Optimal high frequency strategy in an omniscient order book.

Summary The aim of this study is to quantify the low latency advantage of High Frequency Trading (HFT) and to compute, empirically, an optimal holding period of a HF trader. Critics claim that low latency leads to information asymmetry victimizing retail investors. However, objective studies measuring the gain due to this asymmetry are rare. In order to perform the study, new methods are introduced in this paper, in particular, the optimal strategy problem is formulated and ideas are given to compute it in a reasonable amount of time. A new measure, the weighted mean holding period, is introduced and an algorithm to compute it is suggested. Using the previous concepts, a large empirical study based on optimal omniscient strategy is presented and evidence of the low latency advantage limitation is provided. In particular, it is shown that the bid ask spread and the transaction costs lead to a trading frequency much lower than the information renewal frequency.