Trexquant obtains an abundance of data from many different external sources. Many thousands of data variables are refreshed daily and are used by alpha models.
Alphas are trading signals derived from data variables and designed to predict potential excess returns. They are constrained to be market and industry dollar neutral.
Trexquant has built a library of tens of thousands of Alphas and continues to add intelligently to this collection.
Proprietary machine search technology enhances the development process.
Strategies are optimized combinations of alphas which are designed with the intention of generating a stronger signal that overcomes trading costs while satisfying risk constraints.
Strategies can dynamically adjust their parameters based on changing market conditions.
We employ sophisticated machine learning methods to classify alpha signals and develop strategies, such as decision trees, random forest, neural networks and others.
We run multiple strategies to maximize the efficacy of our alpha library and allow flexible capital allocation.