financial markets through spoofing is presented: submitting spurious orders to mislead traders who observe the order book and it is shown that a market with HBL traders is spoofable, in that a spoofer can qualitatively manipulate prices towards its desired direction.
ABIDES: Towards High-Fidelity Market Simulation for AI Research
TLDR
ABIDES, an Agent-Based Interactive Discrete Event Simulation environment designed from the ground up to support AI agent research in market applications is introduced and its use and configuration is illustrated with sample code, validating the environment with example trading scenarios.
The predictive power of zero intelligence in financial markets.
TLDR
This work uses data from the London Stock Exchange to test a simple model in which minimally intelligent agents place orders to trade at random and demonstrates the existence of simple laws relating prices to order flows and suggests there are circumstances where the strategic behavior of agents may be dominated by other considerations.
...
...