Table of Links Abstract, Acknowledgements, and Statements and Declarations Abstract, Acknowledgements, and Statements and Declarations Introduction


Background and Related Work
2.1 Agent-based Financial Market simulation
2.2 Flash Crash Episodes


Model Structure and 3.1 Model Set-up
3.2 Common Trader Behaviours
3.3 Fundamental Trader (FT)
3.4 Momentum Trader (MT)
3.5 Noise Trader (NT)
3.6 Market Maker (MM)
3.7 Simulation Dynamics


Model Calibration and Validation and 4.1 Calibration Target: Data and Stylised Facts for Realistic Simulation
4.2 Calibration Workflow and Results
4.3 Model Validation


2010 Flash Crash Scenarios and 5.1 Simulating Historical Flash Crash
5.2 Flash Crash Under Different Conditions


Mini Flash Crash Scenarios and 6.1 Introduction of Spiking Trader (ST)
6.2 Mini Flash Crash Analysis
6.3 Conditions for Mini Flash Crash Scenarios


Conclusion and Future Work
7.1 Summary of Achievements
7.2 Future Works Introduction Introduction Introduction Background and Related Work
2.1 Agent-based Financial Market simulation
2.2 Flash Crash Episodes Background and Related Work 2.1 Agent-based Financial Market simulation 2.1 Agent-based Financial Market simulation 2.2 Flash Crash Episodes 2.2 Flash Crash Episodes Model Structure and 3.1 Model Set-up
3.2 Common Trader Behaviours
3.3 Fundamental Trader (FT)
3.4 Momentum Trader (MT)
3.5 Noise Trader (NT)
3.6 Market Maker (MM)
3.7 Simulation Dynamics Model Structure and 3.1 Model Set-up Model Structure and 3.1 Model Set-up 3.2 Common Trader Behaviours 3.2 Common Trader Behaviours 3.3 Fundamental Trader (FT) 3.3 Fundamental Trader (FT) 3.4 Momentum Trader (MT) 3.4 Momentum Trader (MT) 3.5 Noise Trader (NT) 3.5 Noise Trader (NT) 3.6 Market Maker (MM) 3.6 Market Maker (MM) 3.7 Simulation Dynamics 3.7 Simulation Dynamics Model Calibration and Validation and 4.1 Calibration Target: Data and Stylised Facts for Realistic Simulation
4.2 Calibration Workflow and Results
4.3 Model Validation Model Calibration and Validation and 4.1 Calibration Target: Data and Stylised Facts for Realistic Simulation Model Calibration and Validation and 4.1 Calibration Target: Data and Stylised Facts for Realistic Simulation 4.2 Calibration Workflow and Results 4.2 Calibration Workflow and Results 4.3 Model Validation 4.3 Model Validation 2010 Flash Crash Scenarios and 5.1 Simulating Historical Flash Crash
5.2 Flash Crash Under Different Conditions 2010 Flash Crash Scenarios and 5.1 Simulating Historical Flash Crash 2010 Flash Crash Scenarios and 5.1 Simulating Historical Flash Crash 5.2 Flash Crash Under Different Conditions 5.2 Flash Crash Under Different Conditions Mini Flash Crash Scenarios and 6.1 Introduction of Spiking Trader (ST)
6.2 Mini Flash Crash Analysis
6.3 Conditions for Mini Flash Crash Scenarios Mini Flash Crash Scenarios and 6.1 Introduction of Spiking Trader (ST) Mini Flash Crash Scenarios and 6.1 Introduction of Spiking Trader (ST) 6.2 Mini Flash Crash Analysis 6.2 Mini Flash Crash Analysis 6.3 Conditions for Mini Flash Crash Scenarios 6.3 Conditions for Mini Flash Crash Scenarios Conclusion and Future Work
7.1 Summary of Achievements
7.2 Future Works Conclusion and Future Work 7.1 Summary of Achievements 7.1 Summary of Achievements 7.2 Future Works 7.2 Future Works References and Appendices References and Appendices 2.2 Flash Crash Episodes During the 2010 Flash Crash, over a trillion dollars were wiped off the value of US equity markets in an event that has been largely attributed to the rapid rise of algorithmic trading and high-frequency trading Kirilenko et al. (2017). The base indices in both the futures and securities market experienced a rapid price fall of more than 5% in just several minutes, after which the bulk of the price drop was recovered nearly as fast as it fell. The staff from CFTC and SEC present a thorough report on what happened during the 2010 Flash Crash event (SEC and CFTC 2010). They identified an automated execution algorithm that sold a large number of contracts as the main catalyst for the flash crash. The Sell Algorithm, which was activated on the E-mini S&P 500 futures market, kept pace with the market aiming at selling around 9% of the previous minute’s trading volume (SEC and CFTC 2010). Even though no negative impact was known previously, this process triggered a cascade of panic selling by market participants that employ high-speed automated trading systems. The consequent "hot-potato" effect, where those market participants rapidly acquired and then liquidated positions among themselves, resulted in rapid and extreme price decline. Flash crash episodes have attracted attention after the 2010 Flash Crash event. Several months after the crash, the staff from regulatory authorities released a report that highlighted the important role of a large seller in initiating the flash crash event (SEC and CFTC 2010). It is reported that although high-frequency traders appear to have exacerbated the magnitude of the crash, they do not actually trigger the flash crash. Though high-frequency traders played a role in creating the so-called "hot-potato" effect, the flash crash would very likely have been avoided without the overly simplistic sell algorithm based on volume alone. There is also a lot of academic research on flash crash episodes. For example, Kirilenko et al. (2017) applied purely empirical approaches to understanding the causes of the 2010 Flash Crash. They use regression analysis on a unique dataset that is labelled with the identities of all market participants. It is demonstrated that in responding to the activity of the Sell Algorithm, high-frequency traders caused the "hot-potato" effect that exacerbated the price drop. This is consistent with the SEC and CFTC (2010) report. Paddrik et al. (2012) develop an agent-based model of the E-mini S&P futures applied to flash crash analysis. A general flash crash in price is replicated in their model. However, they only reproduce