@Bedrockgames @robertsconley
One thing that might be helpful is that the OP of the thread talks about the d&d alarm spell as fiat for a variety of reasons. My belief is that the spell in d&d is actually resolved through many of the same kinds of procedures you’ve mentioned. It would be interesting to see that broken down specifically in response to the points in the OP.
Sure, I will take a crack at it.
One thing that might be helpful is that the OP treats the Alarm spell in D&D as an example of GM fiat. My view is that Alarm is resolved through many of the same procedures we've been discussing. I’ll break that down directly in response to the points made.
The example in @pemerton’s post is not relevant to the issue of GM fiat because the two spells, Alarm and Aetherial Premonition, are not equivalent in any meaningful way.
Alarm creates a magical tripwire. It alerts the caster if a creature enters or touches a specific area. It lasts for 8 hours and doesn’t require concentration. It doesn’t alter the world, shift probabilities, or guarantee a specific outcome. It provides information if a defined trigger is met.
Aetherial Premonition is something else entirely. It manipulates the odds. It alters the outcome of the camp event roll (+1) and directly influences future tests (+1D to avert disaster). It’s not just awareness, it changes reality. That’s the difference between that reacts and give notice versus the active alteration of reality and fate.
A proper 5e equivalent would be a spell that modifies random encounter rolls for 8 hours and gives the party advantage on checks if an encounter occurs. That’s not what Alarm does. The comparison fails, and any argument that hinges on their equivalence doesn’t hold up.
That said, Alarm is still worth examining for how it interacts with referee's procedures in play.
Moving on, I’m not sure what @pemerton’s actual conclusion is. He notes that D&D 5e relies on a referee's judgment and that Torchbearer bakes risk management into its mechanics. That’s fine as a comparison of system design. But his only clear assertion is: “The GM is not at liberty just to narrate things in such a way that the spell makes no difference.”
This raises the question: what is the argument here? That if a system allows GM fiat, then spells cast by players are meaningless because the GM can just handwave or manipulate the context? If that’s the point, it’s weak. It assumes bad faith refereeing and dismisses structured adjudication as irrelevant. In World in Motion play, that’s simply not how things work. The GM doesn’t "decide what happens" based on dramatic effect, they resolve outcomes based on what logically follows from the established state of the world.
Let’s address this idea of “illusion of agency” head-on.
@pemerton implies that Alarm looks like it gives players control, but really leaves everything up to GM discretion. That only holds true in games where the world is being improvised moment to moment. In a sandbox or procedural campaign, player agency comes from interaction with a consistent world model. Spells like Alarm are part of that, tools to engage with known threats, secure known spaces, and respond to plausible events. If Alarm triggers or doesn’t, that’s the result of where and how it’s used, not GM caprice.
Also, the “fair play” framing doesn’t hold up under scrutiny. In World in Motion campaigns, outcomes aren’t dictated by what the GM thinks is fair. They’re shaped by prep, causal logic, and resolution procedures. Encounter checks, travel timelines, NPC behavior, all of it is grounded. The referee’s job isn’t to decide whether a spell matters; it’s to determine whether conditions arise where it applies. That’s not fiat. That’s consistency.
Alarm is not just a camping spell. It’s useful whenever a party wants to monitor access to a fixed area, during a heist, in a siege, in a tomb, wherever. The 8-hour duration invites periodic encounter checks. Many referees using World in Motion structure those checks with dice, schedules, or faction behavior, not improvisation.
For example, if the party camps in a dead-end cave and casts Alarm at the entrance, any encounter approaching the cave will trigger the alarm. There’s no ambiguity. If the same party is raiding a building with multiple entrances, and they only ward the front, then yes, someone might come through the side. But that outcome is based on layout, planning, and scouting. It isn’t “GM fiat.” Its consequences that are based on player decisions.
In this style of play, the spell is one move in a larger strategy. The tension doesn’t come from whether the GM decides to honor the spell’s effect, it comes from whether the party planned well, accounted for the space, and took the right risks. That’s agency.
Responding to some of @pemerton’s points:
Does the player’s character have an uninterrupted minute to cast the spell?
This is trivial to adjudicate. If the party is setting up camp or preparing a location, one minute of uninterrupted time is not difficult to justify. Unless the party is threatened directly, there’s no reason to deny the casting.
Does any potential intruder come within 8 hours, or 8 hours and 5 minutes later?
Many referees have procedures or notes that define what happens hour by hour. Encounters aren’t created on a whim, they’re generated or determined from the current state of the setting. As for getting around Alarm when the encounter happens, that depends on the location and circumstances of the encounter. Again, this is based on the current state of the setting and the characters involved, not by GM Fiat.
Does the intruder enter the warded area, or go around it?
Again, this is based on the environment and the party's choices. Did they scout? Are there known side passages? Is the alarm placed strategically? These are tactical decisions with consequences, not fiat.
If the party is asleep and the alarm wakes them, how much can the intruder do before they react?
5e’s surprise and initiative rules cover this. Alarm alerts the party, likely preventing surprise. The combat system handles What happens next. Nothing about that depends on arbitrary GM narration.
In short, Alarm functions meaningfully in play, not because the GM “lets it,” but because it's used in a world where time, space, and NPC behavior are structured. If a referee is making things up as they go, then yes, Alarm can be rendered meaningless. But the same is true of any tool in that kind of campaign. That’s not a problem with the spell, it’s a problem with the lack of underlying procedure.
In contrast, Torchbearer’s Aetherial Premonition encodes its effects directly into resolution mechanics. That’s a design choice, not a moral high ground. Both styles work. But it’s inaccurate to say Alarm relies on fiat unless the entire game already does.
