False Discovery Rate

Good article on False Discovery Rate – the probability that a significant p-value indicates a true effect and one of the reasons that quant tests and systems generally underperform.

For example, Market Stress and Market Stress Automations generally test out at around 85% wins. In real life they run about 62%.

Here’s the link: http://www.nicebread.de/whats-the-probability-that-a-significant-p-value-indicates-a-true-effect/

False Discovery Rate

Leave a Reply