TL;DR:
When you reach for the Resolution Mechanic is far more important than what the mechanic is on its own.
Only call for resolution if results are uncertain, with interesting consequence for both success and failure.
That’s it, that’s the article!
Alternatively:
roll dice less often; say “yes” and “no” more often
//fixme: thesis is that the single best improvement is to discuss the player intention and what is at stake
Intention and resolving uncertainty
Nail down the intention of your game. If the Core Mechanics of the game are not implementing or supporting this intention well, you may be tempted to mess with them. This is a Smell. Consider adapting your techniques, which this article is about. Try not playing DnD, but something else entirely, especially if you’re at the point of “I’ll just reskin it”. Gygax and Arneson didn’t play or design only one RPG system and neither should you limit yourself.
The Core Mechanic of most games is its resolution system, which is used when the results are uncertain. In the lingua franca of D&D5: (d20 + mods >= target number). It has the largest affect on gameplay, and bad resolution systems are often a source of ludonarrative dissonance. Usually this comes up in the context of what the story of the game is about, vs how the mechanics allow you engage with the system and setting. Here, it may be more appropriate to understand this as incongruity.
Incongruity is when the Core Mechanic is disrupting the story, flow, or fantasy that you’re playing to express.
The worst thing an RPG can do with its Core Mechanic -the resolution system- however, is leave it poorly explained. This is one of my main gripes with “&” games, is they are both too verbose for what those systems actually do, without giving enough advice on how to use the resolution system.
Rule 0 of RPGs is that, as GM, you can do what you want with the rules. The important caveat is that you should do what makes the game better. If the intention is heroism, maybe keep the shopping episodes to your group-chat between sessions.
The verbal adjudication
In practice, RPGs operate as a conversation. Any in-game action involves a discussion of players’ intention, the fictional parameters, and the effects. Some players will say “I would like to roll for X”, when what they mean is “I want to X”. Once you suss out the intention, you can make a judgement, and the overwhelming advice is, let them do that thing. However…
There is more to verbal adjudication than just letting players get their muddy boots all over the fiction with constant “yes”. “No”, saith I, “even with a natural 20 it doesn’t make sense to seduce the dragon”. Drawing those lines is important. Limitations breed creativity, and boundaries helps play -the shared fiction- feel real.
The boundaries go both ways, helping you be a better GM for your players. The better you are about being consistent in this respect, the more the players will trust your refereeing, and IME, the more they’ll respect your game together.
I do not believe in:
- The adversarial DMing style typified by fuck-fuck games of
- tricky magic items
- genies
- underhanded GM rugpulling which illegitimatizes nat20s
- Arbitrary fudging which
- was avoidable
- undermines tone
If a certain die result contraindicates what I as referee feel is most dramatically appropriate, I perhaps should not have accepted the roll at all. Forcing oneself to discuss what actually is at stake in a given roll helps avoid this.
Gambling
…Sometimes, players will say “I want to roll for Y”, and they mean, “I want to roll for Y”, because they’re gambling with you that it’ll work, and because the math-rocks make fun clacky noises.
This kind of desire to win against the house is a marker of, perhaps munchkiny play, but more importantly, it’s a marker of engagement. This is rich territory for sharing narrative control. I encourage you to let go of the reins a little. In keeping with the tone, gamble with your players when this comes up.
Here, also, considering what is actually at stake is important.
Friends of “yes” and “no”
“Yes” & “No”, in combinations with “…And” & “…But” are potent!
“Buts” represent between-spaces, other states than pure success and failure. Failing forward, success with consequence, and other bargaining to compel against or introduce aspects of the scene or PC. Discussing the intention, and introducing complications, are super fun. Many players are willing to get their character into trouble to advance the game. Peak roleplaying.
However, complications should add new things to deal with, rather than deleting the success a PC achieved. Imagine marginally small failures and successes, that grey area where fudging is tempting, or where their skill might have succeeded, if not for the difficulty of the scene:
- “yes, you damage the engine, but now you are very obviously covered in oil”.
- “Your blaster misses the bounty hunter, but it hits the door control, and the blast doors shut! There’s no running away from this fight, now.”
Not all uncertainty needs to involve those riders, but they add a lot of flavor. “And"s are particularly useful for critical successes and failures, ripe for opportunities to compel aspects of a PC or a scene:
- “The brutality of your killing blow on the bugbear causes all remaining goblins to test their morale!”
- “No, your knight’s love of wine has caught up with him, and uncouthly offends the baron’s honor in his attempt to sway him”.
