You are planning an experiment for a mobile game feature and need to decide how much traffic to allocate before launch. The team wants to know how power changes with the minimum effect size you care about.
How do you think about statistical power when planning an experiment?
