Absolute Returns with Machine Learning

Three Sigma

A machine learning quantitative model that applies training and prediction algorithms to astronomical sets of data in the global financial market. The quantitative methods intend to generate absolute returns regardless of the investment climate or market sentiment.

How does Three Sigma's algorithm compares?

We take an example of a notable hedge fund’s performance and plot it against Three Sigma’s returns. The bar plot shows a past 20-year comparison of the returns from Pure Alpha Fund by Bridgewater Associates and the returns from Nested Fund using Three Sigma’s algorithm.

“We search through historical data looking for anomalous patterns that we would not expect to occur at random.”

― Jim Simons

“Economists are criticized for not being able to predict the future, but, because the data are incomplete and subject to revision, we cannot even be sure what happened in the recent past. Noisy data make effective policymaking all the more difficult.”

― Ben Bernanke

Monthly portfolio updates to see how your nested egg is performing

The models run simulations from a universe of past data up to decades, and aim to generate alphas with high Sharpe ratio with minimal drawdown.

The quantitative methods intend to generate absolute returns regardless of the investment climate or market sentiment.

Simulated Monthly Returns for the Past 20 Years

Monthly net profit in thousands of dollar after 2% management fee and 20% incentive fee. The returns assume a portfolio starting with $100,000 in year 2000 with the investment held until 2019.

Medallion Fund vs Nested Fund

A past 20-year performance comparison of the returns by Medallion Fund by Renaissance Technologies and the returns by Nested Fund using Three Sigma’s algorithm.

Performance in Year 2020

A monthly simulated cumulative return for year 2020 starting with $100,000 with Three Sigma’s model. The algorithm is adaptable to macroeconomic factors such as being resilient to negative market sentiments caused by the COVID‑19 pandemic.

The quantitative methods intend to generate absolute returns regardless of the investment climate or market sentiment. The models run simulations from a universe of past data up to decades, and aim to generate alphas with high Sharpe ratio with minimal drawdown.

“One can predict the course of a comet more easily than one can predict the course of Citigroup’s stock.

— Jim Simons

Cumulative gross returns for the past 20 years

Cumulative annual gross returns starting with $100,000 in year 2000 and ending with $15MM in year 2019. The above results are simulated with Three Sigma’s machine learning model.

This website uses cookies to ensure you get the best experience.