GM fiat - an illustration

[hawkeyefan]

This is a big fundamental difference between the two philosophies.

Well, feel free to correct this:

I’m not going to correct it. All I’ll say is that you don’t seem to understand how any game except your own can be functional.

And then you provide examples of play that seem, at best, borderline functional.

You’re just way off.

 

[bloodtide]

One interesting thing about this post is that the first bit, about NPCs and game tokens and so on, appears to be a claim for the richness and sophistication of the fiction in the game.

Guess you skipped the paragraph that stated the other side? Comparing the rolling of 'generic NPC' reaction by the rolls and rules vs the DM creating a full non-player character, setting and world within a simulated reality.

But then the second bit, about the acid lake and so on, seems to describe a completely conventional dungeon-crawl type scenario with very little richness at all.

It does depend on how you define riches. To some it is money and gold, to others it is family and friends.

We could compare richness.....

The characters are in the woods and stop to take the Camp Action for the night. One character casts Alarm to protect the camp, forcing a modified camp roll and a look to the rules to be told the out come of "an old foe strikes for revenge". So the DM picks an old foe and has them send an assassin. Then a roll vs the Alarm has it go off and the characters wake up and kill the assassin, and find the helpful scroll that explains the 'revenge bounty' on their heads from the old foe.

Or....

The characters complete an adventure and foil the plans of Lord D'oom. Well, any player paying attention will know Lord D'oom is out for revenge. He has even said such things many times and made countless threats. And if any character would have asked around (doing the real role play acting type playing) they would have discovered Lord D'oom often makes good on his threats and tries to get revenge. And any character wishing to can learn of the Dark Assassins in Lord D'oom home city of Highport, and how they are for hire and how Lord D'oom often hires them. They can even learn about the assassins and the group.

So when the character set of into the wild woods, they could full well know that an assassin is on their trail. If they were paying attention and took the appropriate actions. But, it is just ask likely the players and characters are 100% clueless. Though even if they are clueless, the DM still knows all the details.

So when the characters stop for the night in a small clearing and cast the alarm spell.......the DM knows the assassin is watching nearby and even sees the casting of the alarm spell. As the spell 'alarm' is the most simplest and pathetic magic to guard a campsite and the spell is a very common spell known to all.....the assassin knows the spell very well. And part of the assassins training is "how to get around the pathetic alarm spell". So the assassin strikes by surprise...

.....Unless the good players outsmart the assassin. If they can figure out that he is nearby, or even if they think one might be nearby, they can take smart actions. They might cast more then one alarm spell....or better cast a more rare spell that they know the 'average' assassin would not know the details of or have a magic item made for this. There are a lot of things the characters can do to be ready when the assassin strikes...even find them and strike first, and all the best ones are not in any rules anywhere.

Toon is the only RPG I know of where this is true. Dungeon World, for example, is a game where the judgement of fiction on action is all in terms of the direction of the story, in favor of the character, mixed, or against. The GM also normally decides what specific move is entailed in a given action declaration, though they are constrained by the game's agenda and principles, as well as potentially the players disagreement.

Yet the GM and players are no more constrained overall, and it's not possible for a player to simply invent some arbitrary fiction, certainly not as a problem solving technique. In any case, it's not going to help them much anyhow, they're always going to be called on by the GM to take action, etc

I will just never get why the whole attraction and reason for playing the game is the whole, as you say, "they are constrained by the game's agenda and principles, as well as potentially the players disagreement."

And then say "Yet the GM and players are no more constrained overall". Well, except the constraints you mentioned? Right?

 

[thefutilist]

But if the GM is not presenting some overall direction, and yet all factors must be determined by said GM then either the evaluation of situation is very thin or the pace must be quite slow...

I'm not sure I am following you.

I posted an example of play that was similar to what happened in a game I played in a few sessions back.

T

Post in thread 'GM fiat - an illustration'

Mar 17, 2025

Resolution Method 2- dice are used to determine how the NPC reacts.
Result 2A- she does not forgive him. The GM considers the same factors as above and explains the NPC’s decision.
Result 2B- she does forgive him. The GM considers all the relevant factors and then explains why the NPC forgives the assassin.

My question to @thefutilist was how Resolution Method 1 resulted in an act of creativity by the GM that Resolution Method 2 somehow lacked. Because from what I can see, they both involve the same amount of creativity. The only difference is the means of the decision made...

• thefutilist

• 

Whenever I state 'that's a conflict' just replace rolling the dice with me deciding.

