THREAT - The Risk Economy Part III

The last of the set of basic design principles for classic Dungeon Crawls are those around the animate opposition, the active threats to the character, commonly called ‘monsters’. The second of the booklets in the original 1970’s edition of Dungeons & Dragons was titled “Monsters and Treasure”, and began with a list of potential antagonists. Little has changed, the taxonomy of Monsters still has a special place in the game that has increased in importance, size and detail with almost every edition. There’s plenty of reasons for this, monsters are fun and interesting, and a big part of the impetus for play is discovering and confronting wondrous creatures. Yet what exactly monsters do in play is also interesting, and not entirely obvious.

David Trampier - Goblin
From the AD&D Monster Manual

5th Edition Dungeons and Dragons places the monster as an adversary at the center of the adventure - a “Credible Threat” is the first element that the 5th Edition Dungeon Master’s Guide demands for a “great adventure”. “An adventure needs a threat worthy of the heroes's attention. The threat might be a single villain or monster, a villain with lackeys, an assortment of monsters, or an evil organization. Whatever their nature, the antagonists should have goals that the heroes can uncover and thwart.” 5th Edition Dungeon Master’s Guide pg. 71. This isn’t a bad starting point for writing a story - determining an antagonist, but it’s a poor way to encourage exploration of a world or environment. The idea that at the core of an adventure is a relationship between the characters and a foe is a narrative structure that impedes exploration, but worse to the ethics of the classic dungeon crawl, it removes opportunity for player choice by presupposing that relationship. In the dungeon crawl, no matter how fun and exciting monsters are, mechanically they serve as the risk that creates time and supply pressure via random encounters and another form of puzzle (spatial or social) when they appear as obstacles or factions.

ON THE USE OF MONSTERS

An initial point of fracture between contemporary adventure design and classic dungeon based adventure design is classic design's dismissal of the claim that “[f]undamentally, adventures are stories” in favor of the idea that adventures are first locations. Dungeon Master’s Guide, 5th Edition, pg. 71. This isn’t to say that campaigns or Table Top Roleplaying isn’t about collective storytelling or stories, only that adventure design isn’t the mechanism to create that story, but rather to create the setting of the collectively told story.

More specifically, classic play rejects the idea that at the core of adventure design is through designing a series of scenes - what is often, and more clearly for the purpose of the distinction made here, called the ‘encounter’. The current 5th Edition Dungeon Master’s Guide begins its advice on adventure design with story creation and the importance of narrative elements such as “a credible threat” and the importance of an adventure structure including beginning, middle and end. Id. at 72. However, this focus on story quickly moves to the more prosaic design of encounters - “the individual scenes in the larger story of your adventure.” Id. at 81. Encounter design is itself is fundamentally narrative in purpose - the creation of “a straightforward objective as well as some connection to the overarching story of your campaign, building on the encounters that precede it while foreshadowing encounters yet to come.” Id. It is also, for 5th edition, almost exclusively a combat encounter. While traps and social situations are not entirely omitted from 5th edition's design advice, both appear as possible rare alternatives to the expected combat encounter which is favored with additional advice on balancing challenge and adding exciting tactical elements (such as terrain) dominating the Guide’s pages as well of those of other more contemporary editions such as Pathfinder and Dungeons & Dragons 4th edition.

Dungeon crawl design is not narrative - it’s spatial, and so the encounter as a scene doesn’t function well. This isn’t to say that story elements are absent from dungeon design, or from its encounters. Even the set-piece encounter of the scene - an encounter designed to occur in a specific place and often at a specific time - isn’t entirely absent from the dungeon crawl, but the majority of story elements in a dungeon are either discovered as secrets or generated through character choice and the set-piece encounter is reduced to a ‘lair encounter’ that doesn’t itself advance plot or story in a planned way.

