TL;DR: This post explores the subtle art of not being a dick GM from a slightly different standpoint than my usual Fate, and if you don’t care about my geeky reasons, you can jump right to it.
Discussions of the “Fiction First” approach to RPGs often equate rolling dice “when it’s interesting only” with “as seldom as possible.” Two narrative RPGs—Apocalypse World (AW) and Cortex Prime (CP)—reverse that perspective. Both obey the logic “if you do it, you do it, so make with the dice” (AW, p. 136). AW does it for PCs only because AW’s GM never rolls but CP extends it to anything fictionally meaningful.
However, (dis)similarities between AW’s and CP’s dice-rolling policies are tangential to this post’s topic. I’ve cooked some Fate-AW weaksauce in the past, but today, I’m taking that to the next level with AW’s Threats and CP rather than Fate, because CP’s “Bronze Rule” makes the sauce hotter.
As customary, there are preliminaries: my interpretation of AW’s GM advice. If yours differs, you may not like what I do with Threat mechanics. Still, you can skip to the mechanical part, a glorified worked example.
Decision-making and agent-based modeling
A few years ago, I shared an office in a cognitive science department with a computer-game programmer hired to add intelligent agents in computer-assisted learning software. His expertise was procedural generation, and, as a side project, he had a vision for a “fantasy realm simulator” with NPC quest-givers who could also be independent actors.
The exchanges we had about that project gave me an “a-ha” moment when I read the “Master of Ceremony” chapter in the AW ruleset for the first time. But for exposition purposes, I’ll start with AW.
Is “disclaim[ing] decision-making” bullshit?
The AW Chapter titled “The Master of Ceremony”—AW in-system moniker for “Game Master/Moderator,” because why not—splits the advice on how to run the game into three categories: the MC’s Agenda, their Principles, and their Moves. The third is specific to AW, but the first two are not. The Agenda works “as is” in any narrative game substituting “the setting” for “Apocalypse World.”
Make [the setting] seem real.
Make the players’ characters’ lives not boring.
Play to find out what happens.AW 2E, p. 80.
The Principles (p. 82) are a mix-bag of generalities (“be a fan of the players”) and AW-specific advice (“make your move, but misdirect”). The last one falls in the “generalities” category and is the one I’d like to focus on.
Sometimes, disclaim decision-making.AW 2E, p. 82.
D. Vincent and Meguey Baker split “disclaim[ing] decision-making” into four options.
- Put decision-making into NPCs’ hands.
- Put decision-making into PCs’ hands.
- Create a countdown.
- Make it a stake question.
On the surface, (2) is the only case of “disclaiming” anything. NPCs are decision-makers only by proxy for the MC, and countdowns are set by the MC, so (1) and (3) disclaim nothing. As for (4), answers depend on interactions between PCs and NPCs, so (ultimately) the MC. Thus, on the surface, 50-to-75 percent of “disclaim[ing] decision-making” is questionable at best, if not outright bullshit.
And before any PtbA-zealot gets to my throat: “bullshit” has a technical sense in philosophy—in short, the misrepresentation of intentions for ulterior motives. Here, that would be the old D&D-like Dungeon Master on the lookout for TPK opportunities under the guise of: “Hey, it’s not me being a dick! I disclaimed decision-making!”
However, a good-faith reading supersedes the “bullshit” interpretation and is consistent with the other principles. This brings me back to my programmer friend.
The agent-based fantasy realm
Imagine a village with a blacksmith who offers a bounty for taking out a rampaging ogre, but could take matters into his own hands if, say, no adventurer showed up before Winter’s end. If the first adventurers to visit the village come with the Spring, they’ll find either a blacksmith with no bounty to offer or village folks in need of a blacksmith.
Capturing that for some software implementation is, in principle, relatively simple. The first step is conceptual.
- First, endow the blacksmith with a preference for upsetting the status quo, e.g., “The ogre must be taken care of.”
- Second, give the blacksmith three options: do nothing, offer a bounty, or go after the ogre.
- Third, specify the blacksmith’s expectations about what would happen to the status quo with each option.
The blacksmith would behave as intended if: (1) he believes adventurers have better chances against the ogre; (2) he is also willing to take chances if no adventurer shows up, and: (3) urgency may outweigh his doubts. Equivalently, if a time parameter reverses the blacksmith’s preference between “offer a bounty” and “go after the ogre.”
