Well yes, at times there is an adversarial angle to the game; the DM is in control of the players' opponents, and should try to run them believably and effectively. But what I mean is deeper than that aspect of the game.
This doesn't directly respond to my key point, though: if, as GM,
I am never bound, then the only limit on how hard I push is
me deciding not to push any harder. That is the situation that I want to avoid. And what lets me avoid it is
being bound - by rules, by mechanics, by principles for introducing opposition, etc.
I don't think I have to be bound at every point. For instance, in 4e I have extreme latitude on introducing adversaries of the PCs into play (contrast, say, Marvel Heroic RP, where this may be limited by the Doom Pool). But there is also a pretty reliable encounter difficulty metric in 4e, which means I at least know how hard I have chosen to push. And, once I am actually resolving action declarations in relation to those adversaries (both my declarations for them, and the players' declarations for their PCs in opposition to them), I am bound.
If, at every point, I had to decide whether or not it was "good for the game" to actually apply the outcomes of some particular episode of action resolution to the fiction, I feel that I wouldn't really be pushing the players at all. In fact, I feel that would be very close to simply free narrating a story in response to suggestions from them.
Why is player latitude to be valued so highly, but GM latitude is so frightening?
Because they have different roles in the game?
In my case, very roughly, the job of the GM is to (i) generate the bulk of the backstory, and handle all of the backstory so that it hangs together and makes sense; (ii) frame the players, via their PCs, into situations that will challenge them in respect of the things they have signalled (formally and informally) that they care about; (iii) adjudicate action resolution, and when narrating failures keep a very close eye on those same signalled matters (= "fail forward").
The players' job is to (i) build PCs who are ready to be challenged; (ii) sense the aforementioned signals; (iii) play their PCs - in particular, that means declaring actions in the face of challenge.
Asymmetric responsibilities means asymmetric latitude.
Others who see the allocation of responsibilities differently will of course take a different view about the degree of latitude that each should enjoy.
Every edition has had a rule or caveat of some sort that ultimately, the DM's word is final.
That doesn't tell us very much about the GM's latitude, though. In Australian law, the judgments of the High Court of Australia are final, in the sense both that it is the apex appeal court for the whole country and that there is no appeal from its decisions. But its latitude is limited - for instance, in most cases it is obliged to apply pre-existing law in the resolution of the controversies before it.
In the D&D context, for instance, the GM has the last word on (say) whether or not his/her dungeon notes included a notation that a particular monster will always pursue fleeing characters. But that doesn't mean that the GM is just at liberty to make it up! Or to lie about it!
I use this example deliberately, because it appears in Gygax's explanation of the rules for evasion and pursuit underground (DMG, p 67):
When player characters with attendant hirelings and/or henchmen, if any, elect to retreat or flee from an encounter with a monster or monsters, a possible pursuit situation arises. Whether or not pursuit will actually take place is dependent upon the following:
1. If the matrix or key states that the monster(s) in question will pursue, or if the
MONSTER MANUAL so states, then pursuit will certainly occur.
2. If the monster or monsters encountered are semi-intelligent or under, hungry, angry, aggressive, and/or trained to do so, then pursuit will be 80% likely to occur (d10, 1 through 8).
3. If the monster or monsters encountered are of low intelligence but otherwise suit the qualifications of 2., above, then pursuit will occur with the following probabilities:
A. If the party outnumbers the potential pursuers, then pursuit is 20% likely.
B. If the party is about as numerous as the potential pursuers, then pursuit is 40% likely.
C. If the party is outnumbered by the potential pursuers, then pursuit is 80% likely.
D. If condition C. exists, and furthermore, the potential pursuers conceive of themselves as greatly superior to the party, then pursuit is 100% certain.
We might think this is a good rule, or a bad rule. But whichever it is (and opinions might reasonably differ), I think one thing is clear: it
only makes sense if we assume that the GM is bound by what is written in (or not written in) his/her dungeon key or in the MM. Thus, for instance, if the GM did not write down that certain monsters always pursue; and if the MM doesn't state that they do; then the GM has to make the d10 roll. And, again, the rule
only makes sense if we assume that the GM is bound by the outcome of that roll in deciding whether or not pursuit ensues.
The reason why the GM is bound is fairly clear, I think - if the GM were not so bound, then scrying and scouting and other information collection by the PCs (which it itself the demonstration of skillful play by the players) would be pointless, because they could not rely on that information.
