Article Header Image
Under the Hood of the DMG
Design & Development
by James Wyatt

What goes into designing a book like the Dungeon Master's Guide?

It's hard to answer that question without using the phrase "the kitchen sink." The DMG is jam-packed with information, and in many ways it's a distillation of the design of the whole edition. It didn't take anything like its final form until the Player's Handbook and Monster Manual were nearly finished, but all the wisdom and math that underlie the game come to the fore when we explain to the Dungeon Master how to run the game and how to design for the game.

The best way to give some insight into the book's inner workings, I think, is to take a tour through the book chapter by chapter. I'll highlight an element of each chapter that illustrates an important aspect of the design, development, writing, and sometimes editing that went into each part. Feel free to grab your shiny new DMG and walk through it with me.

Chapter 1: How to Be a DM

There's a section in Chapter 1 (pages 8-10) about player motivations. The section has a lot in common with the similar section of the 3rd Edition Dungeon Master's Guide II, and that's because they both owe their existence to some heavy thinking that took place in R&D around 2003. At that time, we pooled our collective experience and some objective research and sketched out the reasons we thought different people played and derived enjoyment from D&D. Everyone has observed the spectrum, from the die-hard roleplayer to the tactical mastermind, and I think just about every group has the guy who's only interested in the game as a social activity. It helped us at the time to think about these psychographics as we made decisions about what to put in our books -- trying to make sure that our products appeal to the Shakespeares as well as the Caesars, as we called some of our psychographics.

Our department wasn't the only place this kind of thinking was happening, of course. Robin Laws, an eminent figure in the game industry and one of the authors of the DMG2, is particularly famous for a little book called Robin's Laws of Good Game Mastering (published in 2002 by Steve Jackson Games). First in his own book and then in DMG2, Robin applied the same psychographic principles to help Dungeon Masters understand the other players at the table and craft a game that appeals to those players.

Building on all that groundwork, Chris Sims and I put together two and a half pages in the first chapter of the new DMG intended to help the DM create a fun game for the players at the table. We don't go on for columns of text about each motivation. We certainly don't try to suggest that some motivations are better or purer than others. Instead, we simply present three short bullet lists for each motivation:

  • The key characteristics of different kinds of players.
  • Concrete ways that DMs can engage and reward these players.
  • Some ways that player motivations, taken to extremes, can become disruptive to the game.

That's not where it ends, though. Several places in the rest of the book return to this groundwork to help make the DM's job easier. Check out Chapter 2 (pages 18, 28, 32-33), Chapter 6 (pages 101, 105, 115), and Chapter 8 (page 145).

Chapter 2: Running the Game

One of the things that became clear in early playtests of 4th Edition was that this game required a different approach to DM-player communication than past editions. Because the game is so strongly exceptions-based, it's easy for the players to get hit with some really nasty surprises. It really comes to the fore in combat, but our thinking about the issue expanded it into a more general principle: It's essential that the players have enough information to make smart decisions.

The clearest example in combat has to do with monsters that have so-called "gotcha!" abilities. For example, a boneclaw has threatening reach, an exception to the general rule that characters and monsters can only make opportunity attacks against adjacent enemies. When a character who's not adjacent to the boneclaw takes an opportunity attack for moving or using a ranged attack, it's all too easy for the player to feel unduly punished. It's part of the DM's job to communicate the danger so that players can make informed decisions about what they do. Some DMs will choose to be totally explicit: "Oh, this guy has threatening reach. Are you sure you want to do that?" Others will be more narrative about it: "This creature's claws are in constant motion, and it seems very aware of you. Your instinct tells you you're not safe from those claws, even at this distance." Either way, players have the information they need.

The principle goes beyond combat, though. Many of us have played in games where we spent hours searching for the fun, because the structure of the adventure or the meanness of the DM kept us from finding the information we needed to keep things moving. Nobody has fun as those hours stretch on, so we formulated the "Information Imperative" (page 26): Give the players the information they need to keep the adventure going.

Once again, it's a principle we tried to carry through the rest of the book: in narration (page 22), skill challenges (page 75), and adventure design (page 101).

Here's the playtest feedback that directly led to this section of the DMG:

From: "Baker, Rich"
Date: Thu, 20 Sep 2007 09:42:52
Subject: Too much "Gotcha-ness" in monsters?

So, one of the things that really puts a sour taste in my mouth is the gotcha-ness of our current monster designs.

"Gotcha-ness" I define as random things monsters do to break rules and punish me for taking what would otherwise seem to be logical and well-reasoned actions in the fight.

We should be really, really careful about Gotcha! abilities. At the very least, every time a monster has a gotcha ability, we need to make sure that the monster's appearance or nature somehow offers obvious cues as to the potential Gotcha so that players feel like they've got a fair chance to see it coming and not blunder into triggering it.

