Encounter design in Dungeons & Dragons


Designing encounters in Dungeons and Dragons used to be easy. First level of the dungeon? That’ll be level one monsters then! Giant rats, kobolds, skeletons, or similar. Level 2? That’ll be green slime and ghouls. What joyous simplicity.

In designing adventures for D&D, I have never had much time for the encounter design rules. The math involved simply drains me of all creativity and makes me feel straitjacketed.

To be fair, the rules themselves are not too bad. Third Edition (3E) uses the concept of ‘challenge ratings’ (CRs), and every monster is assigned a CR. A CR6 monster is considered a standard encounter for four 6th level characters. A table modifies this to take account of multiple monsters or multiple types.

Fourth Edition (4E) takes encounters as the primary building block of the game. It provides a huge amount of advice on building interesting encounters with unusual terrain or other effects, and a mix of monsters occupying different roles (artillery, skirmishers etc). 4E introduces the concept of the Experience Point (XP) budget. A standard encounter for four 6th level characters would have an XP budget of 1000 XP. These could then be spent buying the monsters for the encounter. 4E simply referred to the ‘level’ of the monsters, but this was mechanically similar to CR, and provided a guide to which monsters were appropriate for a party of a particular level. The 4E system had the advantage of greater flexibility than that of 3E. It allowed DMs to spend the budget on minions, elites as well as extra monsters or even traps.

Fifth Edition (5E) uses the same ‘XP budget’ system as 3E, and here a ‘medium’ encounter for four 6th level characters has a XP budget of 2400 XP. Modifiers are provided for multiple critters. It also makes use of the CR system as a supplementary guide as to what monsters are appropriate for particular parties.

First Edition (1E) and Second Edition (2E) did not, so far as I am aware, have mechanical encounter building rules like these. In the old days, the accepted rule of dungeon design was that the level of the dungeon roughly indicated the level of the creatures present. First level meant first level critters, and a small chance of a second or third level critter.

In early editions, monsters had a suggested ‘number appearing’, which was linked to meeting them on the dungeon level appropriate to their level. If they were encountered on deeper dungeon levels, then more would be encountered.

The “number appearing” was also used as a figure for estimating the numbers encountered in the wilderness, which was a feature of the procedurally generated hexcrawling supported by early editions, and had strong links to the war gaming roots of the game (More on that another time.)


So, what’s the problem?

First, as mentioned above, I find all of the 3E onwards systems of encounter design dull and a distraction from the creative exercise of writing an adventure. Actually having to sit with my budget on a bit of paper and trawl through monster books seeing which ones I’m ‘allowed’ to use is not my idea of fun or creativity.

Second, these systems rely on one critical fact: that monster XP awards and CRs are accurate. Given how many variables there can be in relation to a monster and encounter, I am very sceptical about the reliability of these scores, or what they offer beyond simply eyeballing the monster stats.

For example, in 5E, a medium encounter for 4 6th level characters would have a budget of 2400 XP. Compare an encounter between 1 wyvern (CR6, 2300 XP) and 6 harpies (at CR1 and 200 XP each, with a x 2 multiplier for six monsters = 2400 XP). A wyvern is tough, but a party could handle one as long as the healing kept coming. By contrast, 6 harpies are able to attack each individual party member with 6 charm attacks per round. They can do this from an altitude of 300 feet, which is out of range of most spells. You can argue this how you like, but I don’t find the idea that these encounters are of equivalent difficulty plausible. How do we even compare them?

So the budgetary systems are crude abstractions, and therefore I’m arguing that they are not worth the candle. I can see that some people find them helpful, and god knows D&D has a rule for virtually everything if you want it.

So should we go back to the old days? I’m attracted to the simplicity of saying that level one of the dungeon is where the CR1 critters live. But do we even write adventures that way anymore? Not so much.


So how should people balance encounters?

This question of course assumes that encounters have to be balanced. Fundamentally there is a clash of two different philosophies here. In OE, 1E and 2E the guidance on encounter design was more naturalistic (in that the number of monsters appearing were based on chance, or what “happens” to be there), but also simultaneously gamist (in that the level of the dungeon dictated the dangerousness of the creatures living there).

In 3E onwards, the system is purely gamist and teleological, in that the creatures encountered are based on the level of the party that turns up.

