GM Guide

The GM's Role

AX.C.14.01

The Game Master's job in AxiomRPG is to run the world, adjudicating the rules when dice are needed, narrating everything the player characters encounter, and facilitating the session so that play moves and everyone at the table has something to do. These responsibilities overlap constantly, at any given moment the GM may be doing all three at once.

None of these roles requires the GM to be an adversary. The dice system generates dramatic variance on its own. An explosion on the wild die produces extraordinary outcomes the GM didn't plan. An implosion on a critical roll creates consequence the GM didn't engineer. The GM's job is to set conditions and interpret results, not to manufacture drama by adjusting outcomes after the fact.

Core Function

Adjudication is the GM's rules responsibility. When a player declares an action and the outcome is uncertain, the GM determines whether a roll is required, selects the appropriate dice pool and threshold, and interprets what the result means in the fiction. This includes knowing when not to call for a roll, trivial actions, impossible actions, and actions where the outcome doesn't change anything in the scene don't need dice. Over-rolling slows play and dilutes the significance of rolls that actually matter.

Narration is the GM's world responsibility. Everything outside the player characters, NPCs, environments, threats, factions, time passing, is voiced and described by the GM. Narration answers the question what do you see, hear, and know. Good narration is specific enough to be useful and brief enough to leave room for play. A room described in two sentences with one distinctive detail serves the table better than a paragraph of atmosphere.

Facilitation is the GM's table responsibility. Pacing, spotlight distribution, and forward momentum are managed by the GM. This means noticing when a scene has given everything it has and cutting to the next one, tracking which players haven't acted or spoken recently, and keeping the session moving toward something. Facilitation is invisible when it's working and obvious when it isn't.

Working with Genre Catalogs

AxiomRPG core rules define the universal mechanics: dice pools, thresholds, attributes, talents, saves, health, and advancement. A Genre Catalog applies those mechanics to a specific setting, naming the lineages, defining the power traditions, establishing the factions, and providing the profession library appropriate to that world.

The pattern is consistent: core rules are the constant, the genre catalog is the variable.

Before running with a Genre Catalog, a GM should have working knowledge of four things:

• Power tradition: How Odd Talents function in this setting, what they cost, and any restrictions on use. The Iron Lattice catalog uses Harmonic Traditions with Resonance Stress as the use cost, the first use per scene is free, subsequent uses roll at disadvantage until the character takes a ten-minute rest.
• Lineage mechanics: Whether lineages carry mechanical differences beyond the core archetypes, Health modifiers, cultural talents, special abilities.
• Faction structure: Who the major powers are, what they want, and where they're in conflict. This shapes NPC motivation throughout the campaign.
• Setting-specific rules: Any mechanics the catalog introduces that have no core analog, vehicle rules, environmental encounter conditions, digital-space encounters, and so on.

Where the Genre Catalog is silent, apply core rules. Where core rules don't cover it, rule by intent. Document rulings that come up repeatedly; they become local precedent for that campaign.

Table Calibration

Before the first session, the GM and players should align on what kind of game they're sitting down to play. It doesn't require a lengthy formal process, just answers to a few direct questions.

Tone: Is this setting grim or hopeful? Is failure likely to be permanent? What's the emotional register of the fiction, tension, action, drama, comedy, or some combination?

Scope: Are these characters beginning their story or stepping into an ongoing world? Is the campaign self-contained or open-ended? How much does the broader setting matter relative to personal character arcs?

Content: What's off the table? Topics, scenarios, or types of content that players don't want in the game should be named before they come up during play, not after. Safety tools, techniques that allow players to signal discomfort or redirect scenes without breaking the fiction, are appropriate at any table regardless of tone or setting. Establish which tool the group will use and how before play begins.

Challenge level: How lethal is this game? Are threats intended to genuinely threaten character death, or is death a rare consequence? This affects how players invest in their characters and how aggressively the GM designs encounters.

These questions take ten minutes at session zero and prevent significantly more disruption later.

