It seems pretty straightforward, to me. It is the basic conflict between a GM's judgement, and resolution through an objective process.
Say my character wants to lift an object. In the GM's judgement, there is uncertainty in the result of this attempt. So, the GM says, "I will use the resolution mechanic the written rules have for lifting heavy objects." You then check these rules, and find that either the character cannot fail (if the object is light) or cannot succeed (if the object is heavy). So, it turns out there *wasn't* uncertainty.
No. What turns out is that this PC is capable enough at that level of uncertainty.
Uncertainty does not append to the outcome of the check -- this presupposes a check -- but to the action -- is this action uncertain to be successful. What you're doing here is assuming a check and looking for uncertainty in the outcome of the check. This is not what I'm saying. I'm saying that you look at the action with regard to the fictional state and determine if it is uncertain, and if so, select mechanics to resolve the uncertainty. A check is not a given.
Well, maybe we say there's no big deal. The GM gave you an even shot, at least.
The problematic case is when the GM determines there is no uncertainty, but a reasonable person looking at the rules thinks there ought to be. This goes south generally when the result goes against the player, and the typical mechanics you'd apply said there was a good chance for things to be different, but the GM decides to not use the rules.
This looks like an arbitrary (or worse, a purposeful, non-arbitrary) fiat on the part of the GM.
Another ecampke of "but what if a jerk does it," as if this addresses the topic rather than just introduces a jerk. Here's a thing, someone can always imagine how you say you play with jerks instead and make a case for bad play. It's not your method, though, it's the jerks.
I use this method and don't have this problem. This is because I'm not a jerk, and I strive to engage my players and not gotcha them. Proposing play that is a gotcha, ir results from lack of effort to be clear on the situation, doesn't actualky address the method.
You say "at least I got a roll" as if this protects against jerks. It diesn't, it's weak, and I'd rather address the jerk involved as the problem.
[Quite]
The core play loop you described is idealized, and does not include subtle factors relating to how both the GM and players are imperfect humans, with social dynamics and expectations, and major expectation failures are a failure mode for play.[/QUOTE]
This is absolute hogwash. Fir one, tge play loop is as much rules as the combat section. It's right there in the front of the book as how to play the game. Two, it doesn't rely on mythical perfect people, nor is it more prone to degenerate play than your preferred more 3.x style of hard mechanics. If this was true, it would be impossible for those of us that report excellent results to be anything other than liars for no perceivable reason.
I used to play as you argue -- I got in some ugly arguments with [MENTION=97077]iserith[/MENTION], refusing to believe he could possibly be honest in how he presents this playstyle. I can't say what started me listening, but everything I believed before - that you have to have strong codification of mechanics and that you go to mechanics first - is wrong. Not that it's wrong to play that way, or less fun, because that is 100% untrue, but that playing another way is just as good and still within the rules. And, I like this way better - I'm more attentive, more productive, my games move faster and have more player engagement. YMMV, and probably does; what works for me works for me, it's by no means universal. But, by golly, does it chap my backside when someone spouts ignorance like this. Which is probably karma for me doing it years ago.