Throw in a few adventurers with preferences (for types of jobs and pay), options (take the job, or pass), and expectations (chances against ogres). Handwave extra parameters, and you have sketched an agent-based model (ABM): a game—in the sense of game theory, minus the mathematical details—where agents, initially defined by preferences, action sets, and beliefs, can be left to interact based on those initial determinations.
There are good machine-learning software packages for implementing ABM simulations, but CRPG software implementation of AI-based ABMs is a bit of a nightmare (you know what I mean if you’re familiar with the development history of Cyberpunk 2077’s open world). So my friend’s ideas were, and still are, borderline pipe-dreamy.
But a TTRPG “wetware” implementation of ABMs is relatively straightforward. In fact, AW’s “disclaim decision-making” boils down to an ABM of the game world—the good-faith reading I alluded to—and the reason AW’s MC chapter is no-bullshit gold, after all.
If you skipped the first part: AW’s Threats are set up so the GM can deduce rather than decide how a Threat will behave when coming into play. Perhaps more accurately: Threats are pre-programmed to act in roughly predictable ways. The GM’s the programmer, so NPCs are GM’s creatures, but AW’s guidelines limit arbitrariness—the temptation to alter the programming to fuck the PCs over. Any translation should preserve that.
AW’s programming language for Threats
In AW, the logic of Threat-programming is: at session n, PCs’ actions lead to introduce an NPC; the GM names the NPC and makes them human (one of the principles of AW); then, between session n and session n+1, the GM fills a Threat Card for the NPC (Fig. 1). You can read the programming language off of fig. 1.
Kinds and their subtypes determine impulses (preferences) and a base list of moves (action sets). I’ll handwave the details, work with a cast-of-one example, and utterly disregard stakes, countdowns, and the map. I’ll add a custom move, so if you’re not familiar with AW, here’s for a crash course on the mechanics. And now for the example.
Monk is a grotesque: pain addict [kind], so his impulse is to seek pain. Maybe the move I choose for him just now is attack someone from behind: “Damson, someone steps up behind you in the line for showers and loops a wire around your throat. What do you do?”AW 2E (p. 115)
The list of Grotesques’ moves (AW, pp. 109-110) defines what a “grotesque” is and functions as a set of implicit if-then instructions. A GM can “disclaim decision-making” by checking a given situation against the list. If a move matches the impulse, they can make it. Otherwise, they can’t. But there’s residual GM arbitrariness that’s worth addressing.
Damson’s default move would be to act under fire (AW 2E, p.136). The failure outcome is “prepare for the worst,” whereby Damson could suffer harm. If Damson’s player fails that roll too, the GM’s options include taking Damson straight out. This “move snowball” is consistent with Monk’s seek pain. But a GM could choose other moves or outcomes. So what if I settle for those, and take Damson out? Am I being a dick? I’d prefer not to leave doubt about the answer. So here’s how I’d do it.
First, I’d clarify beforehand that I’m not a dick, but Monk is—announce future badness (AW 2E, p. 89) is the move for that. Second, I’d add a custom move for Monk. Both additions are displayed Fig. 2 on a custom Fari card. For details about rolls, see here; 3 harm (ap) [armor piercing] guarantees (at least) an injury that “won’t get worse or better by itself” but is only 2-harm away from “life [becoming] untenable” (AW 2E, p. 265).
The Cortexification of Threats
As hinted at in the intro, CP extends AW’s motto, “if you do it, you do it, so make with the dice,” to anything fictionally meaningful. Mechanically, any such thing has at least one Trait, rated by a die (between d4 and d12) that GM and PCs can include in dice pools. However many dice are in a dice pool, the general rule is “roll n, keep two” for total. On most occasions, you’d also keep a third effect die—but for that one, only size matters (insert NSFW joke).
In CP, the Monk-Damson situation would likely begin one step earlier than in AW, with Monk sneaking behind Damson, intending to get A Wire Around Damson’s Neck. That way, Monk’s total would set the difficulty for Damson’s roll. If Damson’s player’s total does not beat Monk’s, there’ll be A Wire Around Damson’s Neck—a Trait, rated as per Monk’s effect die. Mechanically, the conversion of Monk-as-a-Threat is dirt-simple:
- Monk’s impulse in AW, Seek Pain, becomes a Trait, that the GM would add to a (base) Dice Pool to get Monk’s total.