Balancing Challenge and Agency

Across all sessions, the GM holds two responsibilities in tension: the game should be genuinely challenging, and player choices should genuinely matter.

Challenge means that outcomes are uncertain and that failure is possible. A game where the players cannot fail is not a game; it is collaborative storytelling with dice as decoration. The dice pool system supports challenge naturally: even a strong pool can produce a bad roll, and the wild die introduces variance that neither side controls. The GM's job is to present obstacles that test character builds and player decisions, not to punish players for choices made in good faith.

Agency means that what the players choose to do changes what happens. Choices that don't matter, where all roads lead to the same result, drain investment over time. Players should be able to see the fingerprints of their decisions on the world. NPCs should remember what happened. Factions should respond to what the players have done. Events in one session should affect conditions in the next.

When challenge and agency are both present, failure becomes generative rather than terminal. A failed roll that closes one door but opens another, or that succeeds at a cost, keeps the fiction moving. A failed roll that simply ends the scene or removes a player from meaningful participation for several rounds does not.

The GM's ongoing calibration question is: does this outcome produce more story, or does it stop the story? Default toward the former.

Adjudication

AX.C.14.02

Where AX.C.02 defines what the dice do, this section covers when to use them and how to interpret what they produce. Adjudication is the moment-to-moment translation between the fiction and the rules, the GM's most constant responsibility at the table.

When to Call for a Roll

The criterion for calling a roll: the outcome is uncertain, and both success and failure produce something worth playing. If either condition isn't met, don't roll.

Three categories that don't need dice:

Trivial actions, A trained character doing something well within their capability against no meaningful resistance. A skilled climber ascending a ladder. A soldier drawing their weapon. A scholar recalling a fact from their own specialty. Calling for a roll here communicates that the task is harder than it looks, which is either wrong, or a lie the GM will have to sustain.

Impossible actions, No pool size will produce the needed outcome. A character cannot pick a lock they have no picks for, or convince an NPC established as incorruptible to accept a bribe. Rolling and failing communicates the same thing as declaring it impossible, but wastes time and sets a misleading precedent that dice could have helped.

Inconsequential actions, The outcome doesn't change anything about the scene. Whether the character succeeds or fails at identifying the paint color on a wall, the scene continues identically. If neither outcome matters, the roll doesn't matter.

The over-rolling problem: Every roll carries an implicit stake, the table is paying attention because something might go wrong. Frequent calls for inconsequential rolls train players to look for risk where none exists, and to feel cheated when rolls produce nothing. Reserve dice for moments that deserve them.

Setting Thresholds

The Standard Threshold is 2, the default for most non-combat challenges. It produces roughly 75% success on a 6D pool and scales naturally as dice pools grow.

Threshold	Difficulty	Typical Conditions
1	Easy	Favorable circumstances, trained character, no meaningful resistance
2	Standard	Normal conditions, outcome uncertain
3	Difficult	Adverse conditions, active resistance, or missing tools
4	Hard	Severe conditions, strong opposition, multiple compounding factors
5+	Epic	Near-impossible; extraordinary capability required

Threshold reflects the obstacle at hand, not a general desire for tension. The most common modifier is active resistance, when an NPC or threat is actively working against the character, add 1. Unfavorable environment (poor visibility, time pressure, difficult terrain) adds 1. Attempting without required tools or preparation adds 1. These stack, but threshold rarely needs to exceed 4 outside genuinely exceptional circumstances.

Example: Picking a lock is Standard (2). Picking an unfamiliar lock under time pressure without proper picks is Hard (4): +1 for the unfamiliar mechanism, +1 for inferior tools, +1 for the approaching patrol.

Set the threshold before the roll with the full picture in mind. Don't adjust it mid-roll because the fiction shifted.

Reading Results

Clean success (met threshold): The action succeeds as declared. Narrate what happened and move the scene forward. The size of the pool doesn't change the quality of the outcome, meeting the threshold is meeting the threshold.

