What you're trying to articulate is called psuedo-random distribution*. Here's an overview of it: http://wiki.teamliquid.net/dota2/Pse...m_Distribution. The goal of this distribution is simple: eliminate streaks of non-procs that can frustrate players looking for it, and eliminate streaks of procs that can 'break' gameplay and frustrate players on the receiving end up someone's else RNG (supreme dedication to RNGesus).

It works by deciding you want an desired overall N% chance of something to proc. Keep a history of each roll and know how many rolls it has been since a successful proc and do the following:
1) At the very start, or if the proc happened the previous roll, the chance for the proc to is set to lower (much lower) than N%
2) After each roll where the proc doesn't happen, raise the % chance to get a proc.

Very simply, each time you don't get the proc after trying for it it increases the chance you'll get it the next time. Each time you get it, your chance to get it again lowers. It's still not deterministic when you'll successfully proc, but you'll get the procs more consistently spread out. It's great for competitive gaming that have probabilities because you don't want a mega tournament decided in a way that couldn't be foreseen and no one (other than a computer's random) had control over.

I think in FFXIV it would be great if crits become a bit more stable and predictable. This technique's interaction with Bootshine would be debatable as to whether or not it should reset the roll-event history since it is guaranteed. Deep Meditation needs this more than any other skill as it is true RNG stacked behind true RNG making it even more susceptible to long empty streaks. But even if Deep Meditation were in a streak, chakras couldn't come in any more frequently than crit weapon skill events. Effectively, you'd need both at the same time to observe chakra come in like candy on Halloween.

*If you're a software engineer, you call all computer generated random numbers psuedo-random, so this technique is really a more specific implementation of it.