The potential of an ogre encounter is only materially relevant to a meaningful choice being offered to the players if they are predicating the choice, at least in part, around its potentiality.
While this is true, I generally find that
most players feel choices of the type "investigate location X," "take direction Y," or "go to Z first" have such a predicate built-in. That is, if players are presented with a fork in the road and
asked, "Which direction do you want to go?", they assume (a) the two paths go to different places, unless some justification says otherwise (e.g. the paths could converge again later); (b) because you're actually offering the choice, instead of abstracting it out,* it must actually be relevant to their interests; and (c) they will experience
different consequences depending on whether they choose one path or the other.
This is one of the reasons why I work, very hard, to provide supporting justifications for things well in advance, and on the occasions where I slip up, I do my best to empower my players to find out and adapt. E.g., it is a well-established fact in my DW setting that there are Very Nasty Monsters that live out in the desert, especially the deep desert where travel routes don't typically go. Hunters in the wastes (Rangers and Slayers) have a centuries-old tradition of hunting these fell beasts, but there's never really an end to them. The desert conceals many things, and the monsters' origin is one of them. Thus, random encounters are well-established, and the party knows if they're gonna make a journey, they should try to learn about the location so they can be prepared for likely dangers. They
expect "quantum ogres,"** in the sense that they know there's always a chance it could happen and they know to prepare for that possibility consistently.
In other cases, if I "fudge" mechanics or the like, I always make it clear to the party
that something weird is happening, and provide them with the opportunity to learn what it is, and how they can respond to it (whether learning to prevent it in the future, making use of it themselves, weaknesses induced by it, etc.) So, to use the classic "the party got a lucky string of crits and blew away the cool boss fight I made," if I felt it was
that important to keep the fight around long enough for something to happen, I would make up something and the party would be able to observe it. E.g. if it's an evil wizard, "You land crushing, telling blows--the wizard is caught completely off-guard by your attacks. His face contorts into an ugly expression as you land what
should have been a lethal wound, yet somehow, he still stands. He draws from his robe a golf ball-sized globe, which seems to be filled with fire and...possibly blood? He crushes it in his hand and you smell burning blood and the faint trace of brimstone. His wounds are knit shut by sutures of fire and his eyes glow with a baleful light. He's clearly determined to defeat you...no matter what it might cost him."
That makes it so the players know their actions SHOULD have worked, but something got in the way (in this case, the powers of hell). It gives them the opportunity to learn, and later on, prepare.
Note that I used quotes above on "fudging." I do not consider this an example of
fudging proper, because (a) it isn't secret, I'm telling the players something is happening, and (b) the players have the opportunity to learn from it. They might bungle that opportunity. Such is the nature of any dice-based game. But they have
something to work with. It's very, very important to me as both a player and a DM that that opportunity exists.
The existence of an orthogonal rider to the outcome of a player choice (meaningful or not) does not railroading make in and of itself, nor does it make the GM deceitful.
Sure. But, as noted, I find that such orthogonality is relatively rare in the things pro-illusionism GMs care to speak of. The illusory fork in the road, where both paths end up at the same place, is clearly not orthogonal to the choice of which path to take. Fudging the numbers for a rules resolution (be it combat stuff or non-combat), whether for or against the players, clearly isn't orthogonal to the choices they made that triggered the need for resolution. And (for what I consider the worst out of a bunch of bad apples) secret retconning is emphatically not orthogonal to the players' choices, since you're literally changing what the world is.
(The example I usually give for "secret retconning" is the changing killer in a mystery story. That is, you've set up a murder mystery where there are multiple suspects, and you know who is guilty. Say the Crown Prince has been murdered, and his sister the (now-Crown) Princess, his step-father the King-Consort, the Duchess, and the Baron are all implicated, but you know that the Baron is the real killer. You have prepared clues to that effect, and the party has collected some of them. One player then, in a flash of insight, figures out the full story...when you're only a third of the way through. It would be secret retconning to change it so that no, the Baron was always innocent but being
framed by the Princess, even though the party had found evidence that, prior to this change, was 100% legitimate and pointing toward the correct killer. Such actions are justified by (at least some) pro-illusionism GMs as "maintaining the fun" or "making sure the game is satisfying" or whatever. I, of course, don't really accept those justifications.
In the game world, just as in ours, unpredictable things happen that don't relate back to the things our PCs contrive to do. But because we can only ever incompletely model the game world, these things happen as a result of some procedural game mechanic (procedural content generation before or during play, action-resolution mechanics that introduce complications during play, or what-have-you) that someone (usually the GM in D&D) has to willfully invoke, or by GM or player discretion (i.e. simply narrating that the unpredictable thing now happens). In D&D, where the players are not assumed to be able to author the fictional world beyond their PCs (and to their PCs' backgrounds, to some extent), player authorship of random events in the game world is not assumed to be available except perhaps at a superficial level.
While this is all fair, my issue lies in the rather unrestrained use that gets folded into "GM...discretion" here. That is, fudging, secret retconning, and illusionism all depend on the DM having absolute and unlimited justification for changing literally anything, no matter how integral to the play experience, whenever they feel like, for whatever reason they feel like,
and to then prevent the players from ever having even the possibility of knowing that such changes occurred. It is this arbitrary
and secret exercise of power which concerns me so much.
*As one does with a myriad of uninteresting "choices" that occur (e.g. which mundane clothing you wear that day under your fancy stuff, or what specific meal you choose to eat for breakfast, what time you bathe etc.) That is, "choices" which of their very nature definitely don't matter
and don't have consequences worth investigating in over 99% of cases.
**Properly speaking, ogres are part of both civic and non-civic culture in this region, so it wouldn't be quantum
ogres, more likely quantum chimaera or quantum bandits. But that's kinda splitting hairs.