Our mathematicians, statisticians, engineers and computer scientists perform fundamental research and advanced quantitative analysis, and utilize cutting –edge technology to identify and create new strategies in the investment arena.
Our organization and our people are diligent students of several of the best schools of statistical learning (in the tradition of Hastie, Tibshirani and Friedman at Stanford) as well as machine learning. We deploy state-of-the-art software, including many packages from the R statistical language such as caret and mlr along with their underlying boosting, bagging and tree-based classification and regression methods.
It is this core belief in the academic foundation and principles of our work that inform our approach to investing. In “A Practitioner’s Defense of Return Predictability,” a white paper authored by Blair Hull and Xiao Qiao, the issues and opportunities of market timing and return predictability are explored.
The paper discusses a six-month horizon model that leverages correlation screening to combine a constantly changing collection of twenty academically referenced variables in order to demonstrate forecasting efficacy. Using a walk forward simulation in which positions in SPY are taken proportional to the model forecast equity risk premium, this approach indicates simulated strategy yields of more than twice the annual returns of a buy-and-hold strategy and a corresponding Sharpe ratio four times that of buy-and-hold.
The results of the paper can be replicated by accessing the files:
2) the White Paper Indicators file.