Basic concept is in the title. I'm looking for opinions. Details below.

So, the game is supposed to be a sidescrolling action-RPG, and it borrows heavily from a lot of other genres. One of the mechanics I'm looking to borrow is the high, low and mid stances normally found in fighting games. What that means is holding up (which otherwise would just have the non-combat utility of moving the camera upwards) changes your stance to a high stance, making attacks high attacks (which hit higher, clearly) and allowing you to block high attacks. Holding down obviously would do the opposite, changing your stance to a low stance, so your attacks hit lower and you can block low attacks. Holding neither of these means you're in mid stance, so attacks hit at mid-level and you can block mid attacks. (Projectile attacks cannot be blocked by default.) This also applies to parrying, which requires precise timing (your guard must start just as the attack reaches you) but stuns melee attackers, completely defeats projectiles and may have other benefits when done with certain implements. If this sounds difficult, you can use a shield to mitigate screw-ups in this system by providing a portion of its block percentage both against projectiles and when you are in the wrong stance, though it doesn't remove your screw-ups completely and you'd still do better if you knew what you were doing.

Your question is unclear. Are you asking if the three stances you propose seem
appropriate for your game design? Or are you asking about the use of the controller
to activate the stances? What opinions are you looking for?

Okay, I guess that isn't specific enough. Let me clarify.

Mechanically, I know this system works. As does everybody who has ever played a fighting game. I also know it can work in an RPG, as does anybody who has played Mount & Blade (or similar games), as they also used this kind of mechanic (though them being 3d gives them more directions). That's not my concern, I know that as long as the controls are tight and responsive and the block percentages are appropriately balanced this system will work very well and all that stuff can be tweaked later for best results. My concern is with the system's reception. I don't know how well received it is going to be to include a complex fighting game mechanic in a 2d RPG. These genres have wildly different expectations and this game isn't apeing any game that would carry an expectation of such a system being present. I have no idea if most players will see it in play and think "Oh, like in fighting games? Awesome, that makes the combat much more involved." or if most will think "Oh, like in fighting games? That sucks, why are they making it so complicated?". I imagine there will be a lot of each, but which will be the majority remains a mystery and so does how I can communicate the system (through all means, not just text) to make it easier to get a grasp on so players will lean more towards the former perspective, or how well I'm catering to the players who really cannot be bothered. (Though I think this might be the wrong game for them.) That's what I want to know. And that's why I asked for opinions, not critiques.

I swear I've seen action rpgs in the past that have emulated fighting game systems, not that I can think of any off the top of my head. Some of the older 2d RPGs with sidescrolling combat come to mind. (Hah, go far enough back, how about Zelda 2)

Though if you're making an action game with a block percentage as a block chance, that's terrible. If you're talking about the percentage of damage blocked, that's fine.

That's the percentage of damage blocked. It's 25% for natural and ranged weapons, 50% for polearms and hatched weapons, and 75% for blades and shields. Given that this is followed by a point reduction from armour, those are all more than they sound like.