Explanatory Variables
Easy to model, diverse set of predictive time series per asset class, delivered continuously.
Decision Input - Predictive Indices
Our meta-model, covering all predictive relationships we uncovered.
Consists of unpublished mid-frequency asset-class level predictive relationships, sourced from a diverse set of traditional and alternative datasources.
Our Proprietary Research
- Diverse Traditional and Alternative Datasets. Discovered from millions of time series across a wide range of datasources.
- Rigorous Evaluation. Each time-series is evaluated through a multi-step process, with human experts in the loop.
- Our Methodology. Learn more↗
Properties
General properties of our Predictive Indices.
- Market-neutral. Calibrated to have low beta, to ensure that they represent true skill in forecasting returns.
- Risk managed. The risk contribution of individual explanatory variables, as well as the portfolio are continuously re-weighted to a target volatility.
- Well-diversified. Across different parameters and types of datasets to ensure robustness, equivalent to a multi-strategy portfolio.
- Manually tradeable. Predictive Indices decay in days, not minutes, de-risking the trade execution process.
- Interpretable. We provide high level description of explanatory variables, eg.: Logistics / Futures Market Statistics.
- Not capacity constrained. Focusing on the most liquid asset and avoiding latency sensitive strategies enables us to serve large AUM players.
Want more control over modeling?
Use our Explantory Variables to maintain a high level of control over the modeling phase.
-
Public Research
Our research publication series started with the desire to share lesser-known, low-frequency, interesting, intuitive predictive relationships that we believe deserve more attention.