We're hoping this column becomes your window into roleplaying design and development -- or at least the way we approach these things here at Wizards of the Coast. We'll handle a wide range of topics in weeks to come, from frank discussions about over- or underpowered material, to the design goals of a certain supplement, to what we think are the next big ideas for the Dungeons & Dragons game. All of this comes bundled with a healthy look at the people and events that are roleplaying R&D.
This week, Dave Noonan continues his examination of adventure design concerning his own future adventure(s).
Wherein Dave explains why you should say “linearity” like it’s a good thing.
Between our last look at adventures and this one, I’ve finished my design work on one adventure and started another. As promised, I’m going to work through my design notes, applying concealment as necessary so that if you someday play the adventure, it won’t be spoiled for you.
There’s Gnoll Love Like Our Love
Last week I picked on the drow, who often have a hard time justifying their challenge rating. In my adventure, I also added class levels to gnolls, and found it was a far more satisfying exercise.
Here’s why the gnoll makes a good chassis for adding on class levels:
- Utter simplicity in terms of racial features. Adding class levels to a monster sends its complexity through the roof, so it’s good if the base monster is… a little boring. There’s a pretty obvious reason why character classes are the most complex thing in the game (hint: look around your D&D table sometime). But if you add class levels to a creature with a lot of special abilities, spell-likes, and wacko monster stuff, you can pretty quickly create a stat block that’s a train wreck of powers. The DM has enough to keep track of, and an insanely complicated monster isn’t doing him any favors. When you’re looking through a 600-word stat block for one good action to take, you aren’t playing D&D anymore. You’re engaged in a word search.
Tangent Alert!: There’s another practical reason why a complex monster doesn’t need class levels tacked onto it. Most monster encounters look like this: Monster jumps PCs/PCs jump monster, then fight, fight, fight, fight, and the monster’s dead. (I know there are exceptions, but bear with me.) Most D&D monsters have a lifespan of five rounds or less. That begs the question: How many good “moves” should a monster have? Five? Seven, so you have some versatility? Ten, so you have lots of versatility, but you know that half of them won’t get used?
|3.0 marilith: At will—animate dead, bestow curse, chaos hammer, cloudkill, comprehend languages, darkness, desecrate, detect good, detect law, detect magic, inflict serious wounds, magic circle against good (self only), magic weapon, project image, polymorph self, pyrotechnics, see invisibility, shatter, telekinesis, teleport without error (self plus 50 pounds of objects only), unholy aura, and unholy blight.
3.5 marilith: At will—align weapon, blade barrier (DC 23), magic weapon, project image (DC 23), see invisibility, telekinesis (DC 22), greater teleport (self plus 50 pounds of objects only), unholy aura (DC 25). The marilith's polymorph was removed, as noted in the most recent errata.
If you’ve got a circa-2000 Monster Manual (that’s a 3.0 one), take a look at the marilith on page 46. I’m counting 22 spell-like abilities, including animate dead (in case you wanted to pair up a CR 17 marilith with a bunch of skeletons and zombies) and bestow curse (does the marilith think that its 17th-level opposition won’t make a DC 16 Will save?). More to the point, even the better spell-likes on that list take away from the fundamentally cool thing you want the marilith to be doing: slaughtering PCs with all those swords. There are a handful of good abilities on that list, but they’re like the fruits hidden among the “leaves” of all those low-level spell-likes.
Fast forward to the current incarnation of the marilith (p. 44 in your Monster Manual). It’s a much more reasonable eight spell-like abilities (previously nine, before the most recent polymorph errata). This is something we learned along the way. You probably won’t see monsters with long lists of spell-like abilities anymore. We’re pruning them back so you can find a few good ones to use during that marilith’s five-round lifespan.
Besides, don’t you want the marilith saying, “Undead minions? Yeah, I have people for that.”
- Ability score mods that I can work with. You’re looking at Strength +4, Constitution +2, Intelligence –2, Charisma –2. OK, so gnoll wizards, sorcerers, and bards are challenged. But every other class looks good for a gnoll. The gnoll cleric and gnoll rogue would miss having a decent Charisma or Intelligence score if they were PCs, but they’ve got a life expectancy of five rounds or so, so they work just fine.
- Those two humanoid levels. Yeah, you’d think they would hurt spellcaster gnolls, but they really don’t. For example, I made some gnoll clerics. If you add two cleric levels to a CR 1 gnoll, you only increase the CR by 1, because cleric levels are nonassociated (see p. 294 of the Monster Manual). So now I’ve got a 4 Hit Dice creature that’s only CR 2. And getting access to 1st-level cleric spells gives that gnoll a lot of versatility. Cure light wounds, divine favor, and shield of faith? All great spells that help this NPC survive long enough to make an impact.
