The fact that you need to bet is exactly the problem, even when you are right. The article does 'blind monitoring of memory usage', which the comment you respon

The fact that you need to bet is exactly the problem, even when you are right. The article does 'blind monitoring of memory usage', which the comment you respond to points out suggest to the problematic, but in fact could also be benign and even beneficial. So consider you are right and memory usage after boot correlates to being less snappy. That might be just a coincidence, with N being so small, and not a cause. Or maybe there is an interesting, unmentioned cause for both. Or maybe the higher memory usage is due to features that most users actually want, for example better file indexing for search or whatever. It is not surprising to me that offering less features result in faster systems. In this case the comparison doesn't really make sense. But it would also not surprise me if a system that uses a lot of memory is faster, if it uses that memory for caching for example. So we don't know what 'free memory after boot' is a signal of and we don't really know if that is even important. Maybe it is to some who have 4gb or less memory. I have 64gb in my laptop, why should I care? (honest question, I can think of reasons). If you are going to measure a signal, it is only going to be useful if you have a sound story about what the signal represents and why we should care about it. Otherwise its just a blind ranking game, kind of an e-sports competition, which will perversely incentivize useless investments (of time) and choices.