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.)

"Dialectical Bootstrapping" is a simple procedure that may improve your estimates. This is how it works:

1. Estimate the number in whatever manner you usually would estimate. Write that down.
2. Assume your first estimate is off the mark.
3. Think about a few reasons why that could be. Which assumptions and considerations could have been wrong?
4. What do these new considerations imply? Was the first estimate rather too high or too low?
5. Based on this new perspective, make a second, alternative estimate.

Herzog and Hertwig find that average of the two estimates (in a historical-date estimating task) is more accurate than the first estimate, (Edit: or the average of two estimates without the "assume you're wrong" manipulation). To put the finding in a OB/LW-centric manner, this procedure (sometimes, partially) avoids Cached Thoughts.