Let's talk about some terminology regarding the conversation before we discuss the core mechanic.
The Conversation
First, the conversation is the back-and-forth that goes on between a player and the GM to make the fiction happen. It's structured as follows:
- Player Intent. The player says what their character is doing. This declaration of intent puts the character in motion.
- GM Confirmation. This is an opportunity for the GM to reiterate the player's intent if it's unclear, to see if the player had something different in mind, or is acting on a lack of information. The GM then decides whether there's a chance of meaningful failure for the character. If there is, he proceeds to #3 and chooses a mechanic. If there isn't, he skips to #5 and yields a resolution.
- Mechanic Selection. The GM chooses the mechanic he thinks best resolves the character's action. Most of the time, this means choosing between the three core mechanics in the game.
- Mechanic Execution. The dice get rolled. Any wonky dice-modifying abilities or fiction-altering currency (such as fate points) get applied, or forever hold your peace.
- GM Resolution. The GM interprets the result of the roll, and says what happens.
Player Intent
When a player says what their character is doing, that intent needs to be phrased in a fiction-first way, meaning the player is describing what their PC does without reference to any mechanics that resolve their action. They say, "I think back to Professor Ruswick's class on arcane menageries, do I know anything about feral pixies?" rather than, "Can I make a Lore check to see if I know anything about feral pixies?" This is called interrogating the fiction.
Interrogating the Fiction
In order to start the conversation (which may lead to your selecting a mechanic to resolve their character's action), players need to interrogate the fiction. This means having their hero interact with something in the fiction, rather than the player preemptively selecting a mechanic in lieu of their character doing something.
Remember, only you as the GM may select a mechanic to resolve a character's action, because you first have to decide in step #2 that the action has a chance of meaningful failure, in order to warrant rolling for it. If the player starts by asking, "Can I make a Perception check to see if there's anyone in this dark room," it assumes there's a chance of meaningful failure that necessitates the roll.
Whenever a player starts the conversation by selecting a mechanic, ask them how their character is doing the thing they're trying to resolve with the mechanic preemptively.
Keeping ThingS Fiction-First
If the player has discovered what seems to be a trap and they ask, "Can I use Reflexes to disarm the trap?" ask them how they're disarming the trap to short circuit their leading with a mechanic.
The trap might involve rearranging runes they discovered earlier in the dungeon into a particular order (a Smart check) or lifting some heavy object off a trigger plate (a Mighty check), in which case the Reflexes skill wouldn't have anything to do with the roll.
Because the roleplaying game is a game, players have a tendency to make use of their skills (especially the ones tied to high attributes), so they will naturally try to steer the resolution toward a particular roll. It's OK for a player to negotiate with you as you select the mechanic, but make sure she's starting the conversation with what her character is actually doing rather than what she, as the player, thinks is the best solution.
On Leading Questions
Don't let players use the conversation to suss out consequences. It's one thing for a player to ask about what his hero might think could happen, given his knowledge of the scene, but it's another thing for a player to ask the GM what might happen in general. For example, a player might ask: "So if I use my lightning maleficence to create an orb of light, will that attract the feral pixie swarm to a corner of the room? Then I can cross the gymnasium while they're distracted."
You can respond to this line of questioning with "I don't know if it will, I just work here!" or "you can certainly try and see what happens." When players ask hypotheticals like this, they aren't participating in the conversation because they're not taking actions as their characters.
If the player had said "How do the pixies react to the gaslamps in the room?" or "I think back to Professor Ruswick's class on arcane menageries, do I know anything about feral pixies?" these are actions the player is taking as their character, that kick off the conversation. In the former case, you could just skip to #5 and give the PC the information that the feral pixies seem to behave like flies to firelight; in the latter case, you might call for a success check and reveal even more weaknesses about feral pixies if the player rolls particularly well (or yield misinformation, if they roll poorly).
GM Confirmation
Step #2 in the conversation is about making sure everyone involved in the action understands what's at stake. Sometimes players say they want to do something that seems outrageous: the action will put them in obvious danger, or the action makes no sense and ignores the context of the situation. Other times, they're unaware of consequences that their heroes would be aware of. Finally, they may just be acting on a fundamental misunderstanding about the scene—the chasm is 100 feet deep, not 10!
This is an opportunity for you to reiterate their intent with the missing context.
Meaningful Failure
When you assess player intent, you determine whether what the PCs are trying to accomplish has a chance of meaningful failure.
Put simply: if you remain silent, things happen. Players make their characters do things by declaring that they do things, and those things happen unless you intervene. This is implicit. If Mister Loan says he walks into the cave and descends the rickety ladder and wades into the dark pool and then reaches for the gem on top of the pillar (assuming you already described that these things are in the fiction), and you decide there is a chance for meaningful failure in his handling of the gem, then all those other things Mister Loan did happen without incident.
To determine whether there's a chance of meaningful failure, ask yourself if what they're trying to do matters, narratively. Would something interesting happen, if they failed?
For example, it may not be interesting if the professional assassin fails to scale the building he's trying to infiltrate to assassinate his target. Perhaps the building isn't particularly tall, or even if it were and he took falling damage, he can just go home and heal himself and then come back later to try fully healed. What might actually be interesting is if he gets caught skulking around, or doesn't kill his mark cleanly and leaves evidence behind, or if he's unable to get a clean shot. So in this scenario we might skip to step #5 when the player says, "I climb up the building to get to the roof" and just let him get there without a check, but pause and make him roll when he says, "Okay, I'm going to sneak across the rooftop to the skylight" because the rooftop has an unseen patrolman standing guard.
