While my distaste for GSN terminology is fairly well documented, it is clear that you and I have different priorities regarding “realism.”
I've no use for 'GSN terminology' either, hence the small 'g' on gamist.
Put another way: I see the game as more of an exercise in trying to simulate* the game world and how these things would happen in it, rather than as a game for its own sake (as opposed to most games where the game for its own sake is the only reason you're playing it).
* - where reasonably possible, in full knowledge that no simulation can ever be anywhere near perfect.
I just don’t see it as uncertain at all. Again, they know or they don’t, and sure we may not know every moment of the character’s life, but their Background and Proficiencies give me enough of an idea. Or you could take the Iserith route and have the player call out the backstory element that they think might be relevant to the situation as part of the action declaration.
You could, but this very quickly (i.e. immediately!) runs into issues with players taking unfair advantage and always just happening to be able to justify having the required knowledge. No thanks.
I find this leads to weird results, like the strong character of the group being capable of incredible feats of strength some 60-odd percent of the time and randomly incapable of opening a jammed door the rest of the time. And it still leads to the frustration of “what do you mean that was my best effort? There are 17 other numbers that would have been a better effort!”
Yeah, to me this is a problem with 3e-and-forward's overly-elaborate skill system.
And yes, sometimes a strong person can fail on a relatively easy strength check just by bad luck or not getting the leverage right or whatever, and then have the not-so-strong person nail it in one and make the strong guy look like a fool. It happens - I mean, how many times have you struggled and failed to open a jar and then had someone who you know to be weaker than you come along and pop it on the first try?
It’s almost like the d20 roll sets the actual difficulty, to which the DC is a modifier.. It’s weird.
Not quite. The d20 roll sets the limit of how well your character can possibly do in this situation, against a set difficulty that just sits there.
The difficulty of a particular task doesn't change, but the PC's ability to overcome it isn't set in stone. (and yes this means PCs fail far more often than if take-20 was in effect, but I've no problem with that). Put another way, a PC might blow through some DC 17 task (e.g. opening a stuck door; roll adds to 23) and then two hours later completely fail on a very similar DC 17 task (roll adds to 10) and need to find a plan B.
Narratively, all this tells me is that whatever approach was used at the first door for some reason doesn't work here. Were the order of the rolls reversed it'd be even easier: lessons learned from the first door were successfully applied against the second.
I think take 20 is pretty dumb too, but for different reasons. To me it comes across as an attempt to take solid DMing advice - don’t call for a roll when there aren’t consequences for failure - and attempts to turn it into a player-facing mechanic. And in doing so it produces a result that’s the worst of both worlds.
That does match with general 3e design philosophy, though, in that they intentionally tried to turn more mechanics over to the player side. Not that edition's best selling feature, IMO.