My preference is for a mix of the two. I want the game world to make sense, but I know that the point of the game is fun, and overwhelming the party isn’t fun. But striking that balance means following my gut and experience rather than spending a budget.


What Tangled taught me about Non-Player Characters

My favourite scene in the film Tangled is when Flynn takes Rapunzel to a rough tavern (The Snuggly Duckling…ahem) in order scare her into returning to her tower. In the event she wins over the assembled ruffians by her dream of seeing the distant lanterns, and her appeal to their own dreams.

It’s a touching and funny scene, but it also got me thinking. It works because when we see the tavern full of heavily armed warriors we too assume they are one-dimensional thugs. But we quickly learn that each of them has a goal that is important to them, and they suddenly come alive as real people (well, more or less). How can we not see old hooky differently when we learn of his aspirations to be a concert pianist? Or when we learn that Vladimir collects ceramic unicorns?


Of course, I get that this is all played for laughs. But my point is that it doesn’t take much to bring these characters to life. And this is something we can apply to our gaming.

Players interact with lots of Non-Player Characters (NPCs), such as merchants or city guards, and it’s easy to make them faceless. Equally, many opponents can be portrayed simply as a type, such as goblin bandits, or drow warriors.  And they needn’t be. As Tangled shows, in its campy way, it doesn’t take much to give an NPC some extra characterisation. And they should have it. As the songs says “Like everybody else, I’ve got a dream…”.

It just needs the kernel of an idea to light them up, and this seems like something that a random table could help with, so here is one:

Super-quick NPC characterisation table

Die (d20) The NPC’s goal is to…

1 be a performer (e.g. singer, dancer, etc)

2 to practice a craft (e.g. cooking, tailoring, jewellery making, etc)

3 to be an artist (e.g. painter, tattooist, etc)

4 to find love

5 to take up a new profession

6 to explore new lands or make new discoveries

7 to have a more exciting life (or less exciting life)

8 to be rich

9 to be famous or obtain glory

10 to be important or have power

11 to protect something or someone

12 to be the best

13 to find redemption or make amends

14 to be popular

15 to liberate themselves or others

16 to get revenge or justice

17 to find a purpose

18 to serve a being of reverence (god, demon, etc)

19 to protect or restore their honour

20 to complete a collection or find an object

Rolling for fireball

I recently watched the new TV show “Stranger Things”.  It’s about a group of young friends in the 1980s encountering strange happenings, and I highly recommend it. In episode 1 the friends (three boys) are playing D&D in the basement. The culmination of the game involves facing off against Demogorgon, and showcases one of the friends’ decisions to take a risk to help the party by casting fireball against Demogorgon, rather than playing it safe and casting protection (a decision which is significant later on).


To cast fireball the character is depicted as having to roll a 13 or higher (on a d20, presumably). When I saw this my reaction was probably the same as lots of gamers: “What? You don’t have to roll for fireball. The TV writers just don’t get D&D”.

But I kept thinking about that scene, and the more I thought about it, the more I thought the TV writers were right and we are wrong. It’s an exciting moment in the show. Character lives hang in the balance. They are up against terrible odds, and it all rests on one character’s ability to pull off a powerful spell. The dice feel hot in his hand…he rolls.

Unfortunately the reality is nowhere near as interesting. In the real game we simply declare we are casting fireball, and sit back passively while the GM makes one or more saving throws for the enemies caught in the blast. How extraordinarily dull. I think the TV writers knew about the reality and chose to depict something much more interesting. And isn’t it better for the outcomes of the character’s actions to be in the player’s hands?

This goes to fundamental issues about how the D&D game system is built right from 1st Edition. Melee and missile attacks have always been “active” in that the player rolls for the outcome against a target number (armour class), whereas spells have generally been “passive” and the GM rolls against a target number (saving throw) to see what the outcome is.  I can think of various rationales for why the system work this way, but fundamentally the system is not good, and as that episode demonstrates, the dice should be in the hand of the players in determining their character’s actions.

It’s simple enough to change this system by simply having a “magic defence” statistic that works like armour class, and some other RPGs have gone down that route instead (e.g. 13th Age, Dragon Warriors). That represents an advance, I think. But I’ve also noticed a trend in the new wave of games to try to delegate all dice rolling to the players. This requires a system that allows players to not only roll to determine their own actions, but one that allows players to roll to determine the outcome of NPC/monster actions.