Excess successes (exceeded threshold by 1 or more): The character succeeded with margin. This is the GM's invitation to add something, a piece of additional information, a positional advantage, a detail that carries forward into the next scene. Excess successes aren't a separate mechanical system; they are narrative latitude the GM can use or leave alone.

Narrow failure (1 success short): The character came close. This is where success with cost is most appropriate, the GM may offer the character what they were attempting, but something is spent, damaged, or compromised alongside it. The lock opened, but the pick broke inside. The NPC agreed, but overheard the wrong thing on the way out. Not every narrow failure warrants this treatment, but it is always available as an option.

Clear failure (2 or more short): The action failed and it was apparent. Narrate the failure directly, don't soften a clear failure into a near-miss. But the scene should still move. What did the character observe in failing? What condition changed? A failed roll that produces nothing, no new information, no shifted circumstance, no consequence, was a wasted roll. Something should be different after the dice land than before.

The forward principle: A scene should be in a different state after a roll than it was before, regardless of result. Success moves the fiction in the direction the character intended. Failure moves it somewhere else. Both are valid; stasis is not.

The Wild Die as Narrative Signal

The Wild Die produces four cases that warrant distinct narration beyond the arithmetic.

Explosion on success: The character met the threshold and the Wild Die rolled 6, counting as 2 successes. Narrate the success as something exceptional — not just competent, but a moment of genuine excellence. The margin the Explosion produced is real; let the fiction reflect it.

Explosion on failure: The Wild Die contributed 2 successes to the total, but the character still didn't meet the threshold. This is a dramatically close attempt. No mechanical effect beyond the failure itself, but the fiction deserves acknowledgment: the character tried harder than anyone expected and still came up short. That is a different story than a routine miss, and the narration should reflect it.

Implosion on success: The Wild Die removed a success, but enough remained to meet the threshold. The action succeeded, and the Implosion tells the GM to narrate a success with a visible hitch. The door opened, but the noise it made was more than expected. The grapple landed, but the footing shifted awkwardly. This is not a mechanical penalty; the success stands. The attached complication is real but separate from the roll's outcome.

Implosion on failure: The character failed and the Wild Die rolled 1, removing a success from the total. The failure was unambiguous and something was spent in the attempt — a wasted resource, an exposed position, an opening the enemy can read. Narrate what specifically went wrong. An imploded failure should feel distinct from a routine miss.

In all four cases, the GM's narration follows the mechanical result without overriding it. An Explosion does not turn a failure into a success. An Implosion does not turn a success into a failure. The Wild Die adds texture to whatever the pool produced.

Critical Triggers (Optional — see AX.C.02.05)

Groups using the Critical Triggers optional rule add structured mechanical effects to two of the four Wild Die cases.

Critical Success (Explosion on success): The character met the threshold and the Wild Die shows 6. In combat, select one effect from the Critical Success options in AX.C.02.05. In non-combat Challenges, the outcome is enhanced beyond the standard success. Narrate the result as exceptional — the character isn't just succeeding, they are building momentum.

Critical Failure (Implosion on failure): The character failed and the Wild Die shows 1. In combat, the GM selects one complication from the Critical Failure options in AX.C.02.05. In non-combat Challenges, a secondary consequence triggers alongside the failure. A Critical Failure should feel distinct from a standard failed roll — the character is not just back where they started, they are in a worse position.

Scene Framing and Rest Economy

Scene framing is the GM's primary pacing tool. A scene should end at or immediately after its peak moment, the revelation, the decision, the resolution of tension. Scenes allowed to wind down past their peak lose energy and make sessions feel slow. When the most interesting thing in a scene has happened, cut to the next one.

Entering scenes works the same way. Begin with the characters already in the situation, at the negotiation, inside the building, facing the obstacle. Travel, preparation, and approach rarely generate meaningful play on their own; stakes do. If nothing interesting can happen during the approach, skip it.

Rest economy is the GM's control over recovery pacing. The rules define what each rest type recovers; the GM determines whether conditions permit one.