 

[Joanna Geist]

The GM doesn't have unilateral power to interrupt the spell in either version. Your party members are present. Hopefully they are at least smart enough to keep watch for sixty seconds while you cast a spell that serves their self interests.

The GM can fiat up a monster with a 90d covert rank, sure. But the GM can also decide that a mountain falls on you at any time, and this is true no matter what game you're playing. In any game where one party has infinite resources and others do not, fair play is required for the game to function at all.

Personally, as GM, I would be highly incentivised to have something set off the spell so that the player doesn't feel like he need not have bothered. That being said, random encounters don't exist by default in the system I use, so generally this effect wouldn't be used unless the player(s) think / know they're being actively tracked, or they're protecting someone or something, or some other similar situation where they have good reason to think it would be warranted. And there aren't hard limits on power usage, so a wasted use wouldn't feel nearly as bad.

On the other hand, there was another game I ran where I used a hex map to track the movements of party members and various allies and foes. In that game, I would keep track of the locations of monsters and their movements on the map, so no rolling would be required to determine if they encountered a player.

 

[Manbearcat]

You're turning over a ton of control to randomness in this hypothetical challenge focused system. The appeal of the alarm spell is precisely that it does not interact with randomness in that way; it requires specific countermeasures, and if the opposition doesn't use them, they fail. The analogous situation (and presumably desired feeling you're trying to evoke) in a competitive game might be something like the wizard player spending currency to search their deck and then holding an alarm card in hand, that will cancel any ambush subtype card the hunter plays. Rolling 2 modified dice pools to derive the resulting fiction can't produce that kind of byplay except in retrospective retelling. It's a different kind of artificiality than what you're critiquing in "fiat" here, but it's no less a problem.

The easy test is to ask "what would a worse wizard have done here?" And then to explain why it is a worse line of play, and why it's stable that such a line would be worse regardless of how the random breaks went.

Obviously, it would be ideal to have the hunter portrayed by a player separate from the GM, so you can actually put the NPC/PC choices into direct conflict to see what occurs, but that's impractical for most modes of play, so we have the GM simulate competition. It's imperfect, but preferable if you want players to be able to keep discriminating between strategies in unfolding situations.

Just going to fire off some thoughts/clarifications/rejoinders as we have a pretty significant amount of daylight between us:

ON RANDOMNESS:

I'm going to tag @thefutilist here rather than replying to his post to me as some of what I'm saying here engages with his response.

Though my post you replied to doesn't enumerate the entirety of what I would be conceiving in such a system (which would either be Torchbearer flat-out or kindred), I want to be clear that your conception of such a system would be (a) short-shrifting the layered and longitudinal decision-space (giving significant expression and consequence to both tactical and strategic play) inherent to deciding to cast the equivalent of an Alarm spell while (b) also imagining an increased signature on play of fortune resolution ("randomization"...which also isn't total randomization as the decision would entail mustering currency, possibly proactively generating toll on character, while simultaneously engaging with advancement scheme in a thoughtful way). I don't mean to point at you directly, but I definitely feel like there are some confounders to your sense of how this comes together holistically in various games. For example:

* Various forms of Poker feature the randomization of drawn hands, drawn cards, and blind hole cards (which may be blind to opponents or even blind parts of a hand that you are to play). My sense is that your hypothesis on these kinds of games should be that skill is mostly noise and randomization is the signal. Even in relatively "simple" (not simplistic) games of Poker, this is just not true and results bear that out. Better players win more than worse players and the best players in the world and at the table dominate. You'll see a very stray rogue win by someone who isn't on the level of a great or world class player, but overwhelmingly tables are won by better players and the long haul of strategic play over time absolutely bears out that better players (and certainly world class ones) dominate results.

* The game you're imagining (some kind of deck building game akin to MtG or something) has a host of consequential, randomized elements that you're eliminating from your scenario and its analysis. These include (i) match-up, (ii) starting hand drawn, (iii) early game draws, (iv) resource for ramping and "curving out", (v) opponent's own (i-iv) which lets them interrupt your sequences/curve-out/meta-scheme, and even (vi) fortune resolution within play (some novel cards/games feature dice rolling along with the other layers of play). Yet, like Poker, these random elements don't swarm over the Skilled Play valuation of games individually and certainly over a sequence of games. Controlling for ridiculous match-up problems, it is a fairly trivial matter to suss out poor players from good players and that only increases as you move toward a tail (the elite and world class players are profoundly better than average players and examination of play easily susses that out).