Dungeon World delegates all dice rolling to the players. The GM narrates the situation, and the player decides on what “move” his character will make. The player then rolls dice and, if they fail, or succeed partially, the implications of that are decided by the GM (a “GM move”), such as the monster wounding them, or them being put in a bad position.  Likewise, Numenera has players roll to attack, and players rolling to dodge when they are attacked.

Giving players the responsibility of rolling all of the dice has advantages because it enhances the players’ sense of agency in the world, and it frees the GM up to direct the game. This seem to be a genuinely good development in game design that deserves to be emulated. Lucky rolling…and watch out for demogorgon.

Losing control of the narrative

Another trend that I’ve noticed growing in prevalence is the introduction of mechanisms for improvisation and negotiating narrative control with the GM. I realise that there are purely narrativist games out there, but I’m talking about the introduction of narrative exchange systems in games that are otherwise simulationist or gamist.

In FATE, not only can you invoke an aspect for aid, but the GM can also turn an aspect against you in an appropriate situation by using a “compel”. If you accept the compel (a rogue who is Tempted by Shiny Things, might not be able to help trying to lift the noblewoman’s diamond necklace, for example), then you gain 1 fate point. If you refuse the compel, then you have to pay 1 fate point.

Narrative control systems like this crop up more and more often in modern RPGs.

dungeons-and-dragons-master-book-2-wallpaper-001In Trail of Cthulhu each character has to have a core “drive“. Drives act as tools of narrative control. Where a situation arises that would tempt a character with a particular drive then they can either resist that drive (costing them stability points – a bit like sanity points), or they can act in line with that drive (gaining stability points), even if that will cause complications (such as a character whose core drive is Thirst for Knowledge going into the basement for answers). There are soft drivers for situations that are merely tempting, or hard drivers for situations that are irresistible (costing or providing different amounts of stability).

In Numenera, the GM can intervene at any time to create a complication for a character, called a “GM intrusion” (oohh err missus). A GM intrusion is essentially something that could happen and that would add to the story or excitement, even if it makes the character’s life more difficult. For example, a GM might suggest that nearby guards have overhead a character’s attack, or a character is standing in the wrong place when a trap door opens, or that a character is targeted by a thief. The GM decides what might happen and suggests that, and the player decides whether or not to accept it. Its using a narrative judgement by the GM and player in place of what is often determined by dice roll in other games. If the character accepts the complication, he gains 1 XP, and can gift 1 XP to another character. If the character wants to refuse the complication, he must spend 1 XP. GM’s are expected to intervene no more than once or twice per character per session.

All of these are examples of the GM gaining narrative control over the actions of the character (with the player’s acquiescence). But Fate also allows the player to take limited control of the wider narrative. In combat in Fate, a character can concede defeat (if the battle is not going their way), and in doing so they get to narrate what happens to them. They cannot negate the other side’s victory, but they can avoid the worst part of their fate (and they gain fate points for conceding too).

What I like about all of these systems is that they are doing double duty. Not only are they helping you to characterise your protagonist, and to provide the GM with story hooks, but they are also making the game itself more interesting. We all know that stories are made more interesting when things don’t go to plan, but not all players are willing to make sub-optimal choices simply to create a better story. Indeed, some may feel that they are letting the team down if they make things more difficult for them. These game mechanics prevent the player feeling like he is selfishly spoiling the fun of the other players by making the acceptance of complications a central part of there game.

I GM’d a long-running Vampire the Masquerade game, and one of my house rules was that each player had one or more “fate points”. These could be used to save a character’s life (like fate points in the old Warhammer Fantasy RPG), but they could also be used to pull off some heroic feat, or to automatically succeed at something that might otherwise have proved difficult or impossible. I had to agree to allow any fate point to be used, and they could only be used in ways that enhanced or were neutral in respect of the progression of the story. That is, the players could not burn a fate point to turn an otherwise interesting challenge or encounter into something tedious. Fate points were regained by good roleplaying, and in particular by role-playing one’s character in a particularly convincing or effective way.

My use of fate points was pretty successful. Most of the players used them, and understood that these were not part of the game system, so much as tokens allowing them to change the story in unexpected or cool ways.

Is it right for GMs to be able to take control of a character’s actions? When should players have the opportunity to take control of the narrative?

An unusual aspect

I’ve tended to play and GM traditional RPGs, that follow in the vein of D&D, with a traditional GM/PC relationship, classes, skills and dice mechanics.  In the last few years there seems to have been quiet revolution in game design, and it is is one that I think has largely passed me by until now.

