This project involved the analysis of detailed high-frequency trade data, including analysis if a key missing data point: time. Investigating the myriad of clocks and synchronisation structures used in the capital markets, seemingly random behaviour became clearer. Supplementing this with proprietary order flow data allowed trade interactions to be classified in terms of informedness, allowing the identification of low-latency predatory traders.
This analysis formed the basis of a HF-TCA methodology that allows any order flow to be classified in terms of it's interaction with more informed traders. Providing the following metrics:
The above methodology was also used to identify informed order flow, and calculate an overall rent that the market pays to such traders - for liquidity that would otherwise find a natural counterparty.
Model: Custom built
Client: Capital Markets CRC, Confidential
Technologies: C++, Python, Pandas
Website: Click here