Hands-On Machine Learning for Algorithmic Trading

Exclusivity

The likelihood that an alternative dataset contains a signal that is sufficiently predictive to drive a strategy on a stand-alone basis with a high Sharpe ratio for a meaningful period is inversely related to its availability and ease of processing. In other words, the more exclusive, and the harder to process the data, the better the chances that a dataset with alpha content can drive a strategy without suffering rapid signal decay.

Public fundamental data that provides standard financial ratios contains little alpha and is not attractive for a standalone strategy, but may help diversify a portfolio of risk factors. Large, complex datasets will take more time to be absorbed by the market, and new datasets continue to emerge on ...