Because the moment you introduce any randomness you're also introducing the possibility - no matter how small - of something extreme occurring as an outcome of that randomness.
Only if you decide to design it that way. You present this as a guaranteed, absolutely-must-happen kind of thing, but...it isn't. Backgammon and Monopoly contain dice. Those dice never do anything more (or less) dramatic than moving you around on the board--the closest to "something extreme occurring" is being sent to jail for getting three doubles in a row, and even that is hardly an "extreme" occurrence.
Just because the dice are present, doesn't mean extreme results HAVE to happen. You need to explain why having any source of randomness
guarantees hyper-extreme results--and you're gonna have a tough time explaining something that simply, flatly, isn't true.
And the existence of that possibility is why supposedly-foregone situations still need to be played out. Otherwise, you're fudging the outcome by denying the possibility of the extreme.
Well, since I reject the idea that that possibility is guaranteed unless the designer specifically chooses to add it
on top of choosing to include dice (or other randomness), this conclusion does not follow.
Not sure how much you follow baseball, but...
Last season the Chicago White Sox were a bad team. Historically bad. If memory serves, no other team in MLB history has lost as many games in a season as they just did. There's every reason to expect they'll be just as bad this year.
At the same time, the LA Dodgers were a very good team last year; and have since if anything become significantly better.
By your logic here, were the White Sox and Dodgers scheduled to play each other during the coming season a Dodgers sweep would be a foregone conclusion, so why even bother playing those games?
It just don't work that way.
Not even slightly would I have said that. Because you are (pretty blatantly) comparing apples to
kumquats.
Before a battle begins, nothing is a foregone conclusion. Before a sport match begins, nothing is a foregone conclusion.
I'm talking about situations like "the PCs are all at full health, they have defeated 90% of the enemy force, the remaining combatants genuinely
cannot do enough damage to take down any PC even if they crit on every single attack" level stuff. Things where it is literally, completely JUST a procedural thing to wrap up. It saves everyone's time and energy to just skip the two or three turns of attacks and be done with it--perhaps roll to see if anyone might get hit and the piddly-nothing damage they'll take, if nickel-and-dime bean-counting matters that much.
And, further, it is NOT fudging to call a fight early when it's clear the PCs have simply, outrightly won and all that remains is cleanup duty. Fudging is done in secret. That's the whole
point. With fudging, you genuinely need to never reveal that you've secretly altered the game to be what you wanted it to be, rather than what the rules actually said would happen, because if players found out it would ruin the emotional impact of the situation and make everything feel contrived.
Calling a fight early is not, and cannot be, fudging--because you
have to do it openly. There's no other way to end the fight without, y'know,
telling the players that you're ending it there because it isn't worth the effort to grind through it.
Furthermore, separately from the above, doesn't this position you're taking here conflict with another you've taken before? Specifically, you have (IIRC more than once!) mentioned your annoyance with modern D&D fans' failure to consider
retreat as a valid option. By what you've said here, that should never be even a consideration--since there's always a
chance of victory, the party should never surrender and never retreat, no matter what. Yet I know that's
not what you believe; you very much think retreat should always be an option the party is willing to consider. That, too, is an example of ending a combat early--just one fully under the players' control.