- The GM move in AW, attack someone from behind, becomes the narrative description of Monk’s action.
CP has no need for the custom move, and no direct conversion is needed. But I’d convey Monk’s lethality adding Dick Move as a special effect (SFX). I’d have issues with the arbitrariness of introducing Monk behind Damson in CP, so I’d translate the constraint to announce future badness with a cost for introducing Monk. Fig. 2 displays a Threat card with these two options (see The Fine Prints for mechanical clarifications).
Good (+3) question. Here’s a quick conversion.
- Forget about the cost.
- Make Seek Pain an ad hoc Skill Fair (+2)—d8 in CP is better than Average (+1)/ but not really Good (+3)/ or Great (+4)/.
- Make Dick Move a Stunt that gives +2 to Create an Advantage with Seek Pain when Attacking someone from behind.
- Treat A Wire Around Damson’s Neck as an Aspect Monk would create with Seek Pain.
A warning: the conversion would work for other Threats, but with some bronze-plating. Specifically, Afflictions, Landscapes, Terrains, and Vehicles are Threat kinds in AW. Also, some custom moves/SFXs best translate as “Situation Stunts” which are not a thing in vanilla Fate anymore.
Then again, from my experience with both Fate and CP, Fate runs smoother with fewer rolls. Its design principles are almost the opposite of “if you do it, you do it, so make with the dice.” So just because you can bronze-rule Threats does not mean that you should.
Arbitrariness Creep. “Prepare for the worst” means the GM can make as hard a move as she wants, not that she has to. So it’s GM fiat if she picks the worst result—and that’s arbitrariness creep. Still, this is no different than setting circumstances for an if-then threat move to apply. A good-faith GM would not conjure Monk out of thin air but would merely need a good-faith answer to Damson’s player’s question: “Wait, how did Monk sneak up on me?”—that’d be anything better than “Because I said so.” So, in the same vein, here’s a rule of thumb for NPCs’ motives and, generally, hard moves. First, imagine the players asking, “Wait, why did you play [hard move] on me?” Second, picture yourself replying: “Well, why would [NPC] inflict [hard move] upon you?” If you can’t put yourself in the players’ shoes and find an earlier announce future badness that makes the answer obvious, [NPC]’s not being a dick: you are.
Monk’s Dick Move. CP uses the roll-and-keep method exclusively (without flat modifiers) and alters probabilities with “dice tricks” triggered by narrative permissions. Monk’s Dick Move is one such, based on the notions that: (1) all 2dn results are on a Bell curve; (2) more dn in a pool increase the probability of a (two-die) total equal to the average of 2dn; and: (3) more dn in a pool make more dn available for effect. With a d8 rating, Monk is better than average (d6), but not exceptional (d10)—ditto a d8 Wire Around Damson’s Neck complication. The Doom Pool is a GM option whereby, whenever a PC rolls a 1, the GM can “buy” the die for a Plot Point (CP’s meta-currency). Dice bought that way go in the GM’s roll-and-keep dice pool (e.g., to which Seek Pain would be added). Alternatively, the GM can “spend” them for introducing (the equivalent of) new Threats, or for one-time dice tricks.
Wrapping Up: Don’t be a dick
In AW’s implied setting, everything the GM preps for (NPCs, locations, etc.) is a threat. And a good chunk of the GM advice in AW is about making the world’s threats go for the PCs’ throat without being a dick about it. Thus, AW’s approach translates well to any game-setting where the world is dangerous, but the GM’s endgame is not TPK. I love that kind of setting, so I love Threats.
I also like the “if you do it, you do it, so make with the dice” approach as both a player and a GM, and I see expendable resources (for players or GM) as an extension of that. In my experience, the combination can help folks who have issues with the language-centric features of Fate.
But that’s because of how my brain works (quite literally). It’s a matter of preferences, and although they may correlate with how different folks use cognitive resources, I don’t have direct evidence for that. So that’s a conversation for a Theory Thursday, at best.
And so, that’s all for today, folks.