CloudQuant proves value in PA Signals alt date set

CloudQuant says it has proven the value in the Precision Alpha Machine Learning Signals (PA Signals) alternative data set. Its detailed data science study shows a long-short portfolio outperforms the equal-weight S&P 500 ETF by an average of 37.9 per cent per year after transaction costs. 

CloudQuant found that over 91.5 per cent of the total return is pure alpha. The results of the study are significant to the 99th per cent level.

Cutting-edge machine learning is transforming quantitative analysis for portfolio managers and traders. PA Identifies structural breaks and exposes investment signals that market participants are currently unable to see. The PA Signal offers a favourable risk-adjusted return that can be used to create large-scale investment algorithms.

“Backtesting on CloudQuant’s Mariner™ showed that a long top 5 per cent-short bottom 5 per cent quantile intraday strategy achieved overall Sharpe Ratio1 of 5.36 and a very low CAPM beta,” says Morgan Slade, Chief Executive Officer of CloudQuant.

The growing quality and quantity of Alternative Data Sets have created a dilemma for many investment managers. Profitable information is contained in new data but most investors lack the resources to onboard and then research the data. CloudQuant’s quantamental researchers have studied the PA Signals and provide a detailed white paper, and backtesting algorithm with source code (free upon qualified request) that allows any portfolio manager to replicate the research and immediately begin to reproduce the results.

“With CloudQuant investment professionals can jumpstart their research without incurring the cost of dataset ingress and curation. They are able to see the value in the data,” says Mark Temple-Raston, PhD and Chief Data Scientist of Precision Alpha.