Player Intent with Multiple Opportunities for Failure
When you're assessing player intent, the player might sometimes express as a single action an intent to do something that contains multiple opportunities for meaningful failure. Taking the example above, this might be "I want to scale the building and then snipe my mark from the skylight." In this case you want to stop at the first opportunity for meaningful failure, which we determined above is crossing the threshold between the top of the building and the skylight, to avoid detection by the guard.
Mechanic Selection
We already know what rulings are: they are the rules we make up when there's a gap in the rules-as-written. Typically, a ruling will arise at #3 in the conversation, when there's no pre-existing mechanic to reach for. This is when you have to exercise GM fiat to move the narrative forward.
However, when we select an existing mechanic to resolve a character's action, we're also exercising GM fiat in making that choice. This type of GM fiat is the most common you will exercise in OSR+, but the choice is not without guidelines built into the system. When it comes down to it, the most common "tool" you'll be reaching for is one of the three flavors of the core mechanic in the game.
Attribute check
An attribute check yields a pass/fail result. It's either opposed (in which case you need an NPC with attributes to roll against the player) or it's unopposed, in which case you assign a TN. When a PC fails an attribute check, something should be at stake that introduces a complication to push the narrative forward: for example, they fall into the chasm and take falling damage, or the NPC catches them in a lie. Attribute checks should not reinforce the status quo (outside of combat).
Success Check
A success check measures the success of a PC's action. When you choose this mechanic, you're telling the player there is no chance of their hero failing to carry out the intended action, it's just a matter how they performed, and if there were complications.
Scene Check
A scene check is like a success check, except it can resolve an entire scene's worth of actions. Such checks are interested in how the heroes collectively resolved their action (the assumption is that they succeed), and at what cost, in the form of complications. You want to use a scene check to save time out-of-game, or skip past actions that might be tedious to play through, but could yield complications later in the session if they had been played out.
Diegetic vs. Non-Diegetic Mechanics
When choosing between mechanics to resolve character action, consider whether the action is diegetic or non-diegetic. Success checks and attribute checks model actions PCs take in the fiction (they're diegetic): the PC leaps from building to building, so a potential resolution is that the PC falls. However, scene checks model narrative outcomes (they're non-diegetic): the PCs are trying to escape the goons, so the outcomes of that action yield complications in the narrative that may not be directly related to what they're doing (they run into a crowded marketplace which gets them separated, for example).
Order of Operations
If you have to roll against the player (for example, in a contested attribute check, or to otherwise set a TN) there should always be a small pause between when you've informed the player the nature of the roll and before you roll to establish the TN. The easy way to do this is after you've selected and announced the mechanic, prompt the player to declare any tags, then announce that you are going to roll.
- Before you roll. When it comes to negative story tags (such as weakness tags derived from flaws), the player is not entitled to know the TN of the roll before they declare them. If the TN ends up being extremely high, for example, the player may be tempted to come up with a flimsy reason to declare their weakness (since they expect to fail the check anyway). We want to combat this impulse by requiring them to make that declaration before you roll.
- After you roll. When it comes to positive story tags (the kind derived from conflicts that add a bonus to the roll), players are entitled to declare them after they know the TN. This is because the player is dictating what happens in the fiction in a non-diegetic way when they apply story tags, so they need full knowledge of the stakes to be able to make that decision.
- Before they roll. For all story tags (and skills) of any kind, the player must declare them before they roll. This is to prevent players from arbitrarily reaching for bonuses when they haven't rolled high enough, which has the tendency to take them out of the fiction.
Resolution
A resolution (#5) happens when we're past all the fictional negotiating and dice rolling. The player said what his character is doing, you decided there's a chance of meaningful failure, and you choose a mechanic to resolve that action. The dice were rolled. The resolution happens when you narrate what happens. It's the point of no return. Once you've said what happens, it happens, and there's no take backs, even if the player did the math wrong or forgot to apply a skill, etc. Even fate points cannot reverse resolutions.
If you remain inflexible about reversing resolutions, players will quickly learn that they need to get their ducks in order before you start narrating. Keep resolutions sacred.
Rewinding the Fiction
This is going to sound like backpedaling, we promise it's not: it's OK to ask your players if they want to rewind a scene, as long as you have the unanimous and enthusiastic consent of the table.
Sometimes you can sense that your resolution has been disastrous. Maybe it wasn't just one resolution, but a whole series of them that has left a sour taste in your players' mouths. We're not talking about PCs dying, or the party suffering some great loss; those are the sort of losses that everyone at the table savors, even if they're all feeling a little glum. We're talking about some fundamental failing on your part to run the game, whether because you made a bad ruling or because you just didn't stick the landing in a way that felt satisfying.
Riding Rolls
In his article "The Art of Rulings, Part 8: Let it Ride" Justin Alexander discusses the concept of letting the outcome of a check continue forward until the status quo changes:
When a PC is attempting a particular endeavor, make a single check and let the result of that check ride forward.
Justin Alexander, The Alexandrian
The point of a riding roll is to avoid rolling multiple times if the status quo hasn't changed (as this only increases the odds of outright failure). In OSR+, stealth checks are not the only checks you can allow to roll forward; you can do this with any check if it makes sense.
For example, if the character is carrying something heavy and precious, establishing a Mighty check and only rolling a check if things happen in the fiction that would make holding onto it "risky" is a way to "let a roll ride." A failure on this check means the PC must compensate for their loss of control (which can be resolved in a number of ways, depending on what the PC does), similar to when a stealthed character is detected.