When trying to evaluate an issue, there are many, many factors that could be considered. As a matter of logic, the solution space is huge, and the number of things you could look at is huge.
Most factors are irrelevant.
Of the relevant factors, most are not close calls. They aren’t borderline. There’s way more than needed or way too little. (Or for negative factors, there’s way too much or way less than would be a problem.) Most factors either fail by a lot or aren’t even close to failing.
This is unintuitive due to the factors we’re accustomed to paying attention to. We focus on relevant close calls: issues where more precise measuring, or making improvements, actually might matter. We focus our attention on things near the border between success and failure.
And what type of projects are we accustomed to paying attention to? Ones with a reasonable possibility of success. Not ones that seem impossible. So we ignore projects that require something where we fail by a mile. We avoid projects that involve factors or sub-components that are way too hard.
For example, no one is talking about towing the sun closer to make the weather on mars nicer, because that’s so far beyond our abilities. That one could actually be an interesting thought experiment or a science fiction plot, but no one is talking about actually doing it now or soon.
An example no one would even discuss would just sound dumb, e.g. personally eating all the blackberries in Africa and thereby getting elected president of the world. Most logically possible projects are even dumber and more incoherent than that.
So we focus attention on projects that are “maybes”. We do lots of easy projects without even noticing or thinking of them as projects, e.g. walking to the kitchen, because there are zero parts of that project that are near the border between success and failure. We do a small number of more difficult but realistically achievable projects that require attention. And we avoid projects that are way too hard.
Borderline factors are bottlenecks, constraints or limiting factors. Easy factors are non-constraints with plenty of excess capacity. Factors that are way too hard are blockers, dealbreakers or reasons to give up, not bottlenecks.
Optimizing easy factors is pointless because they’re good enough already. Optimizing overly hard factors is pointless because they will fail anyway – they’re too out of reach and can be considered currently, contextually impossible. Only borderline factors – factors within a margin of error of a breakpoint – should be optimized. (A breakpoint is a point on a spectrum where an incremental change in quantity makes a qualitative difference – e.g. changes type, kind or category rather than just slightly changing amount or degree. The most important breakpoints differentiate success or failure at your goal.)