a rough shape of the flash crash price behaviour, and detailed analyses of trader behaviours and market depth are absent. Karvik et al. (2018) developed an agent-based model to analyse the flash crash episodes in the sterling-dollar forex market. They emphasize the important role of high-frequency traders in the emergence of flash crash episodes. The proposed approach in this paper is partly inspired by their work. Paddrik et al. (2017) explore how the levels of information can be used to predict the occurrence of flash crash events. Their findings suggest that some stability indicators derived from limit order book information are capable of signalling a high likelihood of an imminent flash crash event. There are also different angles of view for flash crash episodes in literature. Paulin et al. (2019) design and implement a hybrid microscopic and macroscopic agent-based approach to investigate the conditions that give rise to the "electronic contagion"[3] of flash crash events. Their results demonstrate that the flash crash contagion between different assets is dependent on portfolio diversification, behaviours of algorithmic traders, and network topology. It is also stressed that regulatory interventions are important during the propagation of flash crash distress. Menkveld and Yueshen (2019) look at the flash crash event from the perspective of cross-arbitrage. They find that the breakdown of cross-arbitrage activities between related markets plays an important role in exacerbating the flash crash event. Kyle and Obižaeva (2020) analyse price impact during flash crash events in their market microstructure invariance model. It is shown that the actual price declines in flash crash events are larger than the predicted price impact. Madhavan (2012) argues that the flash crash episodes are linked directly to the current market structure, mostly the pattern of volume and market fragmentation. He further suggests that a lack of liquidity is the critical issue that requires the greatest policy attention to prevent future flash crash events. Similarly, Borkovec et al. (2010) explicitly owe the flash crash in ETFs to an extreme deterioration in liquidity. Their results are consistent with the liquidity provision behaviour in financial markets. Golub et al. (2012) analyse mini flash crashes, which are the scaled-down versions of the 2010 Flash Crash. It is shown that mini flash crashes also have an adverse impact on market liquidity and are associated with fleeting liquidity phenomenon. The above provides various analyses for the occurrence of flash crash episodes. However, despite extensive work on analysing the flash crash episodes, the exact causes of the flash crash episodes are still not clear. In this paper, we investigate and analyse the flash crash episodes through the lens of agent-based financial market simulation. In this sense, our work is similar to the work in Karvik et al. (2018) and Paddrik et al. (2012). Nevertheless, we offer a much more extensive and detailed analysis of the simulated flash crash event which, to the best of our knowledge, is the most fine-grained analysis in current literature. Specifically, we realistically simulate the 2010 Flash Crash event in our simulation, and divide the simulated flash crash event to several phases. For each phase, detailed analyses about traders behaviours and market dynamics are presented. To the best of our knowledge, this fine-grained simulation and analysis is not reported before in literature. By dividing the whole flash crash event into different phases and examine trader behaviours and market dynamics for each phase, we shed light on the cause for flash crash events. In addition, controlled experiments under different model settings and trader behaviours are carried out in the developed agent-based simulation framework, which provide insights about how to prevent the happening of detrimental flash crash events. The basis of our agent-based model is the extended Chiarella model in Majewski et al. (2020), which comprises fundamental traders, momentum traders, and noise traders. We further divide momentum traders into long-term momentum traders and short-term momentum traders, and introduce market makers to the model. The motivation for introducing these types of traders and their interactions will be presented in section 3.1. It is shown that the proposed model is capable of generating realistic artificial financial time series. Within the framework of the proposed realistic agent-based financial market simulation, special types of agents are introduced to trigger flash crash episodes in the simulated financial market. In this way, simulated flash crash episodes are scrutinized and analysed. Authors:
(1) Kang Gao, Department of Computing, Imperial College London, London SW7 2AZ, UK and Simudyne Limited, London EC3V 9DS, UK (kang.gao18@imperial.ac.uk);
(2) Perukrishnen Vytelingum, Simudyne Limited, London EC3V 9DS, UK;
(3) Stephen Weston, Department of Computing, Imperial College London, London SW7 2AZ, UK;
(4) Wayne Luk, Department of Computing, Imperial College London, London SW7 2AZ, UK;
(5) Ce Guo, Department of Computing, Imperial College London, London SW7 2AZ, UK. Authors: Authors: (1) Kang Gao, Department of Computing, Imperial College London, London SW7 2AZ, UK and Simudyne Limited, London EC3V 9DS, UK (kang.gao18@imperial.ac.uk); (2) Perukrishnen Vytelingum, Simudyne Limited, London EC3V 9DS, UK; (3) Stephen Weston, Department of Computing, Imperial College London, London SW7 2AZ, UK; (4) Wayne Luk, Department of Computing, Imperial College London, London SW7 2AZ, UK; (5) Ce Guo, Department of Computing, Imperial College London, London SW7 2AZ, UK. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. available on arxiv [3] Electronic contagion refers to the contagion phenomenon in financial markets that results from interactions between trading algorithms, instead of human traders.

