Pedantic
Legend
I prefer a mechanical understanding of "immersion" to describe what you're talking about here. Mechanics become more immersive the less space they create between player/character decision making. Thus, a character trying to succeed and a player trying to play well can she to the same choice of action, which requires among other things a tight relationship between causality and action declaration/resolution. That's my preferred goal for play, with all the theatre kid stuff consigned to determining what goals the PC is trying to succeed at in the first place.In almost all cases where someone says that something is destroying immersion, that exact same thing is destroying the type of challenge based play that the role-playing medium provides.
The reason this is invisible is because of the whole fish doesn’t see water thing. The sense of vividness and reality is based around the challenge game play loop that in actual play would feel like exploration (if the challenge part was a given).
For instance. You’re searching some study in a Cthulhu mystery game and the players and GM just spend hours doing it. The loop is:
Players ask questions about the fiction stuff > GM provides answers. All the while you’re building up both a very vivid scene that adds to the reality and ‘trying to figure something out.’ It’s the two in unison that makes challenge based role-play exciting*. Part of the challenge is asking the correct questions to elicit further information based on the previously established facts.
That, unfortunately, does not seem to precisely mirror the "immersion" described by low rules visible proponents. There's overlap, but the goal there can't be understood as a design goal for the game system, because they reject a ludic mindset as immersive at all. To acknowledge you're playing a game in the first place misses the point, and is at best a means to an end. This is where you occasional get players wistful for completely obscured rules or dipping into FKR.
The earlier point about just how subjective "realism" ends up being in this context is part of the reason this kind of play doesn't yield a mechanically resolvable design goal. Precisely because it's so unclear what will or will not shake any individual player's connection, concentrating power in the hands of a real time designer is the only solution that can allow for the desired play.