Thursday, June 9, 2011

Adaptive Strategies for High Frequency Trading (research review)

Anderson, Merolla, and Pribula looked at the eminiS&P500 orderbook for their paper,  Adaptive Strategies for High Frequency Trading . They explained some fundamental concepts around applying orderbook volumes, with respect to their levels, to have an idea of where prices (bids/asks) will likely go in the extreme near term future. This is somewhat intuitive as it is in line with general ideas of exchange level supply and demand.

Forecast methods discussed applying best bid/ask Prices using orderbook volumes: 
  1. Mean Squared Error Prediction
  2. Support Vector Machines
  3. Independent Component Analysis
  4. Simple Moving Average
 Tested Trading Strategy

All market making. Basically, using the forecast value for directional bias, and spam the market with limit orders; e.g. if forecast says the inside Ask price will be higher, the strategy would start working the current best Bid and the expected Ask; and vice-versa. That's the rough idea.

So yeah, interesting research paper for anyone looking to learn about high frequency trading.

0 Reflections: