Nice link. Chess and piano performance were the two examples that came to mind for me before clicking the link.

However, increasing the quality that yields the greatest marginal benefit is not necessarily the same as increasing the minimum of the individual actions (assuming you don't define "minimum quality" in terms of marginal benefit). For example, if the lowest quality action only has a small impact on outcome, there is probably something else it would be more beneficial to improve. Of course, some composite skills probably do have performance proportional to min(indvidual_skill), in which case increasing the minimum would always be most beneficial, but most don't.