TO me, the most important reason it is bad is it moves the DM from impartial arbiter of action to a role with vested interest in the outcome of the encounter. Why is that bad? I think that's bad because of all the game roles, designer, DM, or player, the DM has the most covert control over the course of the session. The design is written down and can be reviewed by the DM and others. The player's actions are announced and interpreted by at least one other individual, the DM. No such oversight applies to the DM. The DM operates with as close to perfect knowledge of the situation, ramifications, and consequences to action as possible. If the DM is acting as an impartial arbiter then the events and consequences for player choice evolve 'naturally'. If the DM has a vested interest in a preferred outcome or solution then there is a likelihood that the situations the character find themsleves in with be guided and thus strip some player types of the feeling of success.
I don't think this argument works. I think the point is that the DM is not an impartial arbiter if he's already setting up the preconditions in which he will (allegedly) impartially arbitrate.
Personally, I don't any of the following as being all that different from one another- at most, a difference in degree and not type:
1. Fudging a die roll to avoid a PC death or other unsatisfying event (not all pc deaths are unsatisfying, yadda yadda, if any passer-bys which to launch into a tirade on this point please fork the thread)
2. Fudging a die roll to ensure a PC setback you predict will be fun
3. Secretly granting a bonus or penalty to a die roll that isn't justified by the rules, but which is logical and helps avoid a PC death or other unsatisfying event
4. Secretly granting a bonus or penalty to a die roll that isn't justified by the rules, but which is logical and helps ensure a PC setback you think will be fun
5. Designing an encounter in advance so that certain outcomes are likely, for example, by including lots of places to hide so that a stealthy PC can succeed
6. Designing an encounter in advance so that certain setbacks are likely, such as including bright light to make a stealthy pcs life more difficult
7. Making a decision on the fly so that a PC can succeed, or is more likely to succeed, such as filling a previously unmapped room with hiding places so that a fleeing PC can hide or ambush his pursuers
8. Making a decision on the fly so that a PC can suffer a fun setback, such as a previously unmapped room being filled with a bathing opera star who screams at the PCs arrival with volume that shatters glass, alerting every guard in the area but giving him someone to sweet talk
9. Publicly announcing, in advance of it coming up, that a particular roll will get a bonus or penalty for a plot reason you've invented that will help create an outcome you think will be fun
10. Publicly announcing, at the time of the roll, that there will be a bonus or penalty for a plot reason you've invented that will help create an outcome you think will be fun
11. Publicly announcing, in advance of it coming up, that an entire rule will be altered for a plot reason you've invented that will help create an outcome you think will be fun
12. Publicly announcing, at the time of the relevant event, that an entire rule will be altered for a plot reason you've invented that will help create an outcome you think will be fun
The difference between these, to me, seems to be how seamlessly they fit into the game. Some show more or less of the man behind the curtain, others preserve the illusion of the Great Oz.
I think there's a lot of reasons to view numbers 10 and 12 as the worst, actually. No matter what the DM is exercising editorial control, but RPGs work on the illusion that the DM is being neutral. Numbers 10 and 12 are the ones that makes it most obvious that the DM is altering "the game," and since it happens at the time of the die roll, it makes it seem as if the DM is doing so without forethought.
Number 9 is less bad because it implies the DM considered matters in advance and came up with a situational rule, which is normally the DM's job.
Numbers 5 and 6 are least objectionable, and probably completely unavoidable. There are actually some DMs who claim not to consider PC abilities when designing encounters, but I'm not sure I believe them. In-game outcomes would be awfully strange if PC abilities were in no way considered during the creation of an encounter. But even 5 and 6 can become bad if the players start to get a feeling that the DM is using this ability unfairly.
Numbers 7 and 8 are amongst the least objectionable, though I've had some DMs object to them on the grounds that no one should ever have unmapped regions that the PCs might encounter. I've never quite grasped that objection.
Anyways, this has gone on long enough and I haven't even started to get into my dislike for the attitude that says that a DMs suspension of disbelief is important... let me stick all that and bottom line this:
All of the examples are the DM exerting editorial control. They have different degrees of utility and subtlety. At least for me, what matters most is preserving the illusion and believability of the gameworld, and letting interesting things happen at the table. Each of these are good or bad to the extent that they accomplish these goals.
But editorial control itself is inevitable and inescapable.