High frequency automated market making algorithms with adverse selection risk control via reinforcement learning

High frequency automated market making algorithms with adverse selection risk control via reinforcement learning
复制标题

通过强化学习进行逆向选择风险控制的高频自动化做市算法

DOI:
10.1145/3490354.3494398
复制
发表时间:
2021
期刊:
2nd ACM International Conference on AI in Finance (ICAIF’21
影响因子:
--
通讯作者:
Linetsky, Vadim
Linetsky, Vadim
中科院分区:
--
文献类型:
--
作者:
Zhao, Muchen;Linetsky, Vadim

文献摘要

参考文献

被引文献

相似文献

做市商通过在市场两边(买入价和卖出价)下限价单来提供流动性,目的是赚取买卖价差。他们的长期表现很大程度上取决于他们减轻逆向选择风险的能力,当他们的限价单被拥有相关信息的知情交易员选中时,这些信息将市场推向一个新的水平,导致做市商损失。本文提出了一个高频特征——图书耗尽率(BER),并从理论和经验上证明了BER可以从均衡的角度作为逆向选择风险的直接度量。我们使用芝加哥商品交易所(CME)标准普尔500指数和10年期国债期货的三年限价订单数据,通过强化学习训练了一个做市算法,并证明利用误码率可以使算法避免由于逆向选择而造成的巨大损失,并实现稳定的性能。
Market makers provide liquidity by placing limit orders on both sides of the market (bids and offers) while aiming to earn the bid-offer (bid-ask) spread. Their long-term performance is significantly determined by their ability to mitigate the risk of adverse selection when their limit orders are picked off by informed traders possessing relevant information that moves the market to a new level resulting in losses to market makers. This paper proposes a high-frequency featureBook Exhaustion Rate(BER) and shows theoretically and empirically that the BER can serve as a direct measurement of the adverse selection risk from an equilibrium point of view. We train a market making algorithm via Reinforcement Learning using three years of limit order book data on Chicago Mercantile Exchange (CME) S&P 500 and 10-year Treasury note futures and demonstrate that with utilizing the BER allows the algorithm to avoid large losses due to adverse selection and achieve stable performance.
DOI: 10.1080/14697688.2012.708779
发表时间: 2011-06
影响因子: 1.3
作者:
Fabien Guilbaud;H. Pham
通讯作者: Fabien Guilbaud;H. Pham
限价订单市场中小投资者的最优投资组合:影子价格法
DOI: 10.1007/s11579-010-0027-9
发表时间: 2010
影响因子: 1.6
作者:
C. Kühn;Maximilian Stroh
通讯作者: Maximilian Stroh
通过在线学习进行适应性做市
DOI: --
发表时间: 2013
期刊: Neural Information Processing Systems
影响因子: --
作者:
Jacob D. Abernethy;Satyen Kale
通讯作者: Satyen Kale
DOI: 10.48550/arxiv.1804.04216
发表时间: 2018
期刊: arXiv e-prints
影响因子: --
作者:
Spooner Thomas
通讯作者: Spooner Thomas