With the encounter as the basic unit of adventure an adventure is several encounters where progress is defined by how many of its encounters a party has overcome. I don’t want to entirely stereotype modern adventures as simply a linear string of combat encounters with this description - it’s entirely possible to build an encounter or scene based game that is complex and baroque, but the design principles and ethos are different from a dungeon crawl. When building more contemporary styled adventures one builds and links encounters and scenes, while building a classic dungeon crawl requires the creation of the more holistic level. Classic dungeon crawls thus have their own design principles and the most relevant of these that relate to combat, encounters and monsters are Random Encounters and Faction Play.

An Illustration - Balance and Design, Level or Encounter Based

One way to easily understand the ways encounter and location based design diverge is the way both traditions approach the issue of monster difficulty. Contemporary encounter based design famously focuses on Challenge Rating - a mechanical system for determining how difficult (largely as direct combat) a given encounter will be for characters of a certain level to overcome. “CR” is of greater or lesser complexity depending on the system, but it is always a GM facing mechanism for creating a balanced encounter within a small space - a lair or a scene. It is discreet, the monster a challenge intended to be faced and resolved within the limited space of the encounter.

Location based tradition expands the scope of the encounter to the “Level” where monsters are classified (usually via random encounter tables) by what level of a dungeon they inhabit. This mechanic goes back to the earliest editions of Dungeons & Dragons, with the monster lists by level appearing in the “Underworld & Wilderness Adventures” booklet of the first, mid-1970’s edition. Categorization of monster by ‘level’ is still a balancing method, but less exacting then CR, allowing for only a rough idea of difficulty rather then fine balancing which functions well within the context of the dungeon crawl. These techniques are not absolute, but the distinction between CR and level metrics for difficulty is illustrative of a larger split in design principles. As a practical matter, many designers and Game Masters reject balancing and difficulty metrics, either insisting that they can determine balance without a mechanic, or that concerns for balance are always too gamified, and naturalistic world building precludes balancing efforts. The distinction between CR and level though shows (besides revealing that concerns about balance and difficulty have long been a feature of Dungeons & Dragons) is the frame that the different design principles are using.

Dungeon crawl mechanics are level metered because the location style of play looks at the entire dungeon zone or level as the arena for success and failure in the adventure and interaction with foes. When the locus of play is exploration, fine balancing in the encounter is less useful, and likely impossible - because the risk of encounters will be mitigated or enhanced through means other than combat. The spatial complexities related to exploration can be turned towards victory against enemies. Avoidance, negotiation, trickery, retreat, and stratagem such as false-flight and ambush are all expected, even encouraged solutions in classic play. Of course these player tricks are equally available to monsters and when running a dungeon level, and it’s important for the GM to contemplate how an enemy might use the social and physical space of the dungeon to defend themselves. This makes the difficulty of an individual foe or encounter harder to determine.

Encounter mitigation, or victory through varied and nefarious means is not impossible within the individual set piece encounter, but it can be disappointing because the method of design almost always limits stratagem to tactical consideration by failing to connect each encounter to the larger arena of the level. More modern iterations of Dungeons & Dragons have focused on delivering a tactical combat experience with complex rules designed to showcase the power of the individual character, spotlighting a player’s understanding of the rules and success at building a character potent in combat. There’s a lot of fun in this, and nothing inherently wrong with a tactical combat focus or the encounter based design that allows it to shine, but it’s fundamentally different from the level based design and exploration focus of classic dungeon adventure. Older rule sets largely use rules as a solution to complex risks, or failures of imagination and caution by the players. Classic combat rules are brutal, quick and unpredictable because they exist as an appeal to chance for situations with a large number of random possibilities that the GM can’t be expected to decide on an individual basis and where GM decisions are most likely to be questioned because they result in character death.