ON GMING SIMULATING COMPETITION

While it isn't in your face as a physical randomizer like fortune resolution, I also think you're significantly short-shrifting the significant signature of randomness that is (even an expert) "GM simulating competition" in the way you're imagining. To unpack that "in the way you're imagining" (presumably here, based on this conversation and prior ones), I presume you're referring to (i) mental modeling of extraordinarily complex systems, (ii) articulating the complex systems' dynamics to amateur and expert player alike in such a way that those players can functionally perform their OODA (Observe > Orient > Decide > Act) Loop based on the GM's description of fictional positioning, then (iii) employ some kind of principled opposition regime in a way that sufficiently corrects for personal bias while engaging with and deploying (this GM's perception of...which also must be functionally conveyed to players) fictional positioning & simulation scheme-derived countermeasures.

I can't possibly convey the depths of my skepticism that this GMing and play paradigm won't inject play with a randomness element that significantly outstrips any conceived Skilled Play signal (particularly in contrast with the first form of play as in Torchbearer). My take is simply that deriving skillfulness of play in this proposed paradigm becomes extraordinarily fraught because of the following reasons:

* I've never seen evidence that humans are capable of mentally modeling extraordinarily complex systems in a way that is either sufficiently accurate to be gameable moment-to-moment or consistently coherent with respect to their ability to convey long term regimes of causality in such a fashion that is gameable longitudinally. Complete reliance upon fictional positioning derived from this modeling is a fantasy, a unicorn, imo.

* Inherent to the above bullet point isn't just the mental modeling requirements; the expertise, the cognitive load, the necessary mental bandwidth, the ability to detect and correct for personal bias. Equally as important is the ability to very effectively communicate extremely complicated information sets to both amateurs and other experts. And not just communicate...but communicate in such a way that presents a consistently stable and rich gameable environment. Again, imo a fantasy...a unicorn, imo.

There are probably 5ish areas where I would qualify as an expert. Two of them are physical. Merely communicating with other experts (in which we are presumed to share similar knowledge sets and common language) is already difficult enough. Now add in the requirement to generate and communicate a gameable space? Woof. Now throttle to folks in which I don't share a similar knowledge set or a common language of understanding (amateurs or thereabouts)? Double woof.

IMO, something (like codified rules, principles, and procedures) has to do the work to create a functional User Interface for us.

* That personal bias component? Multivariate and just enormously fraught. Among those many are (a) the social layer and (b) competing gameplay priorities, both of which can also be broken out into many subsections. On (a), you have so many parts of this from social pressure regimes (direct and indirect, spoken and unspoken, aware of and oblivious to) to entrenched social cohort dynamics to various personal traits around disagreeableness and approaches to conflict. On (b), this is a huge area of problem in the TTRPG world since forever. I've broken this out so many times and in such excruciating detail that I'm not willing to do it again. I think people know what I'm talking about here. Any game that is possessed of specific storytelling imperatives (whether it be the GM's or AP's unilateral imperatives or consensus-based imperatives or social pressure-based imperatives) and/or is pressured by pacing imperatives or real-world temporal imperatives (like at a Con or even a home game that "wants some juice" before end of session) becomes a problem when competing with Skilled Play imperatives (and compete they will...not every moment, but they absolutely will collide and compete for control over moments and thereby trajectory of play). I'm not even going to get into (c) intrapersonal vs interpersonal Immersionist dynamics, but that is another can-of-worms here.

---

Anyway, that is a lot of stuff to engage with. Net TLDR:

* Randomness is loaded everywhere, even when not a piece of ephemera. And it can be "corrected for" and/or it can be rendered only a small parameter in play.

* I don't believe governing play via the vessel of mental modeling and fictional positioning conveyance can yield a consistently functional and coherent Skilled Play paradigm. For @thefutilist , with respect to Narrativism, I will agree that it can work to a degree. However, I would say that management of fictional positioning and negotiation of fictional positioning whereby some unilateral ("GM decides") or multilateral (negotiated consensus) derives outcomes of conflicts is not satisfying play for me. I would say that in this paradigm you are balancing on an extraordinarily precarious fault line moment-to-moment; one slip and you lose the necessary dynamic of all participants being surprised by play's trajectory and risk character evolution being an outgrowth of (let me call it) "determinism creep." Those are both anathema to Narrativism.

I hope this entire post doesn't come across too strongly...I've been thinking about how to post these thoughts without coming across as too direct and just douchey...but that is the best I've got. If too direct and douchey, I'll just go ahead and wear it!

 

[bloodtide]

The GM can fiat up a monster with a 90d covert rank, sure. But the GM can also decide that a mountain falls on you at any time, and this is true no matter what game you're playing.

