No considerations are given for the strength of the advantage

I wish this were stressed more often. It's really easy to think up selective pressures on any trait and really hard to pin down their magnitude. This means that most armchair EP explanations have very low prior probabilities by default, even if they seem intuitively reasonable.