What does the distinction between encounter and level design look like in practice, and how are these differences meaningful? Consider how a goblin lair encounter exists in B2 - Keep on the Borderlands is different from a very similar encounter in Lost Mines of Phandelver. B2-Keep on the Borderlands is a widely praised module written by Gary Gygax and published in 1979. It was included in the “Basic” edition of Dungeons & Dragons written by Tom Moldvay and for many classic players, including me, it was the first experience they had with table top role playing games. The adventure consists of a detailed haven (the “Keep”), a small wilderness region, and a dungeon in the form of a box canyon honeycombed with caves. These “Caves of Chaos” are the space where the adventure is likely to happen (though notably the Keep is keyed in exactly the same manner, and rich in plunder) and the goblin lair in the caves is fairly typical of B2’s content.

The B2 goblin lair begins with a discussion of it’s scurrying busy inhabitants and the chance of a random encounter with gangs of wandering goblin warriors. The first areas in the lair are two guard rooms, each occupied by a detachment of goblin guards, and the description mostly relates to how the creatures will raise an alarm by crying “Bree Yark” and induce a mercenary ogre to fight for them if they spot intruders. The text describes how the goblins are militarized and beset by hobgoblin slavers to the South and orcs to the North. The lair exists within the larger context of the level as a whole, and the author’s first impulse is to describe the lair wide defensive measures that the goblins will take.

Lost Mine of Phandelver (“Lost Mines”) is the original starting adventure for 5th edition Dungeons & Dragons, much as Keep on the Borderlands was for 1981 Molday Basic Dungeons & Dragons. It also includes a goblin lair, the first area that the players will encounter. This is itself an example of the difference between play ethos that separates the editions. In “Lost Mines” that characters begin as caravan guards, attacked by goblin bandits, one of whom will flee to the goblin lair. In Keep the goblin lair is but one of many caves in a hillside filled with monstrous humanoid factions. Where Keep makes little assumption about the goals of the characters or how, when and if they will choose to approach the goblin lair, Lost Mines leads the characters to the lair, providing a single clear reason and method for interacting with its residents (a rescue raid).

The lair itself is also very different, for example, while the goblins of Lost Mines have also set up a guard post, the two goblins in it are a discreet encounter, they attack the party rather than alerting their fellows. Additional goblin guards will break a dam against the intruders and they have a pack of wolves that protects one side passage. Notably all of these goblins and wolves, are largely independent. Each encounter exists to be overcome in seclusion without a thought to an alarm or how the lair responds collectively upon realizing it’s under assault. Only after the players overcome the majority of guards and traps does a final gang of goblins think to raise the alarm and summon their leader.

Looking at these different design philosophies, there’s no reason to decide which is better or worse, they intend to accomplish very different goals. Lost Mines wants a string of combat encounters, broken up with a trap, that culminates in a climactic encounter with a goblin leader. It is paced and organized to accomplish this, leading the players and telegraphing the next steps for the characters. Puzzle solving (wolves chained to a wall for example) is limited to specific encounters, and the map serves mainly as an intuitive way to link these encounters. The Keep’s goblins are instead a faction, a monstrous community that the party can encounter, which has needs, dynamics and plans within the larger context of other factions and whose response to player intrusion is not dictated by a player need, but rather the author’s idea of what a lair full of goblins would do. Lost Mines’ goblin lair is narrative, while Keep’s is a location.

With Lost Mines the goblin lair is designed to offer a challenge appropriate to the level 1 characters assailing it, specifically the challenge of combat, which may be made easier by avoiding traps within or engaging in negotiations during one encounter (a goblin sub-chief hungry for power). With the Keep’s lair, level 1 (or 2 or 3) characters are very likely to fail in frontal assault, but the lair’s placement within the level as a whole offers opportunities for risk mitigation through flight, trickery and avoidance, as well as other opportunities for scheming and role play because the party holds a position that can tip the balance of power between various humanoid factions within the Caves of Chaos. The author of Keep is relatively unconcerned with the likelihood of player success, and has instead attempted to naturalistically model a fantastical location and its inhabitants with an internal logic that players can unravel or predict and use as the basis of plans.