- Finally, there’s this tiny little reason: A gnoll’s favored class is ranger, and NPC rangers have good treasure. They typically wear cool light armor (useful to a good chunk of the party), and they’ve either got a good longbow or two good melee weapons. Many fighters and other nonspellcasters spend their money on melee weapons rather than ranged weapons, so when a potent ranged weapon falls into their lap, they’re awfully happy about it. And dual-weapon wielders make for an easier treasure split: The paladin can take the longsword while the rogue gets the short sword, for example. If that were an NPC fighter with a greatsword, that split can’t happen. Like I said, it’s a tiny reason, but it’s one of the things we think about when writing an adventure: making sure a good chunk of the treasure is useful to a typical group.
So the gnolls were great baseline monsters for me, but they weren’t perfect. There’s one little problem: They speak only Gnoll, and with that –2 Intelligence penalty, it would take a substantial chunk of their available skill points to teach them Common. My gnolls live and work with some giants… giants who also speak only their own language, and who also have an Intelligence penalty. It took some engineering on my part to at least get the leadership to be able to talk to each other.
But there’s still a little problem: I think that it’s better for rank-and-file humanoid bad guys to converse in Common. The game where the PCs are eavesdropping before the fight, trading taunts during the fight, and listening to “Run! The humans are coming!” after the fight is better that the game where it’s all “guttural growls in a language you don’t understand.” Language barriers have their place, but D&D is fundamentally a game of verbal communication. Language barriers should be the exception, not the rule.
We often say that there isn’t much difference between site-based adventures and event-based adventures. It doesn’t take great mental gymnastics to imagine that the dungeon map is actually an event flowchart, or that the flowchart is really a map. But there’s a key difference between event flowcharts and maps: The lines that connect the boxes on the flowchart are usually one-way arrows; cause leads to effect, but then you generally don’t go back to “cause.” But with a dungeon, the corridors that connect the boxes run both ways. The players perceive greater freedom of choice on a map, even if revisiting a room where you’ve already been won’t be exciting most of the time.
But whether you’re drawing a map or a flowchart, there’s a fine balancing act when it comes to junctions: How many to provide? The word “linear” is not one an adventure wants to hear, but I think it’s unfairly maligned. I know from bitter experience that a dungeon with too many choices (not linear enough) is just as unsatisfying as one that’s basically a chain of rooms (too linear).
One of the reasons that linearity is good is the rapid pace of level advancement in D&D. Let’s start with a pretty basic assumption: 12 encounters gets you a level.
|Tangent Alert!: I think the real number might be more like 10 encounters nowadays, for reasons on both sides of the screen. DMs are increasingly likely to throw monsters one or two points of CR higher than the average level of the party. And players are often playing with ability scores that far outstrip the 25 point buy that is the game’s intended baseline. Which happened first is a chicken-and-egg question, I suppose.
If your dungeon has more than 12 rooms, your characters are going to level up. Make a 25-room dungeon, and they’ll level up twice. Particularly in a low-level dungeon, you need some linearity to ensure that players don’t hit the CR 3 or CR 4 monster in room 25 until they’ve got the experience from rooms 1 through 24.
There’s an “analysis paralysis” reason why linearity is a virtue, too. If every room has four undifferentiated doors leading out of it, you’re going to see the game grind to halt as the players argue every time about whether to go east or west. That’s no fun for anyone.
So clearly you want some linearity, but players will feel stifled if they don’t feel like they get to make meaningful choices. Here’s one approach to linearity that worked well for me. I’m going to use a dungeon as the example, but the approach works for any adventure site—or any event flowchart, for that matter.
Rather than start the PCs at one edge of your graph paper, put that first entry staircase in the middle of the map. Drop them into a room that gives three or four choices right off the bat, and your players will revel in the choice. Then, build your dungeon like a bullseye, with easier encounters near the middle of the map and the tough stuff tucked away at the edges and corners. Include periodic branches, especially ones that connect within the same “ring” of the bullseye. Now the players perceive meaningful choice, and you know the PCs won’t get to the corners without the prior experience they need.
Next Time: A Preview of Sorts
One of the most exciting things about this adventure is that it has a new approach to giving the DM the necessary information to run each encounter. I’m going to show off some prototypes we used at last year’s Gen Con Indy so you can see how we fold, spindle and mutilate an idea until it becomes a product.
You Craft the Creature
As mentioned, we’ll be asking for your input in the next You Craft the Creature question. Look for a message board thread in the days ahead, asking for your short concept best describing the Aberrant Mastermind. We’ll provide the specifics for how your concept should be phrased; as with the special abilities, we’ll scour the thread to select the nominees for you to then vote on.
Until then, start planning your concept!
About the Authors
Design: David Noonan is a designer/developer for Wizards of the Coast. His credits include co-designing Dungeon Master's Guide II, Heroes of Battle, and numerous products for the Eberron campaign setting. He lives in Washington state with his wife, son, and daughter.
Development: Jesse Decker is the development manager for RPGs at Wizards of the Coast. His credits include a two-year stint as editor-in-chief of Dragon magazine; design work on Complete Adventurer, DMG II, and other RPG titles; and development work on numerous D&D products, most of which he can’t talk about yet.