A Short Rest (10–15 minutes) requires no active threat. Deny it when threats are ongoing, time is critical, or the characters genuinely cannot stop safely. Granting Short Rests freely reduces attrition between encounters; denying them entirely turns every session into a war of resource exhaustion. Both extremes flatten the texture of play.

A Long Rest (8 hours) requires shelter, food, and water. Use environmental and narrative pressure, not arbitrary refusal, to pace when Long Rests are available. A campaign that provides a Long Rest between every encounter removes the resource management that gives individual encounters weight. A campaign that never allows one turns survival into a grind rather than a challenge.

Rest is a pacing lever. Use it deliberately in both directions.

Resolving Edge Cases and Rule Gaps

Every session produces situations the rules don't explicitly cover. When they arise, work through the following in order:

Intent first: What was this rule trying to accomplish? Apply it in the spirit of that purpose. Most edge cases have a clear answer when the rule's function is considered rather than its literal text.

Precedent second: Has this situation come up before in this campaign? Apply the same ruling. Inconsistent rulings on the same situation erode the table's trust in the adjudication process. Consistency matters more than perfection.

Dramatic logic third: When intent and precedent don't resolve it, ask what outcome makes the most interesting result without undermining the players' investment in their characters or the fiction they've built.

When a ruling is made at the table, note it. Apply it the same way if the situation recurs. If a ruling proves unworkable, revise it between sessions, not mid-session, and tell the players it changed and why.

Genre Catalog mechanics are local rules. When a Genre Catalog defines a mechanic with no core analog, that mechanic governs its domain and does not imply changes to core rules outside it. A digital-space encounter system in a sci-fi catalog doesn't affect how physical combat resolves. Treat catalog mechanics as authoritative within their stated scope and silent outside of it.

NPC Framework

AX.C.14.03

A non-player character is anyone in the world the players interact with who isn't trying to kill them, at least not yet. This section provides a minimal construction framework for creating and running NPCs without full stat blocks, and a tracking approach for making recurring characters feel like part of a living world.

Quick NPC Construction

A playable NPC needs three things: what they want, what stands between them and that want, and how they respond to the players.

That's it. A name, a physical description, and a backstory are useful when they serve the interaction, but they are not what makes an NPC work at the table. The three-element minimum is sufficient to run any NPC through any kind of scene: social, investigative, or transactional.

Build the three elements before the session. Fill in color during it, or between sessions as the character becomes established.

Want, Obstacle, Response

Want is the NPC's motivation, what they are pursuing independently of the player characters. It should be specific enough to drive behavior. "Power" is too abstract to be useful at the table. "Become head of the guild before their rival does" generates behavior the GM can play consistently across multiple scenes.

Want operates independently of the players. An NPC with a want is doing something between sessions. When the players encounter them, they are catching the NPC mid-process, not waiting for the players to give them a purpose.

Obstacle is what stands between the NPC and their want. This is the source of tension that makes the NPC a character rather than a fixture. An NPC who already has what they want has no reason to engage with the players except as set dressing. The obstacle is the opening through which the players can become relevant to the NPC, and through which the NPC becomes interesting to the players.

The obstacle doesn't have to be the players themselves. A merchant wants to expand their operation; the obstacle is a rival who controls the only available routes. The players can help, complicate, ignore, or exploit that situation, but the obstacle exists regardless of their choices.

Response defines how the NPC reacts when the players enter their situation. Five common types:

• Cooperative: The NPC is openly helpful. They share what they know, facilitate what the players need, and ask for little or nothing in return. Useful for established allies and incidental contacts, but too many cooperative NPCs flatten the social texture of a setting.
• Transactional: The NPC helps, but at a cost. They are not hostile; they simply expect something in return. The exchange defines the relationship and creates a web of ongoing obligations.
• Hostile: The NPC is actively opposed to the players' goals, presence, or faction. Hostility doesn't require personal animus; it can be professional, ideological, or circumstantial.
• Indifferent: The NPC doesn't care. They have their own problems and the players are not among them. They can be made to engage, but it requires the players to make themselves relevant to the NPC's want or obstacle first.
• Layered: The NPC presents one response and holds another underneath. The cooperative contact who is feeding information to an opposing faction. The hostile official who privately wants the players to succeed. Layered NPCs require the players to work past the surface response before the real shape of the relationship emerges.