The several differences discussed above exemplify alternative approaches to adventure design. With Keep on the Borderland the dungeon as a whole is the locus of play and design depends on creating a location that has a logic or ecology, admittedly not an especially detailed or even entirely believable one, but a set of relationships between the monsters within and the environment that players can understand and turn to their advantage. This principle of design is sometimes called “Gygaxian Naturalism” and it’s typical of classic adventures or more broadly a simulationist approach to design.

Naturalism and simulationism aren’t the goal of classic play’s approach to monsters, though they may be the original impetus (See AD&D’s Dungeon Master’s Guide, pgs. 87- 93). Most importantly, they’re a tool to create a specific relationship between the characters and monsters. When placed with concern to a fictional ecology, either natural or social, monsters can be both unknown and unrestricted in power. Ecology and relationship to the setting mean that smart players, or players engaged in the puzzle aspect of the game, will use rumor and scouting to understand the strengths and weaknesses of the setting around them. Because the monster fits within a comprehensible structure, the players don’t need to rely on the GM to create an artificial one - “CR”. Monsters that are independent of an ecology are much harder to place - the player can’t compare them to their prey or enemies to gauge strength, they have only the GM assurance that the unknown isn’t too deadly or dangerous for the strength of their characters. This doesn’t mean that there’s no balancing, only that the balancing occurs at a higher level then the individual encounter. The strength or weakness of individual monsters can be unknown and unrestricted, because within the context of the level as a whole players can determine it - alpha predators hunt the level, dangerous and best avoided, numerous factions must be met as equals or picked off slowly, while weak factions or outcasts hide and struggle seeking any ally, even invaders like the party.

THE LAIR ENCOUNTER

 Despite differences, encounter based and level based design sometimes use monsters similarly, most commonly in the form of a lair encounter. Lair encounters are monsters that have a fixed location where they remain (usually, nothing here is absolute of course, just general categories with considerable overlap). In encounter based design these encounters, even if not set in the actual home of the creature are most often the basis of specific encounters. In level based design lair encounters represent fixed obstacles or problems, but the ease of avoiding them means the level based adventure usually provides additional reasons for braving the encounter. What’s exciting about lair encounters for the designer is that their fixed nature offers a high likelihood of a specific experience in a specific location - the location can be tailored to the encounter in a way that’s difficult with random or faction encounters. The GM can include environmental hazards and special mechanics as a way of making the encounter more exciting, and because it is pre-planned or known this effort won’t be wasted. In tactically complex systems, where a combat encounter requires a special battlemap and environmental mechanics need to interact with multiple complex subsystems, the sure nature of the lair encounter (or the use of illusionism and plotting to make other sorts of encounters into lair encounters) is almost a necessity.

In level based design though the lair encounter is not especially common, usually only a few will exist on each level because they don’t actively interact with the remainder of the level or its inhabitants in a significant way. A classic lair encounter would be a forgotten and avoided tomb filled with surly undead. The skeletons or wights don’t stalk about the rest of the level, they simply wait, protecting their graves, and other then stumbling upon them by happenstance there’s no significant reason for the players to interfere with their eternal vigil. As part of the level lair encounters can serve the obvious purpose of being an obstacle, shortcut, guardian or wrong turn - a part of the levels spatial puzzle to avoid. More complex level based design will create reasons for interacting with the lair: rumors of treasure, a potential ally, the subject of interest for factions in or beyond the dungeon.

Encounter based design usually takes a different approach to the lair encounter, beside the tendency to use simple maps where lairs/encounters act as necessary choke points, encounter based design, almost always provides mechanical benefit for overcoming enemies. Hunting monsters for experience points becomes a focus of the game, and there’s always a reason to enter a monster lair.

