Knowledge is great: I suspect we can agree there. Sadly, though, we can't guarantee ourselves infinite time in which to learn everything eventually, and in the meantime, there are plenty of situations where having irrelevant knowledge instead of more instrumentally useful knowledge can be decidedly suboptimal. Therefore, there's good reason to work out what facts we'll need to deploy and give special priority to learning those facts. There's nothing intrinsically more interesting or valuable about the knowledge that the capital of the United States is Washington, D.C. than there is about the knowledge that the capital of Bali is Denpasar, but unless you live or spend a lot of time in Indonesia, the latter knowledge will be less likely to come up.
It seems the same is true of procedural knowledge (with the quirk that it's easier to deliberately put yourself in situations where you use whatever procedural knowledge you have than it is to arrange to need to know the capital of Bali.) If your procedural knowledge is useful, and also difficult to obtain or unpopular to practice or both, you might even turn it into a career (or save money that you would have spent hiring people who have).
Rationality is sort of the ur-procedure, but after a certain point - the point where you're no longer buying into supernaturalist superstition, begging for a Darwin Award, or falling for cheap scams - its marginal practical value diminishes. Practicing rationality as an art is fun and there's some chance it'll yield a high return, but evolution (genetic and memetic) didn't do that bad of a job on us: we enter adulthood with an arsenal of heuristics that are mostly good enough. A little patching of the worst leaks, some bailing of bilge that got in early on, and you have a serviceable brain-yacht. (Sound of metaphor straining.)
So when you want to spend time on learning or honing a skill, it makes sense to choose skills with a high return on investment, be it in terms of fun, resources, the goodwill of others, insurance against emergency, or other valuable results. Note that if you learned a skill, used it to learn a non-customized fact, and do not anticipate using the skill again, it's not the skill that was useful; the skill was just a sine qua non for the useful fact, and others don't have to duplicate the research process to benefit. A skill that yielded one (or more) customized facts - i.e., facts about yourself, that you can't go on to share straight up with other people - might be a useful skill in this way, however.
For practical daily purposes, what is your most valuable skill (or what most valuable skill are you trying to attain now)? Post it in the comments, along with what makes your skill valuable, tips for picking it up, and what made you first investigate it.