12 August 2007

DM Ramblings (Part I)

Okay, so here’s the first in what I hope will eventually become a series I call “DM Ramblings”, the aim of which is to vent my spleen concerning my philosophy on how the Great Game (by which I mean Dungeons & Dragons) ought to be played. As with any unsolicited, unremunerated maunderings, feel free to take this, and the posts that follow, for what they’re worth.


Back in the good old days, when I used to teach military history and classical strategy, I used to use the following, stripped-down, definitions:

BATTLE: a battle is a meeting between two opposing forces. Commanders use men and equipment to win battles. It may result in a victory for one side over the other, or a stalemate.

CAMPAIGN: a campaign is a series of battles fought to achieve an overarching strategic objective. Generals use battles to win campaigns.

WAR: a war is a series of campaigns fought to achieve an overarching political objective. Sovereigns use campaigns to win wars.

From these, the following definitions flow: TACTICS is the use of men and equipment to win battles; OPERATIONAL ART is the use of battles to win campaigns; and STRATEGY is the use of campaigns to win wars.

If we apply these general principles to the Great Game, we can draw a few parallels. An ENCOUNTER, for example, is analogous to a BATTLE; the PCs use their abilities and their equipment to triumph over a given situation that may comprise physical obstacles, enemy creatures, and traps. A DUNGEON, therefore, is analogous to a CAMPAIGN, in that the PCs employ OPERATIONAL ART to emerge victorious from a series of ENCOUNTERS that must be defeated en route an overarching strategic objective. The PCs gradual progress from first level through (if they survive) to epic levels of experience is the WAR, in the sense that their strategic objective – survival and advancement – can only be attained through victory in a succession of DUNGEONS.

(Things are unavoidably confusing here, because in military terms, a exists at the Operational level, whereas in D&D the “campaign” generally refers to the Strategic level).


NOW…the whole purpose of this article is to discuss what happens when you turn the map around and look at it from the DM’s perspective. Here you have to consider two points of view: the GAME, and the META-GAME. Simply put, the Meta-Game is why the DM does what he does; the GAME represents the impact of the DMs decisions within the artificial imaginary construct of the world he has put together for the PCs.

The two are fundamentally different and cannot be confused without threatening the foundations of the game itself. The ontological difference between the Game and the Meta-Game paradigms boils down to nothing more than this: within the Game World, causal logic must remain intact, whereas from a Meta-Game perspective, there is no requirement for causal logic.

Allow me to explain. Imagine you – you yourself – are being bullied in a schoolyard by a trio of thugs. In scenario (A), you’re just about to lose the fight when a teacher walks out of the school, stops the fight, and orders your assailants to the Principal’s office. In scenario (B), a boulder drops out of the sky and crushes your assailants. You look up, notice that the sky is empty, and wonder where the boulder came from.

From a Meta-Game perspective, there is no difference between the two scenarios; a PC (you) got involved in an encounter and were about to lose, but survived. The mechanism isn’t important, because the story flow is uninterrupted – and the mechanism doesn’t have to be explained, because the DM decides what happens.

From a Game perspective, however, there is an enormous difference between the two scenarios. In scenario (A), causal logic remains intact: the appearance of an authority figure wielding that authority in a situation where such an action would be entirely reasonable means that. The question, “why did that happen”, can be answered on the basis of causal logic )(“X happened because of Y”).

In scenario (B), however, causal logic is jeopardized, because there doesn’t seem to be any reason for the boulder to have crushed your opponents. Did a passing airship drop part of its cargo? Is a Stone Giant watching out for you? Is one of those floating castles from the Dragonlance modules falling apart? Is a Cloud Giant redecorating? Or did the DM just suddenly realize that he had to make sure you (which is to say, your PC) didn’t die, and so arranged a little miraculous intervention?

I’m not saying this is bad – after all, divine intervention is a not uncommon part of the Great Game (as are, for that matter, redecorating Cloud Giants). What I’m saying is that if the rock falls out of the sky and lands precisely where it has to in order to save your kiester at the right time, then there had better be a good reason why the thing fell where, and when, it did. And “reasons”, in order to make sense in the Game World (as opposed to the Meta-Game World), require two things: possibility, and probability.

Is a rock from the sky possible? Hell, sure, it’s possible even in the real world. In fact, there aren’t a whole lot of things in the Game World that would be considered “impossible”. So the causal logic of scenario (B) now hinges on probability. What is the likelihood of each of the potential answers to the question I asked above: why did the rock fall here and now?

Passing airship: Well, you didn’t SEE an airship pass by, so this answer seems improbable. (“It was an INVISIBLE airship”, says the DM brightly. The PCs would be more likely to believe him if they’d heard about such things before). Especially when you factor in how hard it would be to hit three bullies standing side-by-side with a dropped rock from high altitude. IMPROBABLE.

Decrepit Floating Castle: Are there such things in this world? Did it fall through a special game-world crossover rift? And did you see one go by? They’d be pretty hard to miss. IMPROBABLE.

Redecorating Cloud Giant: Okay, some of them live in the clouds, which are pretty high up. Even if this is a reasonable answer, the “passing airship” problem exists here as well: what are the chances that Cloud Giant construction waste would drop precisely on top of your enemies at precisely the moment you need them to? IMPROBABLE.

Helpful Stone Giant: NOW we’re getting somewhere. Stone Giants throw rocks, we all know this; and we also know that they are particularly good at it, so the “possibility” box is checked. And the “probability” box is in good shape, too, because if a Stone Giant is lobbing rocks to protect you, then he is doing so deliberately, and would deliberately target people trying to harm you, thereby answering the key question, “why did the rock fall precisely where and when it did.” Now the only question left to answer is why there’s a Stone Giant looking out for you – and this is the kind of in-game question that adventures are made of.

