Dungeon Mastering as a Fine Art

[pemerton]

The thread title is taken from Moldvay Basic (p B60). I think Moldvay gives some good advice. Here is one of my favourite bits:

"That's not in the rules!" The players will often surprise the DM by doing the unexpected. . .

One quick way for a DM to decide whether a solution will work is by imagining the situation, and then choosing percentage chances for different possibilities. For example, suppose the DM is running a combat that is taking place on a ledge next to an unexplored chasm. One player suddenly decides that his character has no chance to survive combat. The player announces "My character want to jump into the chasm to excape!" There may be a chance that he will fall to a nearby ledge or land in a pool of water at the bottom of the chasm. The DM thinks about the dungeon for a minute, and remembers that an underground river flows through some of the lower dungeon levels, so there might be a pool below. Even so, the character will fall 60', and a normal fall will do 1d6 hp of damage per 10' fallen. This character has only 7 hp, and seems likely to die even if the water cushions his landing and reduces the damage. However, there should always be a chance to do something nearly impossible. A player should have, at the very least, a saving throw or a state percentage chance of a miraculous occurrence saving the character. The DM answers: "Looking down into the chasm, your character can estimate that he has a 98% chance of dying, no saving throw, if he jumps. If you decide that your character jumps, roll percentage dice. A result of 99 or 00 will mean that your character lives, but any other result will mean that he will die in the attempt. Do you still want to jump?"​

What I like about this is that it recognises the priority of playing the game over the GM's pre-authored backstory. So more important than the question "Where does the underground river really flow?" is the question "Is there something in the established backstory that can be drawn upon or manipulated to help the play of the game?"

That's not to say that it's perfect. Moldvay seems to suggest that the chance of success should be based on the ingame likelihoods - whereas I prefer an approach that uses saving throws or checks against externally-established DCs. These will make miraculous escapes more common in the game than they would be in real life, but that suits what I want out of the game. Gygax explains this in his discussion of saving throws (DMG pp 80-81):

Could a man chained to a rock . . . save himself from the blast of a red dragon's breath? Why not? . . . Imagine that the figure, at the last moment of course, manages to drop beneath the licking flames, or finds a crevice in which to shield his or her body, or succeeds in finding a way to be free of the fetters. Why not?​

Gygax doesn't set the percentage chance of a saving throw based on the "objective" ingame likelihood of a crevice, or of the chains breaking. The chance is set by the saving throw table.

Still, I think what Moldvay says is excellent advice and I think my GMing could have been better if I'd taken it to heart, and applied it, more seriously earlier on in my GMing career.

There are some bits of Moldvay's advice, though, that I once used to follow but now don't. You can see in the passage I already quoted that he frames the situation from the ingame point of view ("your character can estimate that he has a 98% chance of dying, no saving throw, if he jumps") although it gets a bit clunky when the PC apparently can see that he won't get a saving throw - an infelicitous running together of the ingame and the metagame.

Moldvay says a couple of things directly on the point of metagame knowledge (pp B60-61):

"Your character doesn't know that." A player should not allow his or her character to act on information that character has no way of knowing (for example, attacking an NPC because the NPC killed a previous character run by the player, even though the NPC and current character have never met). If te playrs get careless about this the DM should remind them. The DM may, in addition, forbid certain actions to the characters involved. . .

MONSTER HIT POINTS: The DM should never reveal the hit points of the monsters. It is enough to tell the players how a monster reacts after a successful attack.​

I often tell my players how many hit points a monster has left. If an attack leaves it with 1 hp, or low single digit hp, then I like to taunt them. And as well as taunting value, information about hp remaining can also increase the pressure that the players feel. (In 4e you see this when the players succeed in a couple of big attacks, and express the concern that the monster still is not bloodied!)

I think using metagame knowledge more generally can make for a fun game, for similar sorts of reasons: it helps generate emotionally engaging play. For instance, the scene where the new PC meets the NPC who killed the player's former PC will probably be more emotionally charged if the player takes that shared table experience into account in playing his/her new PC in the encounter.