RANDOM ENCOUNTERS

A major design principle in classic play that varies from contemporary design is the use and reliance on “Wandering Monsters”. In the Keep on the Borderland’s goblin lair we this in the chance of encountering goblin warriors every turn the party spends in the active lair, but it’s often part of table of varied “Random Encounters”, especially in a larger level, or areas that aren’t controlled by a single type of monster. There are numerous ways to construct and include wandering monsters and random encounter tables in almost any adventure to many different purposes, but the primary goal of the random encounter is to create risk.

• Random encounters represent hostile or otherwise disruptive threats to the characters, whose appearance is a risk gone wrong on the part of the players - the principle risk being spending too long in the mythic underworld or dungeon. Because of the random encounter is effectively punitive, a failure, a chance taken for the worse (though a somewhat inevitable one) certain very specific contexts and rules apply.

• A Random Encounter must be random. This might sound silly or perhaps a constraint on story, but it’s utterly essential because trust is utterly essential to a game with an active risk economy (also the subject of the next Note on this blog). The players have elected to spend time in the dangerous environs of the dungeon and to determine if they want to take that risk it’s important for them to know what the odds are and that the GM isn’t bending the rules. There won’t be a meaningful decision of the players, or at least not an in game one if they know or suspect that doing well and having an easy time always results in their GM adding encounters in the name of excitement or challenge. Likewise, when escaping from the dungeon and finding shortcuts through it are an important part of the game a GM who allows the party to avoid random encounters when injured or running low on supplies disregards the players’ decision to take the risk of pushing on.

• Random Encounters can’t provide benefits. Players shouldn’t seek out random encounters and this is one of the reasons to forgo XP for killing monsters. With limited exceptions Random Encounters also shouldn’t provide treasure or reduce the population of monsters elsewhere in the level.

• Random Encounters should be unpredictably dangerous. To be truly worrisome, but also interesting, a varied random encounter table including nearly harmless vermin, representatives of level factions, and powerful monsters is useful. Varied foes mean that not only will the Random Encounter’s chance of occurring represent a risk, but that each encounter offers different risks and demand different responses from the players: combat, negotiation or flight. This also provides advantages to characters that explore and think about the dungeon level as they soon come to understand the risks of the Random Encounters they may face, and learn how best to approach each one. 

Preparing A Random Encounter Table

Random encounters are generally the product of a table, the GMing rolling to see the specifics of the encounter in play. When designing the table it’s useful to have both a variety of creatures, with a variety of power levels and responses on it but also to emphasize the level or dungeons themes.

The size of the table is important, but only because it gives a feel for the level. For the goblin lair of B2 there’s only one possible random encounter, and that’s fine - because of the context. With less than ten keyed locations and organized inhabitants whose defense plan is ready to activate very quickly the GM is unlikely to need a large number of random encounters, and it’s both naturalistic and mechanically sound that the one encounter is with wandering guards who will raise the alarm. Exploring the goblin lair is a stealth experience about scouting or reducing goblin resources for a later, almost wargame stye raid or confrontation.

For a more complex level the table will likely be larger, if only because the characters are likely to spend more time exploring and the entire level is unlikely to be the fortified base of a single faction (and even if it is there may be more than one kind of guard or resident to encounter). When selecting what sort of encounters to place on a random table the most obvious to include are representatives of various level factions: patrolling guards, gatherers, hunters, traders and travellers. These groups have the advantage of potentially introducing the party to factions they haven’t encountered yet, starting a new relationship, or building off past one. Faction encounters are more likely to offer an opportunity for roleplay and negotiation and reflect the characters past experiences in the location.

Beyond factions, both weak vermin style monsters and powerful aggressive predators offer a nice addition because they represent combat encounters that are draining but eminently winnable and combat encounters that are best to flee from. Additionally both provide potential for logical restocking of areas that the adventurers have emptied out - vermin will certainly hide there, and some kind of dangerous creature may decide to settle down and create a lair.

