PORTFOLIO MANAGEMENT
PORTFOLIO OPTIMIZATION
- We build well-diversified portfolios of trading systems. Our portfolios’ dynamic composition results from a portfolio optimization algorithm.
- We have read dozens of papers on the subject, and on top of them, we have built our private algorithms.
- These start where the portfolio optimization techniques of Meucci (2006) and De Prado (2018) end.
- We keep our algorithms private in the best interest of our clients.
NESTED WALKFORWARD
To account for training and selection bias, our chosen portfolio is based on a Bayesian model selection approach on top of a nested walk-forward algorithm.
Nested Walk Forward, as the word suggests, is based on a two stages analysis:
- In the first one, the different optimization algorithms are trained in-sample, and the out-of-sample results are collected.
- In the second one, the out-of-sample results of the first stage are combined by the Bayesian model selection algorithm to find the best weights for the second out-of-sample period.