I think effect size is a very useful tool if you know some odds a priori. E.g if you can compare the effect size with some known features/variables. It can then

I think effect size is a very useful tool if you know some odds a priori. E.g if you can compare the effect size with some known features/variables. It can then give you a rough feeling how true it might be. If the features are small as genetic variants then people will have a hard time estimating the supposed effect (unless it's a functional one that alters the encoded protein e.g).