I ran a game last night, a pre-written module (the game is not relevant) with a certain difficulty level (not so difficult in my opinion), and the party is composed of moderately experienced players and a newbie. The party had to face some enemies, and seeing that they had problems with previous encounters I decided to tweak the fight a bit to make it easier: removed some opponents, made the enemy reinforcements arrive late so the PC's had a little more breathing time between turns, made other enemies waste time boasting and taunting, removed abilities from the "BBEG" stat block.
It still ended in a TPK.
Mostly because the PC's rolls were just TERRIBLE, rolling so bad they ended in the impossible range of, like, 0.0009% chance of failure. What can a GM do in such a situation? I know this question is kinda connected to this other question, but suppose there is no logical/plausible way to save the party and the players might feel cheated if saved by stretching the rules, the world, the story or whatever else. Do we just accept the thing and move on? Personally, I always thought that a TPK was an ending too, although not a much satisfying one, but the players might beg to differ.
Reworded for brevity:
The PCs fight enemies, there's a TPK because of incredibly bad rolls and no way to save the group without it looking like cheating, not in the system we're using at least. Do I cheat anyway, talk to the group and laugh it off, explain that it can happen and change campaign?
Since someone believes that in certain games you cannot fail or die, let's say that we're talking about Pathfinder 2e.
Best Answer
Different Systems, Different Solutions
This problem isn't unique to your group, or your game. How to handle failure is addressed in a lot of ways in TTRPGS (or RPGs as a whole). Here are some things you can put into most RPGs. There may be system-specific mechanisms you can employ or have things built in (like Dark Heresy's "Degrees of Success").
Let's take a look at some things you can implement in most systems, if they are not there already.
Fail Forward
This is maybe the best idea: fail forward. In a nutshell, fail forward says: "if you fail a roll, you still succeed (do the thing) but at some cost."
Some examples:
The Prequel
TPKs usually do not feel good, but you can use that to fuel a new adventure. Those dead adventurers can be used as a "prequel" to a second set of adventurers.
Surely these adventurers have families and fiends! After a bit of time, people will wonder what happened to their half-brother/aunt/step-mother/etc. These characters have a built-in plot hook to find out how their sister/uncle/father/3rd cousin/etc died.
Additionally, if this is some sort of regional or larger conflict, so many more people than a rag-tag group of adventurers will be interested or involved in combating... whatever is going on.
This also lets your players have a second chance at the adventure. You, as the GM/DM, give them permission to meta-game a little, too. Placing notable items from dead characters in the new party's ways (or their corpses) can be a great touch. You can even "retcon" a bonus or hidden item in the adventure for the new party to stumble on. (You can take this a step further and let each character leave behind a clue or a bonus item.)
Discretion is the Better Part of Valor
Sometimes your players need to know when to use any other tactic than kick down the door and brutalize everything within. Yes, some games like D&D 5e reward this sort of play, but maybe your system is not like that (like Call of Cthulhu). Having a session 0 can help.