Differentiating NPCs at the Table

The goal is memorability, not performance. A GM running six distinct NPCs across a session cannot rely on extended character work for each one. Three anchors make an NPC recognizable without preparation:

One speech pattern or verbal tic: A character who asks a question at the end of every statement. A character who never refers to themselves in the first person. A character who uses professional jargon from their field even in casual conversation. The pattern doesn't need to be elaborate; it needs to be consistent enough that players recognize the character when the GM slips back into it.

One physical detail the GM returns to: Not a full description, one specific detail that gets mentioned when the NPC appears. The clerk who is always eating something. The official whose uniform is never quite right. The contact who doesn't look at whoever they're talking to. Physical details anchor an NPC in the fiction when the GM brings them back after a long absence.

One piece of information only they hold: Every meaningful NPC knows something the players don't. This creates the information asymmetry that makes social scenes function as actual play rather than exposition delivery. An NPC with exclusive knowledge has leverage, and so do the players, if they realize it. The question of how that information is exchanged, withheld, or extracted is what the scene is actually about.

Information asymmetry applies even when the players don't know it exists. The GM knowing that an NPC holds a particular piece of knowledge changes how that NPC is played, regardless of whether players think to ask.

Tracking Recurring NPCs

An NPC who appears more than once needs three tracked fields maintained between sessions:

Current want: What are they pursuing right now? This may evolve as events progress. An NPC whose obstacle was removed will shift to a new want. An NPC who suffered a setback may have revised their approach or redirected their efforts.

Player history: What have the players done that this NPC knows about and cares about? Not a complete log, only the events the NPC would have observed, heard about, or been directly affected by. This is the input to disposition.

Current disposition: How does this NPC currently regard the players? Disposition runs from actively helpful to actively hostile, and it should shift in response to player history. An ally the players ignored at a critical moment is a cooler ally next time. A neutral contact who received unexpected help without being asked has reason to reciprocate. A hostile NPC who was treated with respect despite the conflict may carry residual complexity into future encounters.

The world should remember what the players have done even when the players don't. When a recurring NPC reacts to events the players have half-forgotten, it signals that choices have weight, which is the foundation of meaningful play.

Tracking doesn't require a dedicated system. A single line per recurring NPC, want, one relevant event, current disposition, is sufficient. Update it after any session where the NPC appeared or where something happened that would affect them.

When an NPC Becomes a Threat

An NPC crosses into AX.C.14 territory when direct opposition requires mechanical resolution, when the conflict between them and the players becomes a matter of dice rather than conversation. This may happen mid-scene or arrive as the destination of a slow-building antagonism. In either case, have a stat block ready before the scene reaches that point. Building one during a confrontation wastes time and produces unfocused results.

Three things carry over from the NPC framework into the stat block:

Motivation drives tactical behavior. An NPC who wants to capture the players alive fights differently than one who wants to eliminate them. An NPC whose obstacle is a time limit pushes aggressively rather than plays defensively. The stat block defines what the NPC can do mechanically; want and obstacle determine what they will do and when they will stop.

Behavior maps to the intelligence tiers in AX.C.14.02. An NPC who has been established as canny and strategic belongs at high intelligence (Wit 4D+); they will use cover, coordinate with allies, and retreat when losing. An NPC who has been played as reactive and impulsive belongs lower on the scale and should fight accordingly. The stat block shouldn't contradict the character the players have already encountered.

Established personality carries into the encounter as the NPC's voice. An NPC who has spoken with the players, made threats, or revealed something about themselves should continue doing so during the confrontation. An NPC who goes silent the moment combat begins is harder for players to engage with, and easier to forget afterward.

Assign a power package based on the NPC's established role and the party's current level. See AX.C.14.02 for scaling guidance and AX.C.14.03 for the complete stat block template.