Part of HackerNoon's growing list of open-source research papers, promoting free access to academic material.

Maker

Anatomy of a Simulated Market: Behavioral Modeling of Algorithmic Traders

Inside the Artificial Markets That Predict Real Financial Shifts

Read My Stories

Too Long; Didn't Read

CTA: Download the FREE AI Productivity Shift report

Recreating the Algorithm That Almost Broke Wall Street

About Author

Comments

TOPICS

THIS ARTICLE WAS FEATURED IN

Related Stories

Untitled Story

A High-Frequency Model for Analyzing the 2010 Flash Crash and Mini Crash Events

A High-Frequency Model for Analyzing the 2010 Flash Crash and Mini Crash Events

Inside the Artificial Markets That Predict Real Financial Shifts

Anatomy of a Simulated Market: Behavioral Modeling of Algorithmic Traders

What Are Momentum Traders?

A High-Frequency Model for Analyzing the 2010 Flash Crash and Mini Crash Events

A High-Frequency Model for Analyzing the 2010 Flash Crash and Mini Crash Events

Inside the Artificial Markets That Predict Real Financial Shifts

Anatomy of a Simulated Market: Behavioral Modeling of Algorithmic Traders

What Are Momentum Traders?

Light-Mode

Classic

Newspaper

Dark-Mode

Neon Noir

Minty

HN StartUps