Sharing narrative control
//TODO: Expand into an article
I encourage also, in the discussion about “and” & “but”, to be enthusiastic about the ideas your players bring. Having a light hand on the reins of the narrative is hard to teach.
The lingua franca of 5e, and the culture of play around it, suggests that GMs are something aside from a player. Quoting 5e DnD directly, the “[GM] narrates the results of the adventurers’ actions.” Absent any text about player narration, this implies that the GM is solely responsible for imagination. This is wrong. The GM is a player, whose role is to referee.
I frequently discuss the mechanical impact of what occurs, and then ask my players “what does that look like?” and let it ride. Mechanical success or failure, I’ve never been disappointed by letting a player narrate.
Tuning your barometer
Considering what is at stake, mechanically and narratively, is the best way to dial in consistent, fair, and fun refereeing. To that end, I present advice lifted from other games about this subject.
Blades in the Dark
John Harper’s Blades in the Dark introduces Position & Effect for its dice-pool Action Roll. These two dials can be addressed independently to give 9 mechanically distinct outcomes. Position guides the consequences of failure or partial-success, and Effect guides potency of successes. This works in tandem with Clocks, which are a reframed health-track for obstacles; Effect -> “damage” to obstacles, and Position -> stress/harm/complications.
Building the Position and Effect for a roll invokes the Conversation about intentions and parameters. But the Action Roll is not the first stop on the road to success. BitD does a good job of explaining “fiction-first” gaming, but the gist is that the conditions of the fiction, and mechanics other than the Action Roll appropriate to the conditions, should guide how you move forward.
Exercising this procedure will make you a better GM.
“Be a fan of the PCs” and “don’t make the PCs look incompetent”
To summarize, this is largely a framing issue. Whenever consequences occur, even critical failure, an opinion I share with Harper is that being told your PC is a big baby who poo-pood their pants is a mood-killer. It is more satisfactory to talk about the difficulty of the challenge, or the competence of the opponent, etc. Push but do not punish the PCs
Complications are not failures, and should not be used to undermine the PCs, but to launch new threads of fiction. This is also an excellent place to turn it around “why didn’t your character succeed?”, and players are often happy to share the fiction and supply a reasonably interesting jusification.
Please read BitD, it has top-tier GM text.
//TODO: write about clocks and UI
MYZ
Free League’s various MYZ engine games use a similar dice-pool-count-successes mechanic across them, paraphrased here:
Roll (ability + skill + equipment + situational)d6, 6s are successes. If you fail at first, you might roll again by “pushing” at the risk of resources.
A common complaint of MYZ games is that they are very punishing. Starting characters tend to suck at everything, with maybe a 50% chance of succeeding at best.
Their Alien RPG takes this a step further, because as you accumulate failures you also accumulate stress dice, which can send your character into a panic. The advice in the Alien RPG is to really seriously only roll in genuinely tough, dramatic situations. The game text explicitly says that nearly anything which doesn’t fit that criteria should not be rolled and characters should just succeed. It can be difficult, from a background where you roll for everything, to grok that approach.
What qualifies something as genuinely tough? What qualifies as dramatic? You must develop your own answer, but it’s somewhere north of “routine”, in spitting distance of “adventuring conditions”.
GURPS
Steve Jackson Games’ GURPS is tuned to a kind of “heroic realism”. Its 3d6-roll-under system is one of my favorites, admittedly. The brief is that you roll 3d6 under your effective skill to succeed. Because of its results curve inherent to multiple dice, increasing deviation, as well as increasing margin of success (or failure) are mathematically meaningful.
So, even given that rolls are more centrally trending and more predictable than a flat die (1d20 is a flat results “curve”), GURPS still has the advice paraphrased here:
The GM should only require a success roll if there is a chance of meaningful failure or gainful success, in particular:
The GM should require rolls when resources (health/wealth/etc) are at risk, or a PC stands to gain (allies/info/wealth/etc) from the outcome.
The GM should not require rolls for trivial/everyday tasks, daily work at a mundane, nonadventuring job.
Here, the difficulty modifiers imply that skills are pegged to a PC’s capability in stressful situations, where +0 is “Most adventuring tasks, e.g, Driving in a car chase”.
That’s a pretty good barometer, in my opinion.
The remainder of the chapter is a reminder that GURPS 4e Basic Set is a reference tome, with a litany of special cases befitting a Universal game.
Gumshoe/Trail of Cthulhu
Call of Cthulhu is often spooky noir detective work, but has the problem, often, that a player rolling badly can miss a clue.
That sucks and “Bear Lore” is stupid.