This is different from my preferred approach to RPGing, but I find it closer to my preferred approach than the "GM as storyteller" approach.
In my opinion, a good DM knows when to allow player choices to shape and mold the events, but knows that at times it is equally important to take the reins a bit. Whether it is to steer the story back on track a bit (example: the characters are doing everything except go to Castle Ravenloft) or to avoid a TPK (by having a critical hit that would put the entire party at risk be only a hit, or something similar).
This is very different from my preferred approach, in which there is no "track" to get the story back onto. The players run up their flags; the GM frames the PCs into situations in response to that; the players declare actions for their PCs; we find out what happens, via action resolution; this may well change the flags that have been run up; and the cycle repeats.
That might be a case of the lack of definition of player agency. Players are still at liberty for their characters to attempt anything that they want to. The difference made by doing something like setting the encounter in a different place is to adjust the chances of what they want to do. Its just one of the types of decision when putting together an adventure. No player agency lost any more than deciding that the characters will encounter zombies in the twisting catacombs for a fight where the more defensive characters can shine this time around for example.
To me, that example (zombies in the catacombs) is not uncontroversially consistent with player agency. Why are the PCs in the catacombs? Why has the GM decided that there will be zombies there? How does this related to past actions declared by, or decisions taken by, the players?
Here's an example of what I have in mind, within my preferred style. (This is an
actual play example). One of the players makes a check (Circles in Burning Wheel, which somewhat resembles Streetwise in D&D and Classic Traveller, but is applicable to whatever social stratum the PC comes from, not just underworld types), to try and make friendly contact with a prospective employer from the sorcerous cabal to which he belongs. The check fails, so in fact he is called to a meeting with Jabal the Red, one of the leaders of the cabal, to be dressed down. The details and logic of the dressing down don't matter, other than to say that it involves a mysterious feather the PC had previously bought at a market stall - what does matter is that (i) Jabal lives in a tower, and (ii) the PCs notice a furtive stranger on a staircase of the tower, who fits the description earlier given to them of the person who sold the feather to the market stall-holder.
The reason for (ii) is that this speaks to the players' engagement with the mysterious feather. The reason for (i) is that, when the players later on sneak into the tower to see what they can learn about Jabal and his furtive friend, the PC with the instinct "When I fall I cast Falconskin (ie shapechange into a falcon)" has a reasonable prospect of having a fall. Thus, both these decisions as a GM, about the particular backstory that I introduce into the game, reinforce the players' agency by speaking to and building on the signals they have already sent. They are not arbitrary relative to those signals (nor, even worse, do they actively shut down those concerns).
Within the Gygaxian skilled play context, here is an example of how things might go wrong (this is a hypothetical example):
* The PCs have explored a particular section of the dungeon, checking for metal and minerals with their wand, and detecting magic;
* They have detected a particular treasure horde, and casting Divination tells them that they have a reasonable prospect of success at recovering the treasure, provided they bring along some fire with them;
* They have memorised plenty of fireball and burning hands spells, as well as packing the wand of fire they'd been saving for just such an occasion;
* When the PCs open the door to the horde, ready to burn the anticipated enemies within, the GM tells them that a pit opens in the floor beneath them, and they fall through into some catacombs as the pit seals above them.
* In the catacombs, the PCs are assaulted by zombies; and the damp and water of the catacombs makes the PCs' fire attacks hard to bring to bear.
That is not a "zombies in catacombs" example that respects player agency. Depending on the details, there are different points at which agency was thwarted: if the pit was not written into the dungeon key in advance, then introducing the pit is the problem; if the pit
was written in, then the GM should have somehow alluded to that in the Divination result; if the pit was written in but is shielded by some sort of undetectable magic (so the Divination spell was "unaware" of it) then we have probably the most extreme version of the situation.
Generalising the examples: there are different ways of introducing content into the fiction, and framing the PCs into that content. Some of those ways respect player agency, others don't. You can't tell which just by identifying that what has happened is the GM framing the PCs into a particular situation.
their opinion is just as valid as yours.
There's no one right way to DM.
This is true, but surely it cuts both ways. In the context of this thread, it means that the idea of playing a game with a low degree of GM "latitude" (to borrow [MENTION=6696971]Manbearcat[/MENTION]'s terminology) in which the players have a high degree of ability to impact the fiction via action declarations for their PCs, can't be per se objectionable.