Who else has thoughts on Moldvay, or Gygax, or other GM advice we've been offered over the years?

 

[Nemio]

I will be following this topic closely since I'm a wannabe DM that's going to buy the new D&D edition to play this game with other newcomers for the very first time

Still, I think what Moldvay says is excellent advice and I think my GMing could have been better if I'd taken it to heart, and applied it, more seriously earlier on in my GMing career

Any examples of situations where you think you might have done better?

 

[Olgar Shiverstone]

Yeah, the Moldvay Basic set was chock full of solid gaming advice.

 

[Yora]

My approach is always asking "what would make the better story?" If something is a bit far fetched or goes completely against something I had planned but not revealed to the players yet, I try to imagine if the game would be cooler if I let the players run with it, or shot down the idea.

Which is why I love OSR games so much. I don't need to suspend the rules to do these things, but it's already the default assumption that you decide when the players have to roll and what.

 

[Celebrim]

Well, I generally agree with Moldvay on all of that, and typically only depart from Gygax over the subject of how antagonistic the DM is to be with respect to the PCs and any advice that Gygax gives which comes from the mindset of running nightly games for an ever rotating large cast of players (the former possibly being a subcategory of the later).

The only area in the quoted rules that I disagree with Moldvay over is his ad hoc rules creation of miracles. I generally dislike fudging the rules and doing anything ad hoc at all. It's inconceivable to me that you'd introduce a chasm into play without as a condition of the chasms existence giving some thought to where it goes, it's physical dimensions, and what lies at the bottom. Hence, the DM has put himself in a terrible situation of having to rule on the life and death of a player character with no fiction established. Without a fiction being established, there is no way for the DM to fulfill his obligation to the player to be an unbiased neutral referee. Although Moldvay is attempting to fulfill his obligation in this situation, I'm not a fan of how he does it or how he communicates.

First of all, a fall of 60' into sufficiently deep water is survivable if you enter feet first. Falls into water are lethal generally beginning at around 150', or at a lower distance depending on how you enter. This suggests to me that the damage from falling 60' into water (in BD&D terms) is not 6d6, but say 2d6 or even (as it would be in 3e) 1d6. So the question becomes, what's the chance of falling into deep water? With 7 hit points left, if the PC hits water (well) there is a strong chance of survival (possibly 100%) of at least the immediate hazards. And even if we assume shallow water that reduces the damage from 6d6 to 5d6, that means the chance of survival falling into a shallow pool is about 1 in 1000, and onto rock about 1 in 50000. I see no reason to pick a single number - 2% - as the approximation of this. Either figure out what the player lands on from the map, or randomly generate a result based on the possibilities implied by the map and then apply the rules accordingly.

Moldvay seems to be giving it as 2% based on not knowing whether there is water or not or perhaps even worse knowing that there isn't water but wanting to give the plan some chance of success. However, he deliberately makes this chance of success tiny and then deliberately tells the player his plan is almost hopeless, reveals metagame information that the character could not possibly have, and all but tells the player not to act on his plan. While some of that may be justified in the situation, particularly with a new player, that general approach is basically railroading.

Conversely, if Moldvay decides that there is a high likelihood that a deep slow moving river actually flows through the chasm (as opposed to knowing this when the chasm is established), and assigns a 98% chance of survival, and then communicates this metagame information that the character couldn't possibly have to the panicking player he's all but telling the player to jump, and again deciding for himself what the story should be (presumably because he likes the story of the player jumping into the chasm, or wants to save the player). Again, the DM is failing in his basic duties as a GM. The players aren't really free to choose, and the GM is overruling their actions in a way that means that players aren't responsible for and can't claim their own actions. They win because the GM gave them the win and the story they are creating is the one that is satisfactory to the GM. And in general, I hate pulling numbers out of the air because they 'seem right'. What does a 2% chance of life or death have to do with the rules anyway? For that matter what does a 'saving throw' have to do with this situation? If a saving throw has something to do with this situation, why can't and shouldn't a player receive a saving throw every time that they fall off something? And if they don't receive such a saving throw purely because the DM has decided they don't in this situation need one, how is that not the DM choosing for himself what he wants to happen?