The trend I’m focusing on here began in 2003 with the publication the Fate RPG. Fate is intended to be a generic RPG. It includes a number of innovations that embrace a more narrativist style of play.  I never really took it seriously as I was put off by its use of Fudge dice (6-siders with “+” and “-” instead of numbers), which I viewed as a gimmick. Games using the Fate system include the heroic fantasy game Legends of Anglerre, the sci-fi Starblazer Adventures, and the Dresden Files RPG.

The key innovation of Fate is the use of “aspects”.  An aspect is a descriptive phrase made up by the player that expresses something about their character’s skills, personality, and/or history (which may be positive or negative, but is ideally both). Examples of aspects are: “Tempted by Shiny Things”, “Veteran of the Goblin Wars”, “Knight of the Dusk Blades”, “Honour-bound to Avenge my Brother”.

pzo1117-maguscolorIn Fate aspects can be exploited in an appropriate situation by spending 1 fate point, and this allows you to re-roll your dice or gain a +2 bonus when making a skill roll (as the 4dF die roll in Fate normally gives +4 to -4, then +2 is quite a big bonus). Players usually start with 3 fate points.

As an idea, the “aspect” has been quietly ticking away and I’ve been struck by how many modern RPGs use a similar mechanic.

The 13th Age RPG uses a similar idea called “backgrounds”. Instead of spending points and buying skills, players have 8 points to spend on backgrounds which they make up. Each background is a descriptive phrase like an aspect, such as “Cat Burglar”, “Assassin in Training to the Black Wyrm”, “Imperial Mage”, “Wild Mountain Tribe”, “Transformed Animal”, etc.

Trying to do something in 13th Age means rolling 1d20 vs a target number modified by your level and stat bonus plus any points in a relevant background. So each of the background aspects acts like a whole cluster of related skills.

In the Shadows of the Demon Lord RPG, characters have two “professions” instead of skills. A profession is something like “scholar of magic”, “animal trainer”, or “terrorist”(!). Task resolution in SotDL involves rolling d20 against a target number 10 modified by stat bonus. A relevant profession can give a bonus (a boon: an extra d6 to roll and add to your d20 roll), or an automatic success.

Creating a character in the Numenera RPG involves selecting a character class (warrior, wizard or rogue, effectively), a “character descriptor”, and a “focus”. Descriptors and foci are similar to aspects but they must be chosen from a specific list and each provides specific mechanical benefits or powers.

There are 12 descriptors (charming, clever, graceful, intelligent etc). Each of these gives a stat bonus, a limitation, makes the character skilled in specific things, and gives equipment. There are no defined list of skills in Numenera, and the skill training obtained via a descriptor is broad in scope e.g. graceful makes you “skilled in all tasks involving physical performing arts, and all Speed defence tasks”.

There are 30 foci, such as “Bears a Halo of Fire”, “Exists Partially Out of Phase”, “Explores Dark Places”, or “Works in the Back Alleys”. Each of these gives you a special power which improves as you level, equipment and an extra connection to another character.

The Cue system used in the Valiant Universe RPG uses “cues”, and these are effectively aspects that act as inspiration for narrating your character’s actions.

I’m attracted to the idea of aspects because of the potential they have for really allowing you to express the essence of your character. The use of an aspect feels like it should be a big deal, but I find the Fate RPG mechanic disappointing and limited. First because just getting a bonus (albeit, a big one) isn’t that exciting, and second because you can only invoke your aspect by spending a fate point.

Instinctively I much prefer the 13th Age RPG’s idea of aspects standing in for a whole cluster of skills and experience, and being able to use that whenever appropriate. But what is missing here is the Fate RPG concept of your aspect being used against you (more on that later).

But the downside to aspects is that they are broad and both the GM and the players need to be sensible. For example, arguments might arise about whether a player who is a “Warden of the Land” should get an advantage to navigate in city sewers “because they are like underground caves I’ve explored”.

While I think you should be able to use your aspects as much as possible, Fate’s limiting mechanic prevents these kinds of arguments because every use of an aspect requires a fate point, and allowing a broad interpretation is less likely to set a problematic precedent.

How can we take advantage of aspects? What kind of benefit should they give? How should they be limited?

EDITED: to provide some examples of professions in SotDL.