I might roll a random encounter that would be a TPK to the party in their current state, so I ignore or reroll it. I could decide the magic item I just rolled randomly is too disruptive and roll another. I might decide that a crit that would kill a player in the first round is a normal hit which keeps the PC alive but very wounded (and changes his tactics and expends resources to live) or I might decide that every player going first against a powerful bad guy isn't going to be a challenge and up his initiative so that he gets some hits in early. In all those situations, the dice provided me with an option that would lead to a less fun situation for the players (myself included).
There are two different occasions for fudging that we need to disentangle. The first, which you've mentioned, is papering over cracks in the rules. The better solution is for the group to change the rules to something that suits them. I'd also suggest that discretionary rules like random encounter charts are not in the same league as action resolution.
The second occasion is when the DM wants to enforce their particular vision of how things should play out. This is plainly bad.
<snip>
Who is the DM to decide whether something 'hurts the experience' or not? Don't the other participants also have an interest in this?
I agree with
@gorice that it helps to be clear about the purpose of various dice rolls.
Some dice rolls are just shortcuts or guidelines for content generation - eg some rolls for treasure. Assuming that the treasure tables are properly designed, they should be a reasonable guide to balancing treasure placement. But departing from them, while still having regard to their implicit precepts, should do no harm.
Some dice rolls are content generation but also a consequence mechanism - in classic D&D, this is true of wandering monsters, so fudging these is trickier. Gygax has a discussion in the opening pages of his DMG of when it might be OK to fudge a wandering monster roll; what the discussion reveals is that there is a mismatch between the game mechanics and the game as played, resulting from three things: (i) wandering monsters are primarily on a time clock; (ii) mostly, the use of time in dungeon exploration is the expenditure of a resource; (iii) sometimes, and despite (ii), the use of time in dungeon exploration is just a by-product of "simulation" and has no actual game play significance. (i) and (ii) are tightly related - be careless with your resource (time), and you will suffer an adverse consequence (wandering monsters). But when (iii) arises, then its connection to (i) causes a problem, because a party can suffer a consequence (ie wandering monsters) even though they haven't yet made any significant play decisions (in Gygax's example, they are just moving through the explored parts of the dungeon to the target destination for their expedition).
The real solution to Gygax's problem is to change (i), so that it doesn't apply in (iii)-type scenarios. But given that that is not a trivial design exercise, the fudging "solution" is understandable, provided the GM understands the principles that should govern its application. (And Gygax tries hard to spell them out.)
Some dice rolls are action resolution. Gygax is adamant, in the same opening pages of his DMG, that these should
not be messed with; and even in his later discussion of fudging, he gives as an example in the context of action resolution
the effects of being reduced to zero hp, and gives the GM licence to treat that as something other than death (but still hors de combat), again subject to appropriate principles.
If the impetus to mercy that Gygax points to is coming up at all often, then that suggests it's time to change the rules! If the action resolution rules regularly fail to provide the desired action, that's a problem.
If I as a DM have to fudge the dice to make the game work, something broke. I either overdid things on the encounter scaling side or the game was way too swingy, but what I don’t want to see is something in the rules that undercuts that and says “If you didn’t like the outcome, just fudge the result so it’s what your table would prefer or what benefits the story.”
Right.
From WotC's perspective, though, they have a
strong interest in maintaining, at least in general terms, a consistent set of game mechanics, even where those cease to be suitable for the sort of game their customers are playing. This was true back in the AD&D 2nd ed days and hasn't changed.
The issue of fudging vs fate is as old as D&D itself, and outside of Gygax's notion that all rolls be fair regardless of outcome, most of the advice I've ever seen in RPGs is that in a choice between fun and fair, fun should win.
<snip>
2nd edition (where I cut my teeth) was a terrible mismatch of expectations and rules. 2e wanted narrative focused adventures with heavy role playing and storytelling and then used a version of D&D that was full of instant death, randomized chargen and shackles on PC power. No matter how many essays the Dragon Mag and the DMG made, the rules were at odds with the mission statement.
I agree fully with the second quoted paragraph. The first arises because of the second: and so, in a game that is designed without legacy baggage, and that takeg full advantage of the growth in technical understanding of RPG design over the past 50 years, there should not be the sort of tension between
outcomes the mechanics deliver and
the mission statement that tends to arise for D&D.