Moreover, quantifying the game situation in numeric terms is a direct violation of the principle behind not telling the player the hit points of a monster. The whole point is for the DM to begin to transcend the game and immerse the players in the situation. If your narration is liberally sprinkled with numbers and game rules references, what you are communicating is that this is a game and should be experienced as a game. This in my opinion is a playing an RPG at a lower level of skill than experiencing the game as a story in an shared imaginative world where the rules exist only to arbitrate consequences for the characters of that world. The skillful DM doesn't give a monsters hit points but instead narrates the reactions of the monster and consequences it has suffered - "bleeding profusely", "grieviously wounded", "panting and puffing", "whines in pain", etc. - because he's seeking to draw the players into the experience of imaginative play and at least partially out of a boardgaming/wargaming mindset that might otherwise dominate play. Things like, "You have a 2% chance to survive" or "The monster has 14 hit points left" work contrary to that goal.

Basically, in my opinion the skillful DM creates the fiction according to the standards he sets for his world, and the rules according to the standards he sets for his game, and then he having done so largely subjugates himself to these things - concealing and laying down his absolute power - for the sake of the player's freedom of action. Although the DM has the power and right to override his own rules and fictions, the wise DM applying the best techniques of game mastering does so only rarely. To act otherwise is to be a tyrant and deprive the players of a chance to truly influence the game.

 

[pemerton]

Any examples of situations where you think you might have done better?

It's hard to give specific examples out of context. But I think, especially when I used to GM Rolemaster, that there have been times when I've taken the ordinary action resolution mechanics too seriously as setting the limits of what is possible in the game. In RM this is ameliorated a little bit because it has open ended rolls (most things are resolved on d100, and a 96-100 allows rolling again and adding - over the years there were some memorable moments of double- and even triple-open-ended rolls which saved PCs from disaster). But sometimes it can have the effect of shutting down the game.

Part of the issue is one of fairness - if, most of the time, you resolve falling damage on d6 per 10' fallen, how come this time this particular PC gets a % chance of surviving? I didn't use to have a good answer to that question, and so usen't to give the % chance. Whereas now I think I have a better grasp on how to handle some of these issues.

An interesting, and very strong, version of the Moldvay approach to "There's always a chance" (that's another heading that appears on p B60) is found in 13th Age: the players can always declare that their PCs retreat and make it out alive, but the GM then has licence to respond by declaring a "campaign defeat" - one important goal that the players have is forfeited, going against them. In Moldvay's example, for instance, the player can always declare that his/her PC survives the drop by landing in the water, but the GM then gets to declare that while he recovers consciousness, the baddies get to do XYZ. (I'd think of it somewhat along the lines of when Aragorn goes over the cliff in the Two Towers movie.)

This is one way of handling the fairness issue.

 

[Celebrim]

An interesting, and very strong, version of the Moldvay approach to "There's always a chance" (that's another heading that appears on p B60) is found in 13th Age: the players can always declare that their PCs retreat and make it out alive, but the GM then has licence to respond by declaring a "campaign defeat" - one important goal that the players have is forfeited, going against them. In Moldvay's example, for instance, the player can always declare that his/her PC survives the drop by landing in the water, but the GM then gets to declare that while he recovers consciousness, the baddies get to do XYZ. (I'd think of it somewhat along the lines of when Aragorn goes over the cliff in the Two Towers movie.)

The big problem I have with this is that it is not unusual, especially in old school play, for the party to have but a single overriding goal - survive the dungeon. Outside of that context many players do not care what happens in the fiction nor in my opinion has the DM any right to demand that they do start caring.