Lastly random encounter tables serve to provide clues and information about a level. Both the obvious information of what sort of monsters exist on the level, but even the history and mysteries. Creatures like spectral remnants of the past or other invaders from the surface can make excellent additions to the table and inform interested players about the story of the dungeon.
Random Encounter Procedure

The Overloaded Encounter or Exploration die method of managing the risk economy is discussed In the prior Timekeeping essay, and using it as a basis, Random Encounters occur when a ‘1’ is rolled. The chance of encounter can of course vary, as can some of the other effects on the die, for example  encounters can be added to the '2' replacing omen result or in addition to other effects supplementing them.  Special regional events or clocks (e.g. earthquakes or a countdown to a volcanic eruption) are also easy to add to the die, making it a very versatile tool.

Afterwards rolling the Exploration die the GM rolls for the encounter (or if a previous roll of ‘2’ has recently provided for a hint or sign about a specific kind of encounter uses that one).

Once the GM has established the nature of the encounter the next issue is where and when exactly it will occur. It almost never makes sense for the encounter to suddenly appear to confront the party, rather the wanderers need to come from somewhere: around a corner in the corridor, through the door of the chamber the party is searching, behind the next door the party opens. Dungeon design and the range of party light sources will usually make initial positioning clear (random encounters appear just out of view of the party - behind a corner, when a door is opened or at the edge of the light. In confusing conditions such as dim lighting or a forest a simple roll of a D6 or D20 (depending on the GM’s impressions of the location) multiplied by 10 or 100 equals the rough distance in feet between the wanderers and party.

After determining where an enemy appears from, the next step is determining surprise. While 5th edition’s Player’s Handbook rules that surprise is rarely appropriate “If neither side tries to be stealthy, they automatically notice each other.” Pg. 189. When appropriate 5th edition then uses a relatively complex stealth/dexterity vs. perception/perception check that while not unfair, given the tendency of many monsters to have high stealth and perception bonuses, is far more complex and dependent on character or monster design then earlier editions. In early editions of Dungeons & Dragons the mechanics for surprise are a simple, but discrete, subsystem - a simple D6 roll for each side capable of gaining surprise (generally opening a door or not carrying a light source) and on a 1 or 2 that side gained surprise, though some monsters and characters are immune to surprise or gain bonuses (up to 5 in 6 for camouflaged creatures). The player’s party is thus unlikely to gain surprise (being dependent on light sources) unless they quickly and efficiently open a door on a group of monsters, while monsters, who can see in the dark are more likely to gain surprise. A very useful additional rule regarding surprise, taken from the 1970’s Original Dungeons & Dragons Booklet vol. 3, Wilderness & Underworld Adventures is that “If monsters gain surprise they will either close the distance between themselves and the character(s) (unless they are intelligent and their prey is obviously too strong to attack) or attack.” Pg. 9. This helps balance the tendency of the reaction roll to lead to neutral responses and parley. It also heightens player tension as horror emerges suddenly from the darkness, while the protagonists, the party, are tied to their circle of light. It’s an evocative way to emphasize the alien hostility of the mythic underworld with very simple mechanics.

When surprise isn’t a factor the GM rolls for the demeanor of the wandering monsters. While the idea that monsters aren’t always prepared to attack a party of heavily armed intruders in the underworld has fallen out of favor, in later editions (likely as a result of the rise of Encounter design and the complexity of individual encounters) or become subject of GM interpretation, classic editions treat it as another random variable that affects each encounter. “Some monsters always act in the same way (such as zombies, who always attack). However, the reactions of most monsters are not always the same. The DM can always choose the monster's reactions to fit the dungeon, but if he decides not to do this, a DM may use the reaction table below to determine the monster's reactions (roll 2d6).” Moldvay Basic Dungeons & Dragon, 1981, pg. B24. The 2d6 reaction table used by Basic D&D (and other early systems) only provides a small chance that monsters will attack immediate, a roll of 2, or less than 3%. Far more likely monster will mill about, hostile or uncertain, delaying an attack or waiting for the party to act. This system creates a huge opening for player choice. If the party avoids being surprised they can attack, flee, negotiate or attempt to bribe wandering monsters with food or treasure. The fairly vague nature of the reactions: “hostile, may attack”, “uncertain” and such also offer the GM a good amount of leeway on considering the desire and nature of monster response. What does the “enthusiastic friendship” of an owlbear look like - sated curiosity, beak nudging and disturbing caws or just plopping down and staring amicably waiting for a treat? What do hostile but not immediately violent goblins do and what might they want - a toll? the party to back off? a chance to escape? a prisoner? 

