The performance of selected high-frequency trading proxies: An application on Turkish index futures market

This paper intends to provide evidence for how well high-frequency trading (HFT) proxies capture low-latency activity in rarely explored futures markets. We first run suggested identification algorithms using tick-by-tick order and trade message data to derive models’ HFT estimates. Contrasting these with Exchange-provided classification tags (considered as real HFT messages), we interpret the soundness and consistency of these proxies with regard to various reference metrics through an empirical mindset. Our results suggest that certain proxies track low-latency behavior better than others affirming their given credits of reliable HFT identifiers in practice.