Chapter 3: Combat Encounters

I often don't use a DM screen when I'm running a game session -- it's just a barrier between me and the battle grid, which is where the action is. There are two things, though, that make me look forward to getting my hands on a 4th Edition DM Screen: First is the awesome art, and second is the table that appears on page 42 of the DMG: Difficulty Class and Damage by Level. Having that table easily accessible in front of me might mean that I never again have to open a rulebook in the middle of the game. It's possible I'm overstating the case, but that table does excite me.

Summed up in the ten rows of that table are both the math that makes 4th Edition work as D&D has never quite worked before, and a significant shift in philosophy that's reflected throughout the game.

One of our goals in designing 4th Edition was to extend the "sweet spot" across all 30 levels of play. There's a general sense among 3rd Edition players that the game hits a sweet spot around level 5 and stays good up to level 12 or so. Below level 5, characters are too fragile, and above level 12 they're too complicated. But I contend that another reason for that sweet spot is that, utterly by coincidence, that's the range of levels where a mostly arbitrary system of damage, hit points, and attack and saving throw numbers align to make the game work reasonably well. One of the ways we extended the sweet spot across all 30 levels was by replacing that arbitrary math with a system that's consistent and coherent throughout the whole game.

And some of that math is reflected on this table. We have a pretty good idea what character ability scores look like across 30 levels. Every character uses the same progression of attack and defense bonuses. We have targets for monster attack and defense numbers, based on what we have found is a good hit rate for character and monster attacks. And we've done the same math for character and monster hit points. All that math lets us build a table showing target DCs and damage numbers for improvised challenges.

The shift in philosophy reflected in this table is most evident in the section on terrain in Chapter 4 (pages 67-68). In past editions, we'd describe things like cave slime as if the DC of the Acrobatics check to avoid slipping in it were an objective, scientific measurement of its physical properties. "How slippery is cave slime? It's DC 30 slippery." But setting a fixed number like that limits its usefulness -- cave slime would be too challenging for low-level characters and irrelevant for high-level characters. In 4th Edition, we tell you to set the DC to avoid slipping based on the level of the characters, using the Difficulty Class and Damage by Level table. So when 5th-level characters encounter cave slime, they'll be making a check against DC 22, but 25th-level characters have to make a DC 33 check.

Does that mean that high-level characters encounter Epic Cave Slime that's objectively slipperier than the Heroic Cave Slime they encountered in their early careers? Maybe. It doesn't matter. What matters is that the DM has permission to use terrain that's relevant to the characters, regardless of their level -- and has a table supported by solid math to make sure it's relevant.

Chapter 4: Building Encounters

This chapter gets to the heart of a significant philosophical shift in 4th Edition -- a model of encounter design that emphasizes multiple threats rather than single monsters -- and it took shape years before most of the rest of the book. During the editing stage, though, it got a significant overhaul, illustrating the importance of a good editor (Julia Martin, in this case) in making sure that information is presented in the most useful and helpful way.

When I sent Chapter 4 to Julia, it included six pages on the basics of assembling a group of monsters to make an encounter. First were two pages on "basic encounter building," then two pages of encounter templates (not too different than what's still on pages 58-59), and then two pages on "advanced encounters." Julia's observation was that the so-called advanced encounter-building was easier to explain and easier to understand than what we had presented as basic -- because it boils down to shopping with a budget. With a simple, easy-to-grasp metaphor like that, it became much easier to explain what had seemed like a more complicated method of putting encounters together.

Our old "simple method" started with the basic conceptual model for a 4th Edition encounter: X characters of level Y should face an encounter with X monsters of level Y. That is, actually, the easiest way to build an encounter -- pick the right number of monsters of the right level. But it suddenly gets a lot more complicated if you want to use monsters that aren't the same level as the characters, or you want to use more or fewer monsters, or you want to build an encounter that's more or less challenging than the baseline.

The "shopping" method, explained on pages 56-57, makes all of that a lot easier. It leads off with the idea of encounter level, as distinct from character level, so you can easily construct an encounter that's more or less difficult than standard. It tells you how to set your shopping budget, and then gives you guidelines for spending that budget.

Hopefully, the result of Julia's insight is a system that's easy to understand at a glance and easy to use when preparing for a game.

Chapter 5: Noncombat Encounters

When I was working on this book, I realized a strange little paradox about myself: I claimed that I didn't like puzzles in my D&D games, and yet I love doing puzzles in my free time. I think I got soured on puzzles in D&D when I played in a game a couple of years ago that was just one puzzle after another . . . and I was playing an illiterate halfling barbarian.