Goblin from the 1975 D&D Booklet
"Monsters & Treasure"

FACTIONS

Some monsters aren’t solitary creatures, but exist as part of a more or less organized group. In the context of adventure or setting design these creatures are factions, and they are extremely important to level based design. A creature that is part of a faction has back up and a sense of being part of some order beyond itself, this effects both its immediate reaction to player characters and its goals. A faction member may be willing to fight for its fellows when a lone creature would run, or it might bluster even in a weak position. This isn’t to say factions are only intelligent enemies, or have to be large - a pack of cunning animal predators might be a faction and so might a dragon or single large power. The defining feature of a faction is that it controls some part of the dungeon level and that it has relationships with other groups within or outside the dungeon.

Factions provide the players with more than potential foes. They are key to building the story structure of the game world. Factions provide interaction along with jobs, quests and rumors. These interactions give players a chance to make decisions about the future of the campaign by determining which factions will succeed, fail and whose interests the campaign expands upon.

As factions are so important it’s imperative to have an idea of the leaders, goals, resources and attitudes of each faction. The goals and manners of almost every faction (especially human ones) work best if they are comprehensible, even sympathetic, but also in directly opposed to one or more other faction’s goals. The player characters are a powerful force for change and it’s more interesting if there’s no clear (or no easy) side to pick. Moral decisions are as much a part of play as exploration and battle tactics—faction play offers an arena for them. Factions relate to and interact with the game world and encourage player buy-in—even random encounters should be from somewhere or mean something to someone (e.g. almost everyone will be pleased if the party shows up with a fresh owlbear pelt—those things are a menace).

Factions and Level Design

Factions are a key concept for level based design, because unlike lair encounters the monsters within a faction operate to support each other. Faction affiliated monsters, if only because a single combat is unlikely to annihilate them, react and respond to player activity. A faction can become an ally, or an implacable foe - fortifying and trapping its base of operations and seeking to actively hunt the characters, adapting to their tactics depending on how the players treat it or its rivals. This is hard to accomplish when each encounter occurs more or less separately.

The spatial puzzle of level based design and the dungeon crawl give factions a chance to have both spatial and narrative power - attacking the goblin lair head on is dangerous because the goblins of the faction will support each other, work together and plan, making themselves more dangerous as a group then they would as atomized individuals. Narratively a faction maintains relationships with both the players and other factions, and thus its goals and fate effect the setting beyond its lair or the level it’s part of. The goblin’s in Keep on the Borderlands’ ongoing war with the orcs is the simplest sort of faction conflict and goal - they have an enemy the want to destroy.

Goals are a defining feature of the faction and understanding them, even if they are not something the faction can achieve, even with player help, helps the GM understand and role play the faction. For Example, the dead of the forgotten tomb hate all life and wish to end it because its proximity causes them pain. While living the characters will never be able to befriend these undead, but the GM knows what they will demand in any parley (that the party go far away or die) and what might convince them form a temporary alliance or offer them mercenary aid (massacring other nearby mortal factions for tomb gold). Indeed if the players take an inexplicable liking to the murderous dead they may aid them in building their army of souls, and eventually find themselves the root cause of a growing undead scourge.

