Some statistical and computational aspects on managed futures funds

Tobias Rydén
Brummer & Partners
Stockholm


Abstract:

A managed futures (MF) fund (also called commodity trading advisor, CTA), is a fund that trades futures contracts. In this talk I will first give some basic facts about what futures contracts are, what opportunities that diversification across assets and risk adjustment over time offers, and how performance can be measured. Making predictions of future price movements is extremely difficult, which I will illustrate, but I will also illustrate how little more than a 50% chance of a correct prediction that it needed to achieve good performance.


Traditionally, systematic trading of futures was often done using various so-called technical indicators. I will give examples of such, and then proceed to discussing how they can be used, viewed, and generalized from the perspective of high-dimensional statistical and machine learning thinking. In addition I will also talk about principal components, which is a statistical method that has a stronger motivation from economics, in the context of MFs.