I turned out to really enjoy writing the section on puzzles that appears on pages 81-84. I included a "get a clue check" mechanic that DMs can use to appease players whose high-Intelligence characters should be better at puzzles than they are, and later Bill Slavicsek added (at Julia's suggestion, again), a way of treating puzzles as skill challenges. Combined with advice against overusing puzzles, those rules make me feel better about puzzles in my D&D games, and I've started using them more.

Incidentally, the material about creating riddles owes a great deal to an article by Mark Anthony in Dragon 175, and much of the rest of the advice for creating puzzles is drawn from a pair of Dragon articles by Mike Selinker, who were inadvertently omitted from the credits.

There's also a puzzle in the art for this book, in the sample handout on page 24. It relies on the ability to transcribe Dwarven runes, or else great skill at cryptograms. Can you solve it?

Chapter 6: Adventures

The key lesson from Chapter 6 is this: I am not the expert.

The 3rd Edition Dungeonscape book included an extensive -- one might say exhaustive -- discussion of adventure setting, detailing the many kinds of dungeon rooms and features. Our first inclination in putting this section of the DMG together was to draw on that excellent material. Ultimately, though, I decided that it lacked focus, and it was too much information for the first DMG. So I decided to go ask the best DM I know for some advice on crafting a dungeon setting. Besides being my manager, Chris Perkins is a past editor-in-chief of Dungeon, he runs two weekly D&D games renowned for their depth of story and fast-paced thrills, and he has an astonishing ability to come up with cool stuff on the spur of the moment.

Chris's advice is directly responsible for the section on setting "personality" in this chapter (pages 108-109). It's a difficult concept to describe -- Dungeonscape describes it as "theme," and I also use the words "feel" and "flavor" a lot. It's largely about intangibles and minor details, the things that make the Vault of the Drow feel different from the dwarf-built Forge of Fury or the Hall of the Fire Giant King.

So the lesson I took from Chris's desk back to the DMG is that finding the feel of your dungeon is more important than making sure each room has a clear purpose in its original construction. So Setting Personality is the focus of a two-page discussion, while that exhaustive catalog of dungeon rooms appears in a distilled form as the short list on page 110.

Chapter 7: Rewards

Great Moment in 4th Edition Design: In a conference room sometime in September 2007, Mike Mearls finally managed to communicate a dramatic insight to me and Rob Heinsoo. In effect, he said, we should separate treasure rewards from encounters.

A multitude of problems dissolved into nothingness at that moment. The rewards chapter became a cinch to write. Since we were starting from an idea of how much treasure we wanted characters to acquire over the course of a level, we just presented that -- rather than going through contortions to create tables that would randomly generate an approximation of that over the course of a typical array of encounters.

See, in 3rd Edition we know how much treasure we expect characters to acquire over the course of a level -- it's on a table in the DMG, page 54 (version 3.5). But let's say I'm building 13-1/3 encounters for my 5th-level party. Each one will have an average treasure of 1,600 gp -- but the ochre jelly won't have any treasure, the young black dragon will have a randomly-generated treasure worth 4,800 gp on average (triple standard), and the 5th-level NPC bard will have 4,300 gp worth of gear (skewing heavily toward magic items that the characters will sell at half value). I have to put those encounters together and keep track of both how much XP I'm giving out and how much treasure -- or else just do what most DMs do, trust that it's all going to balance out, and end up with characters that are under-equipped (nine times out of ten) for their level.

When I build eight to ten encounters for my 5th-level 4th Edition game, I build the encounters, and then go back over them and dole out the treasure. I start with the list of treasure parcels I'm going to give out, and it's up to me whether to skip over the ochre jelly or put two parcels in the room it occupies. I could put six parcels in the dragon's hoard. The NPC bard can have a single magic item, actually above his level and therefore useful to the PCs, which I'll cross off the list of parcels. And some of the treasure can be outside the dungeon, in quest rewards.

The treasure parcels presented on pages 126-129 are the final development of Mike's essential insight. They're perhaps my favorite aspect of adventure design in the new edition.

Chapter 8: Campaigns

There's a subtle but really important philosophy change reflected in this chapter. Compare these statements from DMGs past and present:

"What lies ahead will require the use of all your skill, put a strain on your imagination, bring your creativity to the fore, test your patience, and exhaust your free time."

"Creating a campaign of your own is the most difficult, but most rewarding, task a DM faces."

"Planning an entire campaign seems a daunting task, but don't worry -- you don't have to plot out every detail right from the start. You can start off with the basics, running a few adventures (whether published or those you design yourself), and later think about larger plotlines you want to explore. You're free to add as much or as little detail as you wish."

I mean no disrespect at all to the many DMs who do devote countless hours to building their campaigns, sacrificing their free time to give their players a rewarding experience. But one of the goals of this DMG was to help more players move behind the DM screen, and telling them up front that being a DM requires loads of work runs counter to that goal. I ran a long-running "campaign" at lunch hours here in the office that was nothing but one random dungeon after another, and I believe that's a perfectly viable campaign model for DMs who don't have free time to spare. Instead of leading off with dire warnings about how much hard work lies ahead, I chose in this chapter to emphasize the simple truth that being a DM rewards as much work as you want to put into it, whether that's an hour of prep time before each week's adventure or several months spent planning out a detailed campaign.

