We develop our algorithms using a wide range of data-analysis tools to create our trading ‘logic engines’. These draw on various optimisation techniques, including genetic algorithms, neural networks, AdaBoost, support-vector machines, K nearest neighbour, regression analysis, naïve Bayes classification and grid search.
Our logic engines allow us to create a range of trading strategies that we fine-tune with high-performance computing technologies before deploying them in the market under strict operating conditions. By combining these trading strategies, we are able to create sub-portfolios of diverse algorithms that can generate superior risk-adjusted returns.