Effect of Market Spread Over Reinforcement Learning Based Market Maker

Abbas Haider; H. Wang; Bryan Scotney; Glenn Hawe

doi:https://doi.org/10.1007/978-3-030-37599-7_13

Effect of Market Spread Over Reinforcement Learning Based Market Maker

Abbas Haider, H. Wang, Bryan Scotney, Glenn Hawe

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Market Making (also known as liquidity providing service) is a well-known trading problem studied in multiple disciplines including Finance, Economics and Artificial Intelligence. This paper examines the impact of Market Spread over the market maker’s (or liquidity provider’s) convergence ability through testing the hypothesis that “Knowledge of market spread while learning leads to faster convergence to an optimal and less volatile market making policy”. Reinforcement Learning was used to mimic the behaviour of a liquidity provider with Limit Order Book using historical Trade and Quote data of five equities, as the trading environment. An empirical study of results obtained from experiments (comparing our reward function with benchmark) shows significant improvement in the magnitude of returns obtained by a market maker with knowledge of market spread compared to a market maker without such knowledge, which proves our stated hypothesis.

Original language	English
Title of host publication	Machine Learning, Optimization, and Data Science - 5th International Conference, LOD 2019, Proceedings
Editors	Giuseppe Nicosia, Panos Pardalos, Renato Umeton, Giovanni Giuffrida, Vincenzo Sciacca
Chapter	13
Pages	143-153
Number of pages	11
Volume	11943
ISBN (Electronic)	978-3-030-37599-7
DOIs	https://doi.org/10.1007/978-3-030-37599-7_13
Publication status	Published - 3 Jan 2020

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11943 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Keywords

Market making
Market spread
Reinforcement learning
Reward function

Access to Document

https://doi.org/10.1007/978-3-030-37599-7_13

Link to publication in Scopus

Fingerprint Dive into the research topics of 'Effect of Market Spread Over Reinforcement Learning Based Market Maker'. Together they form a unique fingerprint.

Cite this

Haider, A., Wang, H., Scotney, B., & Hawe, G. (2020). Effect of Market Spread Over Reinforcement Learning Based Market Maker. In G. Nicosia, P. Pardalos, R. Umeton, G. Giuffrida, & V. Sciacca (Eds.), Machine Learning, Optimization, and Data Science - 5th International Conference, LOD 2019, Proceedings (Vol. 11943, pp. 143-153). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11943 LNCS). https://doi.org/10.1007/978-3-030-37599-7_13

Haider, Abbas ; Wang, H. ; Scotney, Bryan ; Hawe, Glenn. / Effect of Market Spread Over Reinforcement Learning Based Market Maker. Machine Learning, Optimization, and Data Science - 5th International Conference, LOD 2019, Proceedings. editor / Giuseppe Nicosia ; Panos Pardalos ; Renato Umeton ; Giovanni Giuffrida ; Vincenzo Sciacca. Vol. 11943 2020. pp. 143-153 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{d31647a3d1b846459f43f049b5ebeec2,

title = "Effect of Market Spread Over Reinforcement Learning Based Market Maker",

abstract = "Market Making (also known as liquidity providing service) is a well-known trading problem studied in multiple disciplines including Finance, Economics and Artificial Intelligence. This paper examines the impact of Market Spread over the market maker{\textquoteright}s (or liquidity provider{\textquoteright}s) convergence ability through testing the hypothesis that “Knowledge of market spread while learning leads to faster convergence to an optimal and less volatile market making policy”. Reinforcement Learning was used to mimic the behaviour of a liquidity provider with Limit Order Book using historical Trade and Quote data of five equities, as the trading environment. An empirical study of results obtained from experiments (comparing our reward function with benchmark) shows significant improvement in the magnitude of returns obtained by a market maker with knowledge of market spread compared to a market maker without such knowledge, which proves our stated hypothesis.",

keywords = "Market making, Market spread, Reinforcement learning, Reward function",

author = "Abbas Haider and H. Wang and Bryan Scotney and Glenn Hawe",

year = "2020",

month = jan,

day = "3",

doi = "https://doi.org/10.1007/978-3-030-37599-7_13",

language = "English",

isbn = "978-3-030-37598-0",

volume = "11943",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

pages = "143--153",

editor = "Giuseppe Nicosia and Panos Pardalos and Renato Umeton and Giovanni Giuffrida and Vincenzo Sciacca",

booktitle = "Machine Learning, Optimization, and Data Science - 5th International Conference, LOD 2019, Proceedings",

