Okay so...you do realize that subtracting or adding hit points "because you feel like it" (whether or not you have justifications made up) is mathematically equivalent to fudging damage, right? They're literally exactly the same operation, just shifted to the opposite side of the equation. Reducing a player's damage dice is
exactly the same thing as adding HP to an enemy after that player has attacked, and vice versa (increasing their damage roll is exactly the same thing as reducing HP). "[Roll] + bonus - penalty = target" is identical to "[roll] + bonus = target + penalty" in terms of the roller's success rate.
And replacing a die you rolled just now, with one you rolled at the start of the session, is
still fudging. It's just fudging to a number you didn't have total control over.
Neither of which, in my opinion, is "fudging." Unless the players fore-know and expect the subtracted encounter to happen, it's not "real" until the "minis hit the map" (or...whatever terms would be appropriate for TotM)--and thus free game for the DM to change. And as long as the on-the-fly rewritten/bonus encounters don't break information the players already have or earned, it's no problem. For example, consider this situation: "Thanks to Rina the Rogue's quick moves while Polly the Paladin was running interference, we've got a copy of the duty roster for the guard shifts--we KNOW nobody's coming through here between midnight and the first bell." That's information they
earned (successful Bluff or Diplomacy roll on Polly's part, successful Thievery roll on Rina's part), which they've then used to make a sound plan. Breaking their plans by adding an extra, unexpected guard shift is supremely unfair--a form of fudging in my book (even if it is meant to "balance out" their fantastic luck at making difficult Bluff and Thievery checks).
If the group gets real, genuine intel about things, the DM no longer has carte blanche to change them--unless there is a meaningful opportunity for the players to learn differently BEFORE the change is sprung on them. In the above example, as the group plots their infiltration, Wendy the Wizard might make a passive Investigation check to realize, "Hey...we actually STOLE this copy of the duty roster...even if they don't know it was us, they'll know it went missing. This info is a trap now!" I'd still call that a somewhat






move on the DM's part, but at least it means the players can make new,
properly informed choices--instead of choices based on wrong
Yeah I don't think any of this is really "fudging" either, unless you throw a nasty encounter at the whole group and try to prevent the death of someone *else* who *wasn't* ready to change characters. Though I do wonder what you'd do with someone like me, who will stick with a single character 'til kingdom come if there's still interesting stories to be told...
This...well, again, as long as it's something the players
could have actually learned during the course of play, I don't mind it at all. But if the threatening-guy
really was the murderer a first, and they got
actually correct evidence pinning the guilt on him, I wouldn't be okay with switching who the real murderer is. Adding a twist--"his shirtcuffs were bloody because he was trying to heal the victim!" or "he's been angry and weird because the 'poison' given to the victim is actually the potion he uses to control his werewolf transformation, which was stolen to frame him for the murder!"--is perfectly fine. Revealing that the evidence is actually fake--IF the players could actually learn this, somehow--is also fine. But a sudden ass-pull "woops, the guy you were correct to chase this entire time is ACTUALLY innocent" would HUGELY piss me off as a player--first because it comes out of the blue, second because we WERE right until the DM changed his or her mind.
And this is precisely the kind of problem that I believe will solve itself, if the players stop to think about what they're doing. "Beating the odds" isn't a thing--happening to end up in the unusual upper (or lower) 5%, however, is. Being *aware* of the risks you're taking, making
calculated risks, investigating and hedging and learning how to exploit the resources available to you is what allows you to
warp the odds to be in your favor. And then there is no need for hidden cheating--if everyone, the DM and the players, is as well-informed as they can be, and exploits their resources to the fullest, then there need not be any cheating at all, as long as the other safety nets (removing or weakening as-yet-unseen encounters, using non-lethal interpretations of total-party knockouts, etc.) are in place.
What's happened, then, is that your players have been trained to think that their odds are a particular thing, that they will have success if they do X, because their past experience says it will. But their past experience is
wrong. It won't lead to success--it only did because someone else (that is, you) ensured that success would result from plans that were either wrong-headed, overly-risky, or both. They have developed a mistaken idea of how to evaluate their situations, which means that even if they
are well-informed, they can't make wise choices; the choice-making apparatus itself is busted. As I've said to others: can you really say that what you're doing is
unequivocally "good" when you
know you must take great pains to hide it from your players?
Obviously, you can and should continue to play in a way that makes you and your players happy. But, at least as I understand this situation, your own meddling is what has
created the very situation that requires you to meddle again, and again, and again, constantly concealing and obfuscating the hand you had in "their" victory. My understanding of the situation has no bearing on what you should or should not do. I only offer it as another perspective on the situation, with the hope that it helps you game the way you like best (which may very well be the way you game now--but may also very well
not be that way).
As I read this, I can't help thinking it sounds an awful like, "So besides that one little steak on Fridays, no meat EVER! THAT'S JUST NOT COOL!" In other words, methinks the poster doth protest too much--and in fact actually fudges a fair amount, as long as nobody can find out. Which nobody can, because your "proof" is indecipherable scrawl, and nobody knows you use a hidden stock of pre-rolled dice!