Robin Laws’ “Gumshoe” is specifically built to address investigations and mysteries, and Ken Hite’s “Trail of Cthulhu” modifies it to the tentacled milieu. Missing critical clues causing a Halt and Catch Fire, among other problems, are sidestepped by just… giving you the clues with investigative abilities. No random chance involved. Spending points with those abilities gives additional clues, in addition to incurring opportunity costs.
This is my prefered method of providing clues and setting information, at this point. It’s only when my players want to gamble with me, or when information is not critical to adventure, that I gate information behind rolls.
See also “Night’s Black Agents”, also Ken Hite, for vampiric investigations.
Burning Wheel & Others
A rule which I know from Luke Crane’s RPG, Burning Wheel, is called “Let It Ride”. The players get 1 attempt, singular, at a given obstacle or roll. Unless the situation changes dramatically (either in the sense of story-drama or in the sense of a large change), that roll stays in effect. It can be reasonably framed that the attempt given is the best that can be mustered at the time.
OSR, Dungeon Crawls, and Time
Lockpicking without a time pressure is pointless to roll for. What’re you gonna do if they fail? Break a pick? Roll again? meh
Dungeon Turns not only give a game structure, everything that plugs into them makes actions in the dungeon risky, and therefore, something which might warrant rolling.
Turns bring about:
- impending threat of your torch guttering out
- hunger setting in
- gelatinous cubes or other wandering monsters stalking the halls
- time pressures of the adventure of ritual sacrifice or whatever
Turns are an important resource. Time is always marching forward, and the Turn abstraction puts it constantly at risk. Now, failed lockpicking is not just a repeat-until-successful endeavor, it is one that drains resources.
I run most actions with a bent toward competence in this way:
- Engagement with the fiction is best, and if it can, it “happens” in real time. So do conversations.
- Abstracted but reasonably quick abilities roll not to succeed, but to avoid taking a whole Turn to do it. Describing it still happens in real time.
- Abstracted and reasonably involved abilities roll to happen quickly, rather than at all. Searching an office might take only a Turn… or it might take 2d3 Turns.
This is what “taking 10” and “taking 20” rules are for, skipping past the possible failures at the expense of time.
Effects on gameplay
By using good judgement about character strengths, one can help protect and enhance the fantasy in play.
Matt Colville points out in a video, Skill Dogpiling, that failure to protect specialization creates a muppety tone. “[skill dogpiling]’s dumb… it creates dumb outcomes.”
As Matt mentions in his video, integrating that with sensible mechanics and turning it back to the players is probably good practice. Where “no” may meet resistance, “…unless you can justify it or are proficient” may not.
Horror games
Call of Cthulhu or Mothership rolled differently creates different tones. Rolling often, especially with respect to sanity/stress events, can create a spiral of hilarious doom. Rolling too infrequently may pop the suspension of disbelief as uncharacteristically steely characters face inhumane horrors. It can also betray characterization if the hardened FBI agent or the doctor retch at gore, or the android freaks out about interfacing with alien technology.
/**
Aside: that’s a reference to the Mothership 0e rules, in which each character class could gain experience in a unique way, encouraging them to play to the peril of the crew.
Mothership 1e Warden Manual (Tuesday Knight Games) will absolutely make you a better GM for games besides it, btw. To the point about fear/panic rolls, I love the advice that it can be rolled, in addition to the defined list of events, whenever the player wants to. Giving mechanical hooks like that to players to characterize is good design.
*/
Fantasy Heroism/Pulp
Bending towards competence feels better. PCs are adventuring, and are assumed by the game system to be unique to that end. I encourage you to say a simple “yes, you do that” when it is in a PCs wheelhouse and not much is at risk.
PC character sheets are not just a blob of stats, they are a wishlist. BitD encourages thinking of the shared Crew sheet as a wishlist from the players about what types of scores or stories they want to explore. Fulfilling that wishlist by giving them challenges that their PCs are uniquely qualified to succeed in is a cheap hack to player engagement.
Yes, even the minmaxer and the munchkin obsessed with busting the game over their knee. Give them challenges which satisfy their system mastery, those which challenge it, and those which thwart it. If your player wants to build their wizard as a glass cannon who loves fireball, give them things to blow up.
Conclusion
The number 1 technique to better GMing is to call for dice rolls less often. How much less often entirely depends on what is important to your game at your table, and critically, how you communicate that.
Ultimately, this is the space in which RPGs stand out the most from other games. Not only is this core a big differentiator between RPGs, how and when you mechanically adjudicate uncertainty, but this is a unique infinite gap between rules which requires a human to interpret. Board games lack this infinite gap.
For some, perhaps it is best described as an intuition. But a sense for it can be learned, if you’re open to it, even if that learning is rote.