The Colors of Magic - Available Now

May 17, 2013

Building Encounters Angry DM Style

The Angry DM has an entertaining and wise post again.  Go read it!  This post on Run a Game discusses his instructions for making awesome encounters.  You can read mine first since a lot of what I'm saying summarizes and comments on his points, but you owe it to yourself go read the original.

You'll also find he has an entertaining and persuasive writing style.  Speak of which, let me quote the thesis of the essay I'm referring to, directly.  He urges DMs to build encounters (whether on the fly or in prep) and run them with the following process:
"Figure Out the Dramatic Question, Sources of Conflict, and Structure; Adjudicate All the Actions, but Watch for the End of the Encounter; then End the Encounter"
Angry gives us this procedure.  Despite the title of the post, the process is more like 6 things than 4.  Let me break it down into a list, then go item by item, briefly:

  1. Figure Out the Dramatic Question
  2. Determine the Sources of Conflict
  3. Add Structure if Needed
  4. Adjudicate Player Actions
  5. Watch for the End of the Encounter
  6. End the Encounter
Figure Out the Dramatic Question

Angry's "dramatic question" is the objective of the scene, written in yes/no question form.  Just having a fight in a dungeon is fine, as long as you know why it's happening.  And once you know why, it may be that the players will decide to resolve the fight differently perhaps with some other means than a fight.

In addition to what Angry says, I would argue that the GM needs to make sure the players know the dramatic question.  Let me cut your obvious objection off with an example:

"Assassins leap from the shadows, gunning the envoy down with incredibly precise blaster fire, then turning on you as he drops to the floor.  They are prepared to fight their way clear, but they also seem prepared to die here."

This is a random attack and a fight to the death.  The players don't know why the scene is happening.  But if you stop and think of the dramatic question, this random combat scene suddenly becomes more interesting...  But to make sure the players have the dramatic question in their mind, it helps to add something to the introduction of the scene to communicate it:

"As he collapses, the envoy gasps, 'Whyyyy?'  Apparently even he didn't know who sent the assassins or why.  Curious..."

The dramatic question is "will the PCs get a lead as to who sent the assassins and why?"  This question constrains the combat, preventing the PCs from using plasma guns, grenades, or other weapons that will destroy evidence.  And they might try to subdue one or more of the assassins to interrogate them later.

Determine the Sources of Conflict

By this Angry means "what does everyone want to achieve in this scene?"  This creates more options for alternate resolution.  After all, if everyone gets what they want, nobody has to die.  The wolves want to protect their territory, so if the heroes can get through avoiding wolf bites by running like hell or sneaking through cautiously, they win.  He hints that social scenes need internal conflict -- that is, the NPCs need multiple motivations, so that one can play off the other.  This also works for combat:  Bandits want to get rich.  This leads them to rob the PCs, which means they have to overcome their resistance with violence.  But the PCs aren't the only people on the road, and they also want to live.  Once it's clear that it's more costly to fight than to run away, the bandits will do just that.  Tracking conflicting motivations can be done intuitively, or the DM can use game system or a kludged system Anrgy invented to track this...

Add Structure if Needed

Some games have morale systems, or systems for determining when combatants are beaten down.  Others lack these systems, and even these systems don't cover all possible situations where NPCs have conflicting motivations.  He suggests, if you feel the need for structure, writing down the NPCs' motivations, and giving them scores from 1 to 10, starting somewhere from 2-9.  The players' action choices can influence the NPCs' motivations.  Set an end condition.  His example is pretty good:

The orcs begin with a ferocity of eight and a fear of two. Each time something happens that makes them want to run away or leave, increase the fear score. Each time something happens to make them more dedicated to killing the party, increase the ferocity score. If the fear ever equals the ferocity, the orcs flee.
That's a kludged on morale system.  Note that 4e D&D's morale system is entirely up to the players -- it takes a Standard Action to Intimidate an opponent, rolling a skill check against his Will defense plus TEN to force him to surrender.  That's a terrible system.  First of all, it has no suggested situational modifiers.  Say the monster is the last orc left out of ten that attacked the party, he's bloodied, and surrounded.  Wouldn't he be thinking about surrender without it even being suggested to him?  3rd edition was even worse.  Successfully intimidating a monster gave it -2 for a round.  That's it.  The old morale systems from 1st and 2nd edition were erased and not replaced.  Pathfinder did not correct this problem.

Something Angry doesn't say, and I feel needs to be said, is that the GM needs to communicate his structure to the players.  The NPCs' motivations don't have to be obvious to the players, of course, but if they have the opportunity to move footballs in a scene, the players won't deliberately try to move them unless they know about them.  Unlike the dramatic question, the GM doesn't have to explicitly communicate this to the players, instead he should make it clear when their actions change the score somehow.  Whenever a player action changes the score, even if you're handling it intuitively instead of with Angry's 1-10 system, let the players know with some narration.  If there's an opportunity to change the score in a big way, hint at that too:

"The orcs attack with a fierce cry, 'Death before dishonor!'  It appears that they would rather die in battle than be seen as cowards in front of their warleader."