What are you going to do, force the players to behave as if the lives of the orphans are meaningful to them? It's not reasonable to expect that a player is going to be highly motivated to take moral responsibility for the actions of something they themselves see only as a playing piece in a game. This general approach fails in the generic case because it makes assumptions about the motivations and aesthetic interests of the players. In some cases there is an actual defeat I suppose (of a sort). In other cases, players generally don't care what NPCs do to other NPCs, but instead treat it as background color to their story.

The most common alignment I see in play is Chaotic Self-Interest, which is followed closely there after by Neutral Pragmatist. The only campaign defeat that is meaningful in that context is the death/maiming of a PC. Expecting players to view other goals with great concern is I think naïve, and even to the extent that some will, all I think you are doing is punishing players for having goals beyond killing things and taking their stuff and specifically punishing those with story goals more severely than those with less prosaic concerns. If everyone is a thespian with shared dramatic goals, ok, fine. However more likely you've got a thespian, a power gamer, a clown, a casual gamer, and a couple of players with an amalgamation of motives playing together. I also suspect that this would encourage a lot of 'don't throw me in the briar patch' behavior, with stated goals being actually less important to the player than their unstated goals so as to avoid ever actually being thwarted.

Beyond that, this approach once again amounts to railroading, with the GM deciding what story he'd actually prefer and then implementing it regardless of the rules or fiction. What justification other than your preferences as an author do you have for causing something to occur during the 5 minutes that a PC is 'out of it', if 5 minutes taking a bath room break, binding wounds, polishing a sword, or taking 20 to search a chest for traps wouldn't have the same consequence? You're no longer adequately sharing the story in my opinion. The PC's now live in a world which is unreliable, unpredictable, unknowable, and changeable. In such a situation, no choice can truly matter and the only thing that really matters is subtly influencing the master of the world.

Moreover, outside the context of a well established fiction, most 'campaign defeats' are probably rationalizations by the GM. That is to say, unless the GM knew ahead of time exactly what resources were available to a foe, he could not meaningfully make the situation worse. If you are constructing your fiction post hoc, if you have 'no myth', or if your challenge ratings are mostly being dictated by what the rules say a challenge should be like, you are not really going to follow through on 'In Encounter #62, the BBEG will be aided by 20 servants rather than 4, resulting in a high probability of defeat' because Encounter #62 doesn't exist yet in any concrete way. Unless you've established rules that say event M occurs on Wallsday the 24th of Clement, and concretely know how to adjust the BBEG's time table based on each potential victory or defeat, you're just playing with smoke and mirrors. The theoretical set back to the PC's interests is largely existing only in the mind of the GM, and IMO solely for the GM to rationalize his intervention in the situation to himself. The players themselves are used to living in an unknowable world that spawns encounters according to the GM's whim, so they themselves are never going to know 'this is bad' compared to any standard.

Now, there is nothing wrong with having smoke and mirrors and for that matter, it's not always bad to railroad. But in my opinion artful dungeon mastering requires you consciously know what you are doing when you utilize these bags of tricks. Obfuscating from yourself what you are doing via smoke and mirrors strikes me as a rather poor use of illusionism.

As a bit of an aside, Aragorn going over the cliff in the Two Towers movie was terrible story telling and is not the sort of thing you should implement in your games on purpose. It's also in a gaming context illustrative of the problems with the approach I outlined above. It's bad movie making because first it was redundant. Aragorn falling off something and being semi-conscious happens multiple times in that movie alone, each with a slow motion pan to close up, and each with no sense that the hero is really in danger (especially the second or third time it happens).

