Of course, that kind of reflects how I feel as a player about GM fudging; it all depends on how I trust the GM's judgment.
Same. I've had DMs that feel
very cheaty -- a player takes a trip build and suddenly every enemy has 4 legs, or a player creates a pyromaniac and suddenly every enemy has fire resistance, etc. Back in my Seattle days, I had a DM that would throw down 20 enemies onto the battle mat and then just pull a bunch right off the mat mid-battle if it was too much for the PCs.
No matter what we did -- smart or stupid, planning or spontaneous -- every fight would be very hard but not hard enough to kill us. At a certain point you realize that it doesn't matter what you do; these fights are destined for a specific outcome. So, no need to invest time or energy into a plan that might make things go better. In fact, after a while you start to realize that you might even be *protected* from death, so you act stupid just to crash up against the railroad and see how forced it will be.
I myself have never done that as a DM. I run a sandbox and let the dice fall where they may,
almost always. In fact, my players would suggest that I am probably too impartial, as I've allowed them to be TPK'd repeatedly. However, they would be TPK'd probably a little more if it wasn't for the little fudging I
have done.
I had to create a golem to guard a treasure. I couldn't predict if the players were going to storm the ruined keep or if they would send the rogue in to scout -- so I needed a creature that would be a match for all of them, but not so tough that it would insta-kill the rogue if he encountered it solo. I also needed the golem to have the see-invisibility property, which a lot of the weaker ones don't have, and a lot of the tougher ones were too tough. I finally found one that was
almost perfect, in the Tome of Horrors. However, I calculated that
if the golem were to hit for max-damage, it would put the rogue at -1, at which point the rogue would be beaten to death.
Anything other than max damage was fine. So I resolved right then & there that since the golem was otherwise perfect for my game, if I managed to roll max damage, I was re-rolling. Any outcome was fine except one -- I did not feel that certain death with not even a chance to react was "fun" for the players. If the golem hit hard and the rogue lived but was dumb enough to try to solo it anyway, then by all means let's kill the rogue. But killing him out of the blue simply because the monster wasn't a perfect match for all the constraints I had? Nah.
Of course, just as fate would have it, the rogue scouted ahead invisibly, the golem saw him, and walloped him for max damage -- 4 or 5 d6s, with a 6 rolled for every one. Damn it. So I secretly re-rolled and the game went on.
I could have picked a different gaming system that wouldn't have required a re-roll, if I were willing to spend weeks or months investigating such a solution. I could have made a custom monster if I were willing to invest time in that, too. I could have simply said beforehand that the monster would do a set 10 points of damage because that's what I wanted to dictate and then proceeded to railroad the hell out of the game because I'd become some power-tripping evil DM. However, the solution I used took 5 seconds and still left my game with a huge realm of possible outcomes. I was happy, and to this day the players still talk about that insane golem fight, so it appears I provided a good time to them, too.
That is the "mental place" that I hope the DMs I play with can find. I want a DM that leaves things alone, and lets the outcome be the outcome... with rare exceptions for miserably lame crap. Tinker too much and I can see it, which sucks. Tinker not at all and the statistics (at least for D&D) will eventually dictate that something ridiculous, stupid, and annoying happens. In such (hopefully extremely infrequent) cases, I'd rather a secret change than simply discarding or overhauling the entire system.