Transformations  in
Business & Economics

A METHOD FOR TESTING HIGH-FREQUENCY STATISTICAL ARBITRAGE TRADING STRATEGIES IN ELECTRONIC EXCHANGES¹³
Mantas Vaitonis, Saulius Masteika

ABSTRACT. New technological advancements with high performance computing in cryptocurrencies and traditional electronic markets as well as the need for fast trading decision making lead the market members to high frequency trading. The analysis of literature has revealed the absence of the tested formalized methods and technological solutions that would allow for testing the algorithmic high-frequency trading (HFT) strategies in financial markets. The need for testing the HFT strategies is provided in the EU MiFID (Markets in Financial Instruments) II Directive, which aims to oblige the HFT trading operators to test all the trading algorithms and HFT strategies for possible errors.
The authors of the paper attempt at creating and implementing the testing method for the automated high frequency statistical arbitrage trading, which would enable the analysis of a large amount of tick-by-tick data received from the electronic market in nanosecond time-stamp precision. The research task was to prepare a testing method for the automated HFT statistical arbitrage trading that would work with a large amount of high frequency data and offer the hardware prototype. The prototype demonstrates how the use of the graphics processing unit memory, code vectorization, parallel calculations and multidimensional arrays allow to analyze trading data on financial exchanges, to achieve the required calculation speed and therefore to make trading decisions faster than data transfer from electronic exchanges.

KEYWORDS: statistical arbitrage, high frequency trading, futures, cryptocurrencies, multidimensional matrixes, pair trading.

JEL classification: C63, C80, G12, G17.

¹³Acknowledgement: Funded by European Regional Development fund. Funded as part of the European Union's response to the COVID-19 pandemic (project No. 01.2.2-LMT-K-718-05-0081).