At last, we have a PROBABLE answer to the key Meta-Game question: “Why did that just happen?” If you don’t have an answer to that question, then what you have is a Deus Ex Machina – and that’s no fun for anybody, because when the god comes out of the machine, he can do anything. If the DM is simply going to drop a rock out of a clear sky to wipe out the bad guys, then why should the PCs do anything other than hang back and wait for the rock to appear?

(This, incidentally, is why I never enjoyed Superman comics. What’s so special about a guy achieving victory when he has infinite speed, infinite strength, and is invulnerable to any and all forms of attack? The worst that could happen would be that he’s late for a crime because he was picking up bagels, resulting in some innocent person getting killed – which is great for emotional character development, but that means that innocent dead guy’s whole reason for existence was so Mr. Invulnernable could learn that with great power comes great responsibility. Except Superman never had to learn that because he was always responsible and careful and on time. The most basic plot questions of the entire Superman series boiled down to two: either somebody with the same powers, i.e. another Kryptonian, is going to show up to kick Superman’s ass; or we’re going to find out how the villain manages to score some Kryptonite. Superman is boring because he is a walking Deus Ex Machina.)

The importance of keeping Game and Meta-Game thinking separate, and the fact that one relies on causal logic while the other does not, lies at the heart of a lot of the humour in the “Order Of The Stick” webcomic, I might add. I might ALSO add that, in over 480 editions of the comic, the writer has never – not once – resorted to using a Deus Ex Machina to get the PCs out of a jam (although he came pretty close to violating the probability criterion in this one).


The same Game vs. Meta-Game argument can – indeed, MUST – be applied when designing encounters and dungeons.

For encounters this is not too difficult. Random encounters are, by their very nature, random, and therefore are unlikely to be scrutinized by the PCs for deeper meaning. But they still have to have a reasonable answer for the causal logic question: “why here, and why now?”

Even random wandering monsters have to make sense. If there are no dinosaurs in your game world, why would one pop up as a random encounter? If the PCs are travelling in daylight, what’s the chance they would be attacked by nocturnal hunters like wolves? Would they REALLY be attacked by apes in a plain or by cheetahs in dense jungle? Bottom line – if the PCs run into a group of bandits while they’re on the road, the DM had better have a good explanation as to why the bandits were there at the same time as the PCs, and why they attacked the PCs. I’m not saying that logical answers to these questions are difficult to come up with; I’m just saying that the answers have to be there.

The same rationale applies to dungeon creation, only tenfold. Dungeons are a key – some would say, THE key – part of designing an enjoyable adventure. From a meta-game perspective, a good “dungeon” is nothing more than an encounter flow-chart designed to offer players an interesting and varied series of challenges against which they can test their PCs, and which should be set at a level of difficulty (the “encounter level”) sufficient to provide the PCs with a dire challenge, but not so easy that the affair is a cakewalk. Too easy means too few XP and not enough decent treasure; too difficult means massive casualties and possibly a TPK. No causal logic is necessary at this level, because the answer to the question “why is this dungeon HERE” (and all subsidiary questions) is “because the DM designed it that way”.

But that answer falls flat once you enter the game world, because in the game world, the causal logic criterion reigns supreme. This time, the question is “What are this dungeon and all of these monsters DOING here?”

Let’s recall what a “dungeon” is: it is a series of encounters, each consisting of some combination of physical obstacles, creatures and/or traps, the overcoming of which will allow the PCs to make progress to some sort of strategic objective. The strategic objective may be something as simple as gaining experience (and therefore advancing in levels), or acquiring useful or powerful items, particularly magical items. It may be fulfilling a minor story goal (like rescuing a damsel from a dragon, or defeating a villain threatening a village), the reward for which could be RP or story experience, receipt of a valuable gift, information, or any other form of recompense likely to push the PCs forward towards some predefined goal. Or completing the dungeon may actually result in an important strategic advance, e.g. by defeating a key villain, or finding a key artefact.

But that’s all meta-game thinking. The real concern is providing a reasonable explanation for why all of the traps and monsters are there. Why is this a problem? Well, some dungeons are harder to explain than others. Take some of the old First Edition classics. The “Against the Giants” modules offered some excellent answers to the “why” question. In “Glacial Rift of the Frost Giant Jarl”, for example, the Frost Giants live in enlarged ice caverns within a glacier because winter is their natural environment, they don’t suffer from cold, and the ice caverns needed only minimal work to make a comfortable and easily-defended home. And all of the animals found with them made sense – they were all from frigid environments and all had logical reasons for being there.

Now take one of the true monstrosities of the game – Gary Gygax’s “Tomb of Horrors” (which, incidentally, you can still find on-line, updated for Edition 3.5). All of the complex back-story notwithstanding, there is simply no way to produce an answer for the “who built this thing, and why” question that can be justified in any rational way. T of H is nothing more than an excuse for sadistic DMs to indulge in a guilt-free TPK. If that’s what floats your (or your Players’) boat, then good for you, and knock yourself out! But there’s no way that any rational DM could integrate that abomination (or it’s junior cousin, the egregious White Plume Mountain) into any campaign based on causal logic.

So whenever I’m considering a dungeon, the first question I ask myself is, “Why is this thing here?” If it’s an artificial construction, the next question is, “Who built this thing, and why?” And if the current denizens are different from those who designed and built the place, the third question is, “What historical chain of events resulted in this place being the way it is now, instead of the way it used to be?”

If you can answer those three questions, you’ll be able to provide a rationale for your dungeon that is grounded in causal logic and that will not, therefore, violate the probability-based separation between the Game and Meta-Game worlds that is vital to creating an enjoyable and believable game – and one where you won’t have to rely on Deus Ex Machina explanations for why you didn’t come up with reasonable answers to simple questions ahead of time.