Secondly, it's bad story telling because it adds nothing to the story. At no point do any of those scenes actually add anything to our understanding of the character or develop the story. Aragorn continually falling down or getting separated from the group isn't symbolic of anything and isn't meant to reference anything. (Unlike for example, foreshadowing Boromir's moral fall might by doing the same thing in the first movie). There is nothing artful here. All they do is pad the story length out meaninglessly. They go by without ever being referenced again. The audience can almost see the mental workings of the writer, "Add some action here. Maybe falling from a high place. That's always exciting." Ironically, all that padded action does is slow the pace of the story down and make it less exciting.

Thirdly, they are truly meaningless. Nothing about the story changes because Aragorn goes over the cliff. The same set of events still happen and still happen in basically the same way that they would have whether Aragorn took a time out or not. No resources are gained or lost because of them. Aragorn sacrifices nothing of meaning to him. The Battle of Helm's Deep still plays out in the same way. In the context of a post about 'campaign defeats', Aragorn going over the cliff is a quintessential example of how meaningless declaring a 'campaign defeat' that doesn't involve Aragorn's death actually is. Likewise, it shows I think that declaring a 'campaign defeat' that doesn't in some way involve Aragorn's eventual futility - the Battle of Helm's Deep is lost, preventing Aragorn from stopping the sacking of Minas Tirith, one of his major campaign goals - and moreover Aragorn's player KNOWING somehow that if he hadn't fallen off that cliff he would have saved Minas Tirith (rather than the fall of Minas Tirith being DM fiat), is itself pretty meaningless.

And all that assumes Aragorn's player deeply cares about anything other than Aragorn surviving the campaign, preferably while retaining Anduril and that sweet chainmail he got from Théoden.

 

[Yora]

The big problem I have with this is that it is not unusual, especially in old school play, for the party to have but a single overriding goal - survive the dungeon. Outside of that context many players do not care what happens in the fiction nor in my opinion has the DM any right to demand that they do start caring.

What are you going to do, force the players to behave as if the lives of the orphans are meaningful to them? It's not reasonable to expect that a player is going to be highly motivated to take moral responsibility for the actions of something they themselves see only as a playing piece in a game. This general approach fails in the generic case because it makes assumptions about the motivations and aesthetic interests of the players.

And that's why I think dungeon crawling is a terrible way to play the game. If you have fun rolling battles, I won't stop anyone. But if you want to have meaningful development and interaction, you simply have to come to an agreement what the actual goals and priorities of the party are. Which really isn't a big deal. All you have to do is tell the players before they make their characters, that this will be a campaign about heroic knights, a band of ruthless mercenaries, a gang of pirates, or whatever, and if everyone is okay with that and willing to play along with this premise.
Almost all the time, it's a game between people who know and like each other, and get together to have fun. Asking the players to go along with a general concept isn't asking much, and they will almost certainly be happy to agree. And if they really don't like the premise, than they can come up with another one that everyone can agree on.
The situation where random strangers get together with random characters for a few hours of dungeon crawling is probably exceptionally rare and there really is no need to be a slave to some of the conventions this special style of playing requires.
Just ask the players what style they want the campaign to be, and if they don't have any preference (as players almost universaly do), just make something up yourself and ask them to play along.

 

[Bawylie]

Yes, you're going to force your players to create characters that belong in the world. That's the game. People don't show up, fully formed & geared to go out of nowhere. Characters, with classes, ages, races, backgrounds, etc., should also have families, friends, rivals, romances, and all that.

Sometimes, players what to use "Chaotic Neutral" or "practical" as an excuse to avoid RP altogether. Whatever you think of alignment, if the players aren't interested in RP, then they're not interested in setting or pursuing goals (for the most part).

So the fine art of DM-ing starts in character creation. It starts with interviewing the player about the character and rejecting characters that don't fit into the world. "He comes from nowhere, has no family, and cares about nothing." That's not a playable character. It has no motivations nor ties to the setting. You have to lead the player into creating those ties. Because role playing a character is one of this game's biggest features and strengths.

As a DM, you create a rich fictional world with adventure and intrigue. Your players must at least come up with a character with motivations, goals, desires, and some kind of history.