Not true. Many games have rules saying what a DM can do in the game, as they are just a "player-DM". And sure...in theory....they can 'do anything', but they believe in the game rules SO much, that they never would as they believe they must always follow the rules.

 

[hawkeyefan]

But the GM can also decide that a mountain falls on you at any time, and this is true no matter what game you're playing.

I’d like to point out in a more rational manner that not every RPG works this way. There are many games that limit GM authority, and do so not to hamstring the GM, but to enhance the play experience.

In Apocalypse World and many of its derivatives, the GM can only make a move at certain times, and the severity of the move is based on what triggers that move.

This isn’t to limit the GM in some arbitrary way… it’s to keep the game focused on what the players want their characters to do.

 

[Manbearcat]

Not true. Many games have rules saying what a DM can do in the game, as they are just a "player-DM". And sure...in theory....they can 'do anything', but they believe in the game rules SO much, that they never would as they believe they must always follow the rules.

While I think your usage of "player-DM" and some of your other descriptors are clearly meant as pejorative, I do want to say that, as someone who has never played these games, you've actually a fairly solid job of talking about these games in this thread. Not perfect by any means (more on that below!), and a little incendiary, but solid enough to help further discussion and clarify some stuff. So here is to you bloodtide of the internet:

Now a quick clarification.

GMing Torchbearer (what the lead post was discussing and you were responding to, so let us go with TB) is nothing like being a "player DM." While they are rather different than the Traditional or Classic model of GMing as a whole, you do have some overlap on that Venn Diagram, and you do have extremely significant inputs on play as a TB GM. To wit:

* You're either using TB's stock setting of Middarmark or you're coming up with your own in a procedural, principled manner just as you would devise a hex map and stock it with setting/game content. You're talking "civilized" locales and their (fiction and mechanical) dynamics, distances between places (relevant for inter-Adventure site journeys or Toll if using that system), and Adventure sites (which folds in all the typical things a GM must resolve including the TB-specific things).

* While Adventures have very specific, bounded instructions whether Short/Medium/Long (on # of Obstacles, on Obstacle ratings, on Camp site dynamics, et al), the GM devises them.

* While the GM should be using either (i) Condition w/ Success or (ii) Twist in equal proportion when a player Test generates a Failure, (a) it is the GM's decision and (b) while the nature and number set of Twists are bounded by fictional positioning, there is still a lot of play for the GM here.

* While players have a large amount of input on types of Conflict, stakes, and Compromise post-conflict, the GM has equally as much and, in some cases, more than the players. This is a pretty significant component of play, but there is nuance here and breaking this down would be more than I have time for (and I would need to see some interest and investment from folks before I care to do so).

* Like Twists, Town and Camp Event Tables and their results are bounded. Nonetheless, there is some interpretation here giving GM's some (principled and constrained) "play."

---

Net, a TB GM isn't even close to a "player-GM." TB is by far the most difficult game I've ever run for players. The attrition rate and the duress upon players is so profoundly is beyond Classic D&D (and I spent '84-99 overwhelmingly running dungeoncrawls and hexcrawls) that it is difficult to even contrast the two. It is just that (a) TB is also an incredibly vital Story Now engine along with being a Gamist masterpiece and (b) the expression of agency and decision-space a TB player is managing is quite different from Classic D&D.

 

[bloodtide]

This isn’t to limit the GM in some arbitrary way… it’s to keep the game focused on what the players want their characters to do.

Though note it is a Limit.

In my game, as DM I can do whatever I wish on a whim with absolute power.

In the Limited game, often enough, a DM-Player will just sit there and say something like "I can't take any actions in the game as the rules on page 11 say I can't!".

And sure, lots of players love this. They can dance around the table and say "nah nah, the player-dm can't touch us!". I'm sure this makes for a great, fun game for the players.

 

[Joanna Geist]

Not true. Many games have rules saying what a DM can do in the game, as they are just a "player-DM". And sure...in theory....they can 'do anything', but they believe in the game rules SO much, that they never would as they believe they must always follow the rules.

I’d like to point out in a more rational manner that not every RPG works this way. There are many games that limit GM authority, and do so not to hamstring the GM, but to enhance the play experience.

In Apocalypse World and many of its derivatives, the GM can only make a move at certain times, and the severity of the move is based on what triggers that move.

This isn’t to limit the GM in some arbitrary way… it’s to keep the game focused on what the players want their characters to do.

It doesn't actually matter what the book says. The GM can still do whatever he wants. Certainly, he may find himself without players and/or with additional black eyes, I'm not disputing that. But he still has the capability.