Tucker’s Kobolds

Faction membership increases the character and power of individual monsters (though lair monsters often also have goals and character - perhaps making them factions as well - it’s a continuum, not a list.) Increased characterization always runs the risk of GM affection, and that in turn can create antagonistic GMing as the GM tries to save their preferred monster or NPC or the creation of a darling GMPC when the GM uses meta knowledge to always keep the NOC or monster one step ahead of the party.

An example both of the power of faction and these risks is the concept of “Tucker’s Kobolds”. Back in the early days of Dragon magazine, the house organ of TSR (Wizards of the Coast’s predecessor) Roger E. Moore, an early Dungeons & Dragons designer and writer, discussed a game with a GM named Tucker when he was in the army (as a therapist/counselor) where a tribe of lowly kobolds wrecked havoc on a group of high level characters through the use of clever traps, tactics and (antagonistically - but not entirely so) special rules regarding the difficulty of moving and fighting in tiny tunnels. Tucker’s kobold’s use a variety of common sense traps available to smart adventurers (spiking doors shut, attacking in formation, flinging flaming oil and generally using the spatial environment in the manner of Die Hard’s John McClane). While Tucker’s kobolds are too precious and intentionally designed to defeat the players as described, Tucker’s kobolds represent a faction maximizing its advantages. Faction’s should maximize their advantages within the scope of Gygaxian naturalism if perhaps not an intense study of modern small unit guerilla tactics.

In one sense this is where the level in level based design, the imperatives of Gygaxian naturalism, and the principle of the GM as arbitrator rather than antagonist all come together. A lair of goblins (or kobolds) should fight and scheme and work to the best of their ability, but that ability is not the same as the best of the GM’s ability, because the GM has knowledge about and power in the setting that a band of goblins never can.

Goblin 5th Edition D&D

COMBAT AS 'FAILURE'

Following from the naturalistic level based design, the idea of random encounters as a negative result of risk taking and coupled with combat mechanics that make encounters dangerous, one of the oft repeated catchphrases, maxims or cliches of more modern practitioners of classic play is “combat is a fail state”. With plenty of caveats and conditions, this is true in the classic dungeon crawl, but perhaps it’s better phrased as “combat is the near inevitable and dangerous result of player risk taking, lack of perfect knowledge, excessive caution or incaution.”

For Random Encounters this is most obviously true, random encounters impede progress and threaten character supply and survival. In the classic dungeon crawl they do so without providing mechanical benefit. Random Encounters are and should be a near pure risk as possible, enabled by the XP=GP mechanic. Among the advantages of separating character advancement from combat success is that it removes the necessity of combat. Players who can succeed at their goals (or even collect sufficient treasure) without combat gain levels and accumulate power at less risk than those that fight their way through. In turn the GM can place more powerful monsters because there’s no expectation that the adventure will proceed through a series of direct confrontations. When the intent, the ethic of play, is no longer to test the player’s tactical abilities but their puzzle solving ones, and thus the dungeon level is no longer “safe”, or at least the combat is no longer likely to lead to success.

In this context combat is most often the result of a series of suboptimal decisions by the players: failures of stealth, orienteering, negotiation, or luck. Even when players (inevitably) decide to violently confront a monster there’s also less impetus for direct combat. The players know that going into a fight they may be outclassed, and so trickery, traps and ambush take on greater importance.

A potential downside of using monsters as the adventure's primary mechanism of risk with direct confrontation as the high risk solution to the obstacle they represent is that it becomes necessary to include unbalanced and dangerous encounters.  It's no longer optional so the dungeon crawl won't deliver the fun and comfort many players get from knowing that they can overcome any foes they face.  This can make players used to encounter based design or with expectations that Dungeons & Dragons follows computer game style design principles, very frustrated and often drive them away from classic mechanics, design principles and ethics.  Yet, the risk economy is essential to exploration play, at least classic exploration play, and it's far harder to change the way its mechanics interact then to help unfamiliar players adjust to its ethics of play.