1. Home
  2. Resources
  3. Special reports
  4. Synthetic Reality: Synthetic market data generation at scale using agent based modeling

Agent Based Modeling (ABM) for financial markets

Synthetic Reality: Synthetic market data generation at scale using agent based modeling

By Natraj Raman, Jochen L. Leidner, Krishnen Vytelingum and Geoffrey Horrell

Agent Based Modeling (ABM) offers advances over traditional model testing applications in the financial markets. In this paper we introduce the topic of Agent Based Modeling and illustrate its application in a number of scenarios.

Overview

Download this paper to:
 
  • Learn about different models for simulating financial market dynamics and performing what-if scenario analysis
  • Find out how models can replicate well-known statistical properties of the financial markets
  • See the simulation results of different scenarios e.g. when asset prices drop suddenly

 

This project allowed us to tackle the technical challenges and begin a dialogue around this emerging field of synthetic data and ABM with the financial modeling community.

Read on

Complete this form to access the full report

Key content

Four prototype agent-based simulation models

Powerful enough to produce markets in equilibrium and perform what-if scenario analysis.

Calibrated using Refinitiv’s historical tick by tick pricing data

Sourced from more than 500 trading venues and third-party contributors.

10,000 traders simulate price evolution

Trading agents interact with a stock market by posting orders at each time step to a limit order book.