}

Haider, A , Wang, H , Scotney, B & Hawe, G 2020, Effect of Market Spread Over Reinforcement Learning Based Market Maker. in G Nicosia, P Pardalos, R Umeton, G Giuffrida & V Sciacca (eds), Machine Learning, Optimization, and Data Science - 5th International Conference, LOD 2019, Proceedings. vol. 11943, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11943 LNCS, pp. 143-153. https://doi.org/10.1007/978-3-030-37599-7_13

Effect of Market Spread Over Reinforcement Learning Based Market Maker. / Haider, Abbas ; Wang, H.; Scotney, Bryan ; Hawe, Glenn.

Machine Learning, Optimization, and Data Science - 5th International Conference, LOD 2019, Proceedings. ed. / Giuseppe Nicosia; Panos Pardalos; Renato Umeton; Giovanni Giuffrida; Vincenzo Sciacca. Vol. 11943 2020. p. 143-153 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11943 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Effect of Market Spread Over Reinforcement Learning Based Market Maker

AU - Haider, Abbas

AU - Wang, H.

AU - Scotney, Bryan

AU - Hawe, Glenn

PY - 2020/1/3

Y1 - 2020/1/3

N2 - Market Making (also known as liquidity providing service) is a well-known trading problem studied in multiple disciplines including Finance, Economics and Artificial Intelligence. This paper examines the impact of Market Spread over the market maker’s (or liquidity provider’s) convergence ability through testing the hypothesis that “Knowledge of market spread while learning leads to faster convergence to an optimal and less volatile market making policy”. Reinforcement Learning was used to mimic the behaviour of a liquidity provider with Limit Order Book using historical Trade and Quote data of five equities, as the trading environment. An empirical study of results obtained from experiments (comparing our reward function with benchmark) shows significant improvement in the magnitude of returns obtained by a market maker with knowledge of market spread compared to a market maker without such knowledge, which proves our stated hypothesis.

AB - Market Making (also known as liquidity providing service) is a well-known trading problem studied in multiple disciplines including Finance, Economics and Artificial Intelligence. This paper examines the impact of Market Spread over the market maker’s (or liquidity provider’s) convergence ability through testing the hypothesis that “Knowledge of market spread while learning leads to faster convergence to an optimal and less volatile market making policy”. Reinforcement Learning was used to mimic the behaviour of a liquidity provider with Limit Order Book using historical Trade and Quote data of five equities, as the trading environment. An empirical study of results obtained from experiments (comparing our reward function with benchmark) shows significant improvement in the magnitude of returns obtained by a market maker with knowledge of market spread compared to a market maker without such knowledge, which proves our stated hypothesis.

KW - Market making

KW - Market spread

KW - Reinforcement learning

KW - Reward function

UR - http://www.scopus.com/inward/record.url?scp=85078427273&partnerID=8YFLogxK

U2 - https://doi.org/10.1007/978-3-030-37599-7_13

DO - https://doi.org/10.1007/978-3-030-37599-7_13

M3 - Conference contribution

SN - 978-3-030-37598-0

VL - 11943

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 143

EP - 153

BT - Machine Learning, Optimization, and Data Science - 5th International Conference, LOD 2019, Proceedings

A2 - Nicosia, Giuseppe

A2 - Pardalos, Panos

A2 - Umeton, Renato

A2 - Giuffrida, Giovanni

A2 - Sciacca, Vincenzo

ER -

Haider A , Wang H , Scotney B , Hawe G. Effect of Market Spread Over Reinforcement Learning Based Market Maker. In Nicosia G, Pardalos P, Umeton R, Giuffrida G, Sciacca V, editors, Machine Learning, Optimization, and Data Science - 5th International Conference, LOD 2019, Proceedings. Vol. 11943. 2020. p. 143-153. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-37599-7_13