Chapter 9: The World

One of the topics that drew a lot of debate through the course of 4th Edition design and development was the economy of magic items, which is summed up on page 155. There are some folks, both in the department and among the community of players at large, who are dead-set against the idea of magic item shops where characters can stroll in and buy a +5 holy avenger off the shelf, and others who argue just as fervently that such shops are essential to the function of the game.

The fact that magic items appear in the Player's Handbook in this edition of the game reflects the triumph of the viewpoint that these items are fundamentally under player control. Characters need something to spend their gold on, and we want characters to be able to equip themselves with the items they consider necessary and important to their identity, just as they choose their own feats and powers. So that part of the economy was never really up for debate.

What was debated quite a lot was the question of how characters translate gold into items -- and vice versa. The solution we ended up with is sort of a compromise, but I think it's an elegant one. The natural cycle looks like this:

  • Characters find magic items as treasure. Most of these, they keep and use.
  • Sometimes characters get rid of old items they don't need any more, or they decide not to keep an item they find. Usually, they sell these to interested buyers. As an alternative, they can use the Disenchant Magic Item ritual to create residuum.
  • Characters use the Enchant Magic Item ritual to create items they want. Alternatively, they can buy items, with the question of who they buy the item from entirely up to the DM's discretion.

DMs who don't like magic item shops can take the buying and selling out of that cycle without altering the economy significantly. DMs who prefer magic item shops can do the reverse, if they choose.

Chapter 10: The DM's Toolbox

We knew very early on that designing NPCs for 4th Edition had to be easier than it was in 3rd Edition. If we failed in that, our editors would revolt, and our DMs wouldn't be far behind!

In 3rd Edition, NPCs were part of a tangled mess of rules that wreaked havoc with the economy of the game, among other things. DMs would put hours into designing NPCs with a lifespan at the table of about an hour. They weren't as tough as their CR indicated, and in order to get even close they had to carry as much treasure as a dragon of their level. In campaigns that featured a lot of combat against NPC opponents, characters got a lot more treasure than they deserved for the challenge they faced. Worst of all (and I'm speaking as the designer of City of the Spider Queen, here), there was a whole type of monsters that you just couldn't use unless you were willing to do that work.

Now, NPCs are a lot like monsters -- and the Monster Manual includes monsters like drow, hobgoblins, and (gasp!) humans that aren't built as PCs. When you build an NPC, you don't go through all the steps of building a player character, and you don't necessarily use the same rules. The result is a balanced opponent -- and one that needn't be carrying a single magic item, unless you want to give him one (and mark it off your list of treasure parcels!).

This whole chapter, to me, showcases the elegance of the new system and the joy it is to design for. Advancing a monster? Could hardly be easier. Making a troll fighter? Easy as pie -- and you don't even have to change the troll's level. Want a spellcasting dragon? No problem. Applying a template? The design goal for the templates in this chapter was that the changes to the base creature should fit on a sticky note you could put in your Monster Manual as you run the creature from the book.

When I was at D&D Experience last month, a player talked to me about the process of creating NPCs for 4th Edition. "Say," he suggested, "the characters start a fight with the bartender, and I want the bartender to be an 8th-level fighter. How hard is it for me to make up his stats?"

Here's what I'd do -- and what I did, to demonstrate how easy it was. Open the Monster Manual to the Human entry (page 163). Take the human berserker, a level 4 brute. Turn to the DMG (page 174), and advance him to 8th level. He'll get a +4 on attacks and defenses, +2 damage, and (flip to page 184) 40 hit points. We want a fighter, so we'll give him one fighter encounter power (turn to page 80 of the Player's Handbook) -- let's say reckless strike, since that feels very appropriate for our brute. Done.

Took me 2 minutes.

And that's our whirlwind tour of the Dungeon Master's Guide. I left out Chapter 11, ably written by Rich Baker with assistance from Mike Mearls. But it's my hope that, along the way, you've gained some understanding into the design process that brought this book together, the philosophy that underlies it, and the way that the game has evolved with this edition.

About the Author

James Wyatt is the Lead Story Designer for D&D and one of the lead designers of D&D 4th Edition. In over seven years at Wizards of the Coast, he has authored or co-authored award-winning adventures and settings including the Eberron Campaign Setting, City of the Spider Queen, and Oriental Adventures. His more recent works include Expedition to Castle Ravenloft, Cormyr: The Tearing of the Weave, and The Forge of War. His second Eberron novel, Storm Dragon, releases this month.