A common pitfall is to have a player create a loner badass who lacks reasons to cooperate. But even the classic loner badasses have teams and support systems. Rooster Cogburn, Batman, Wolverine. Even alone Wolf and Cub has Cub (and a moving personal story).

So if this is an art, part of that means drawing out what makes any PC a person and not merely a collection of stats on paper.

(Unless your game is endless kick-in-the-door action, which is fine, but even that tends to have some dimension to it).

 

[Celebrim]

And that's why I think dungeon crawling is a terrible way to play the game.

And once again, we have a definition of what it means to be an artful dungeon master which is situational and limited. Unless your definition of artful dungeon master takes in to account the many styles of player and the many different aesthetic goals that a group of players can bring to the table - narrative, self-expression, discovery, fellowship, challenge, sensation, etc. - then you have a definition that is merely self-serving and self-centered. You are basically saying, "This is what works for me, but I can't vouch for whether anyone else likes it." That's fine, but I'd like to think we could talk about, "Given any group of players, what can the artful dungeon master do to ensure everyone has a good time."

If you have fun rolling battles, I won't stop anyone. But if you want to have meaningful development and interaction, you simply have to come to an agreement what the actual goals and priorities of the party are.

I disagree with the 'have'. You certainly can have meaningful development and interaction by coming to an agreement about what the goals and priorities of the narration are to be, but that's hardly a prerequisite for having meaningful development and interaction. It's just one path to the goal.

Which really isn't a big deal. All you have to do is tell the players before they make their characters, that this will be a campaign about heroic knights, a band of ruthless mercenaries, a gang of pirates, or whatever, and if everyone is okay with that and willing to play along with this premise.

Which actually has nothing to do with the actual goals and priorities of play any more than "We are going to be delvers in an ancient dungeon" as a premise sets the goals and priorities of play. There is no reason to believe that you've necessarily done anything but change the drapes by saying, "We'll be a gang of pirates on the high seas", and every sort of conflict or goal or interaction possible in "We'll be a gang of pirates on the high seas" is possible in, "We'll be delvers into a ruined dwarf city."

Almost all the time, it's a game between people who know and like each other, and get together to have fun.

So? That's hardly the point. Even amongst a group of people who know and like each other, you'll have that player who is a power gamer, that other guy that loves amateur theatrics, that one players that just likes hanging out with friends, that other guy that sees every situation as an opportunity for mad cap hilarity, that one guy who just naturally gravitates to living out heroic fantasy, and that other guy who decompresses by being an amoral ruthless assassins with every imaginable vice. Coming up with a theory that only serves the viewpoint of the guy who is 100% method actor is pretty useless in actual play.

The situation where random strangers get together with random characters for a few hours of dungeon crawling is probably exceptionally rare and there really is no need to be a slave to some of the conventions this special style of playing requires.

I'm not even convinced you know what conventions are needed to run random characters for a few hours of dungeon crawling are. I certainly didn't until I spent a summer running games for a random ever changing cast of players. After that, I opened up the 1e DMG and saw things I'd never understood before. And I certainly agree there is no need to be slave to the conventions that requires, because it implies that you aren't conscious of why you are doing things and have become hidebound. But I'd equally argue that this is true of being slave to generic non-solutions like, "Have them fail forward." or "No myth" or a bunch of other overly simplistic rules people try to promote as 'the answer'.

But that's hardly the point. The point is that a theory of good DMing has to encompass both your home game with players you've been with for 30 years and who all have identical aesthetics of play, and that random group of 12 strangers at a Con or local gaming store. You have to be able to understand what your constraints are, what the goals of your players are, and how you can meet each of those goals without sacrificing (as much as possible) the other considerations. To the extent that at first you can't necessarily know what goals your players have (they might not even know themselves), as a first strategy I think it best to try to meet a mixture of common goals like challenge, narration, discovery, self-expression and so forth. You can then pull the levers as you discover more what your players need, providing of course you even know what the levers are.