Verisimilitude is an end, not a means. This distinguishes it from things like "elegant design" or "nifty mechanics". There is often a hierarchy of means where one things is clearly better than another, even if the comparison to a third thing is more questionble. (For example, "d20+mod versus target number" is not inarguably the "best" means for task resolution. It is almost assuredly better for D&D than "5d4+27 different mod categories versus an opposed check with variable basis." And it is inarguably better than "slap the DM until he agrees that you succeed" for any normal game.
)
But "verisimilitude" by itself is not just any end, but also a conceptual end, not a discrete one. This distinguishes it from from more concrete ends like "a system with combats that take place quickly" or "a system with combats that provide lots of meaningful decision points."
Saying you want "verisimilitude" is thus almost meaningless. It is like saying, "I want some nifty mechanics that produce the goals I want for combat." OK, what are your goals for combat? You want fast combats, meaningful decision points, or some hybrid, and how much?
For verisimilitude to have meaning, you have to say what reality you want approximated, and how much, and at what costs. Then
those things can be usefully discussed, one after the other, classified, weighed, etc. But if your argument for any one of them is that it is a good choice
because it will aid "verisimilitude" then you are trying to shift the ground from an end to a means. Instead, your argument should be the popularity of this particular item, and how many people want that "reality" established, and how little it will impede on people who want something else. If you can't, then it is your mere personal preference for a design goal, and carries no weight.
That's like me saying that everyone here should give me $100. Then when someone ask why they should give me $100, I said, "because then I'll be richer." Well, yeah! But I need something else to work with before I part with the cash.