Effect Size Estimation

Effect size estimation measures the magnitude of a phenomenon or the strength of a relationship, going beyond simple statistical significance. While a p-value tells you if an effect exists, the effect size tells you how much it matters for the bottom line.

In trading, a strategy might be statistically significant but have such a small effect size that transaction costs and slippage make it unprofitable. Traders use metrics like Cohen's d or R-squared to understand the impact of their variables on portfolio returns.

In crypto, where transaction fees on certain networks can be high, effect size is a critical filter for strategy viability. It helps in prioritizing which signals to trade and which to discard.

By focusing on effect size, traders can avoid the trap of chasing statistically significant but economically useless signals. It provides a clearer picture of the potential return on capital.

Integrating effect size into the research process is a hallmark of a mature quantitative approach.