DraQ
Arcane
A well designed game should provide enough reasons to have functioning NPC communities around and have things like random murder disrupt this, even if it doesn't affect player directly.
A society with lots of unsolved murders will naturally function worse economically (hampering typical RPG shopping routine), will have more distrust towards outsiders (such as PC), more guard activity and less tolerance for overall bullshit.
Nevertheless...
Actually yes. Game shouldn't be unlike any other work of fiction in this regard. In a well constructed work of fiction the audience tends to care about characters.
Realistically. Off-screen actions can be abstracted pretty much at will. If you don't have good detailed mechanics to represent something (such as NPC investigation) you can just collect whatever modifiers you can realistically implement and silently roll for it. It won't deprive player of control or detail because the player won't be there to react or observe (and if they try to observe you can auto-bust them for being there using the old "return to the scene of crime" trope).
Hell, even Skyrim has a sort of delayed detection mechanics in the form of hired thugs, even though it isn't tied to it's actual crime system.
Implementing basic workable "medieval CSI" system isn't fucking rocket science.
What for? You're already seem to be baking a lot IRL, judging from your posts.
Game's premise should be constructed in such way that it works as an adventure hook. PLus sufficiently advanced magic and/or technology can be used to carefully tune realism to the level required by premise and gameplay to work.
Because your character doesn't know how and the game doesn't give you tools to actually translate your IRL knowledge (you presumably don't have either) into mechanics - which isn't much of a problem given that your character isn't supposed to know how to make gun in the first place. OTOH pickpocketing evidently does exist in the gameworld so people should know how to deal with it, and it's the obligation of the designer of the world to design it so that this way is actually conductive to the gameplay - or don't implement thief oriented skills at all because it wastes time and resources that could be spent on something useful.
Presumably because even BW animators might have objections to creating necessary content, but there have been games where you could exploit a guard taking a piss so there is no clear reason no to do it.
Because it's hard to implement. There is no good reason not to if you can and it doesn't compete with some other feature that's going to have more/broader impact.
Anyway, none of those reasons matter if you have already decided to have stealing in your game.
Probably because player would keep wandering into their own poisonous clouds.
Anyway, there is no good reason not to as long as there is some way to detect or expect such gas.
For example Skyrim has colorless flammable gas in some areas that can only be seen because of refraction effects and is easy to not notice against some backgrounds or in a hurry. Wandering in with a torch or while slinging fire spells (or having some slung at you) ends with FOOM!.
Why not? If it exists in the world it's accessible in some way.
Since realism in this case simply means that the fluff and the crunch are consistent with one another, realism stops wherever mechanics stops.
If you for example implement your pre-combat special pickpocket idiocy it means that you should have either stopped before making any pickpocketing mechanics and made something like another magic school or made a proper pickpocket mechanics working like pickpocketing.
Mechanics is your model, fluff is your description of what you want to model. If you keep modelling after your description has ended (or in a completely different direction) you are just making unrelated bullshit.
Else why not make game about cubes firing rainbow octahedrons at each other and exploding with candy? After all realism needs to end somewhere.