So they're going to want to take out the warleader first...

"As the warleader falls, his soldiers stop screaming blustering threats.  They turn their full attention to the thrust-and-parry of the melee, fighting for their lives now, instead of their honor."

So they're expecting that when the orcs start to see their lives threatened, they're going to look for ways out...

"Your blow opens a bloody wound on the orc's arm.  He shifts his grip on the falchion, taking more of its weight on his other arm and steps back to a more defensive posture, looking worried.  A sudden change in the tide of the battle now could spook him."

Now you gave them a clue.  They can do something flashy and scare the orcs off.  Maybe a critical hit, or a "nova turn" or a single potent spell...  Their options are wide open.  They can also decide these orcs don't deserve to live and fight them down to the last, slaughtering them to a man.  But it's up to them.  By employing multiple conflicting motivations, the players' option space has opened up

"After the fireball, half the orcs lay motionless on the floor.  They seem to be glancing around, as if looking for a clear escape route.  Now it's Orc #3's turn.  He shifts back and then runs out the door, leaving just two singed and wounded orcs facing you.  They back out of reach of your weapons cautiously, inching toward the door.  What do you do?"

If the players want to chase down the orcs and slaughter them to a man (orc), you misjudged what the players thought the dramatic question was (it was actually "will the PCs kill all the orcs?" not "will the PCs survive the orc ambush?" or whatever).  That's not bad.  All it means is the players are about to start a new encounter, whose dramatic question is "Will the PCs hunt down all the orcs that ran away?"  A creative DM might come up with interesting consequences for failure there.  Or if you're not feeling like running that scene, you can hand-wave it and let the PCs catch the orcs automatically without running it as an encounter.

Adjudicate Player Actions

Go read Angry DM's post about this.  Remember, you can prep the encounter, but you can't prep the players' action choices.  So go ahead and come up with some structure for likely contingencies, but be prepared to be flexible if the players go outside the box.  This tends to be hard on prep GMs like me.  Prep GMs invest a lot of time into designing encounters to be fun and interesting.  This is why I hate plot killing magic in games.  It's hard to predict when you've got a good dramatic scene, if a player with access to that stuff will just say "I cast X" and shortcut everything.  More on the continuum of player empowerment next week when I discuss what I call The Hunter-Horror Ladder.

Watch for the End of the Encounter

Angry says the encounter is over when one of the following happens:

  • The dramatic question has been answered
  • The conflicts are clearly resolved
  • Most of the players have run out of decision points

The first two are already covered, and you should read the original article for more information on them.  The third is interesting.  Decision points for him are like my definition of risks -- consequential decisions without knowing all the facts.  If everything is known, then the choice is obvious.  If the decision has no real consequence, it's not important.  But decision points can also be narrative decision points:  When the players have control over which way the story goes, that's still in play.

Angry has a great analysis of a scene as a bunch of unknowns (barriers to resolving questions and conflicts).  As the scene progresses, the players take actions to resolve the barriers.  They also use up their own resources.  As their resources are consumed, their options dwindle.  As the number of barriers are reduced, their decision points dwindle.  It is possible to reach a point where the question has not been answered, and the conflicts aren't all resolved, but the majority of players have run out of decision points.  The fighters are surrounding their enemy.  The wizards are spamming the same spell (or are out of the good ones), etc.  In this case the encounter is over because the players' choices have little impact on the outcome, so they're just going through the motions.

Often the players seem to be out of decision points because they're repeating actions over and over (regardless of whether they have other choices).  This is usually a signal that the dramatic question has been answered:  The players are using the same weak attacks over and over because they know they're going to win anyway.

End the Encounter

The Angry DM recommends you use no more than three rounds from the point where you realize the encounter is over and the point where you end it.  It's usually easy to end it sooner.  This is easy with non-combat encounters, because you don't have a rigid order of actions.  Combat encounters are a little more tricky.  In my experience running D&D, the players realize they're going to win an encounter about one or two combat rounds after the DM does.  This gives the GM time to drop strong clues that the enemy is demoralized and ready to flee or surrender, and have them disengage.

In Sum

Go read the original post!  Actually it's part of a series of posts on DMing that seems to be the Angry DM's magnum opus, so to speak.  Go read them all.  They're very good.


  1. My question in response to this is: Does D&D really support answering questions much beyond "Do you (and your opponents) live or die?"

    1. Here you go!

    2. I sure hope it does. Otherwise, I've been doing something unsupported for a really, really long time. I (The Angry DM, the writer of the post that inspired this post) am currently running both Pathfinder and D&D Next right now. Previously, I've run a lot of games, but none so much as D&D. In fact, the reason I am writing these articles is because that D&D is GREAT for this stuff, but it does a really piss-poor job of training DMs and giving them tools. It is great for getting out of the way of the story, but it means all the pressure is on the DM to handle EVERYTHING the story needs except action resolution.