Immersive Story Methods for Tabletop Play

NOTE: This was an essay contribution for the magazine PUSH. It is now released free both here and on the PUSH website.

         This is an outline of methods of preparing for tabletop play, based around a particular style of play. There are many functional styles of play, but this is a specific prescription based on my concept of "Immersive Story" (Kim 2004). In brief, that analogy suggests that each player sees the game as a dramatic story with her own PC as the protagonist. This includes all events seen, including those of other PCs. Because the player also controls the actions of her PC, this results in a deepened emotional identification.

         This is a deceptively simple concept with many consequences. The most striking difference from traditional dramatic forms is that the player is not required to externalize her PC's inner conflicts. That PC is not the protagonist to anyone except the player in question -- i.e. the other players need not be drawn to emotionally identify with it. However, that PC is a supporting character in the perceived stories of all the other players. Thus, play should be structured so that these supporting roles are strengthened.

         The concept of a singular protagonist is a limited and even simplistic model of drama, going back to Aristotle. There are many more advanced drama concepts that I do not cover here. However, at a broad level I believe this is a useful approach. Role-playing theory is at an early stage, and we first need to emulate Aristotle before we can have a Brecht. Following the singular protagonist model, we can define other styles of role-playing which contrast with Immersive Story as I have defined it. I will suggest three:

These differ not only in methods, but also in the ideals that they try for. The ideal in Immersive Story is different. It is for each player to come out of the session having consistently identified with their own PC, and felt that PC go through an interesting dramatic arc.

         I will give a hypothetical example first, to illustrate. Assume a traditional medieval fantasy game. Take a scene where a medieval fantasy hero is rescuing a princess, with the forced help of a scoundrel. The player of the hero views the game as something most like a traditional heroic narrative. To the player, this scene is a climactic conflict. The player of the princess views the game as something like a modern feminist re-telling, perhaps like Mary Stewart's The Crystal Cave. To her, this scene may function as a shift, which is the start of an arc about her reintegration and/or romance. The player of the scoundrel, who is a rival of the hero, views this as a satirical twist, such as John Gardner's Grendel or Gregory Maguire's Wicked. To him, this scene is tension-building towards a future confrontation. The overall result is that none of the players will experience a well-structured story in the dramatic sense, but this is balanced by the power of their interaction.

         In this essay, I want see how this theory implies concrete suggestions for game preparation. I will draw some examples from an actual tabletop campaign which I game-mastered, the Water-Uphill World campaign. In that game, the PCs were schoolchildren from the modern world who find themselves in a bizarre fantasy world where water literally runs uphill. The world is rocky and has scattered "geysers" where water falls in jets up into the sky. There were four kids: Noriko, Kate, Martin, and Steve. They were wandering in the basement under their school, and then inexplicably found themselves in a tower of the Royal Palace. The Palace was a floating island -- a huge upside-down bowl of rock with a reservoir of water underneath holding it up. I will describe more of the campaign as I consider different parts of preparation.

Story Soup

         The ideal of Immersive Story is inherently chaotic. Because each player interprets the scenes differently, nearly any formal structure is unworkable. There are as many different goals for the scene as there are players, and thus there is no best way to handle or resolve it according to traditional dramatic principles. This multiplicity of function makes story planning essentially impossible, and even judging a single scene based on dramatic logic is difficult. However, there are methods to increase the chances that as a whole, the scenes will have impact for all the players.

         One analogy for this approach to play is making a stew. Rather than trying to arrange ingredients in some sort of structured narrative, you as GM toss them all into the pot and see what comes out. The dramatic direction should not come from what the situation is, but rather from the PCs. In this approach, the GM is responsible for the balance of ingredients, and occasionally stirring the pot, but not how it is arranged.

         In game terms, this can be broken down into several pieces:

The purpose of these preparations is to give a foundation from which to run your game. At any point in the game, you as GM should be able to give an answer that both maintains continuity and continues the action. Having a narrow scope means that it is possible to prepare all the elements in that scope. I have earlier work that discusses the use of background (Kim 2003). However, the narrow scope also means there is much greater demand for continuity.

         Note that I am emphasizing the role of the GM. However, players can be involved in the design of all the background elements. Play in this style is powered by interactions and developments, not revelation of secrets. So even if the players know and/or designed the elements, they can still be engaged by interaction with them. In Water-Uphill World, Liz (who played Kate) was involved in the world design.


         The first key is defining a Scope. This is the boundary line of the stew pot -- the set of locations and characters and objects that the GM has prepared. The scope needs to be small enough that the GM can have it adequately detailed, while still large enough to be interesting. As a rule of thumb: consider the PCs and perhaps a dozen key NPCs. The scope should be small enough that they are required to keep bumping into each other regularly. By keeping the game within the scope, the relations become deeper and more meaningful. The limited scope means that the games are more static in location, rather than adventurers constantly wandering.

         The scope could be a limited physical location: such as the small home village of the PCs, or a covenant and its surroundings in Ars Magica (Tweet 1987). A city as such is too large to be an adequate scope -- i.e. characters can live the city without regularly bumping into each other. However, the scope may also have conceptual boundaries rather than just physical boundaries. For example, it could be everything related to a set of particular characters, like an extended family. In my Vinland campaign, the scope was a set of five extended families, each with their own homestead. Or it could be a set of things all related to some unusual feature or event in history. With conceptual boundaries, you need to be able to improvise setting details.

         The scope needs to be highly detailed. Because the people and places in it will keep getting re-used, there are much greater demands on continuity -- compared to, say, a wandering quest. Each element in the scope may be looked at repeatedly from different angles and under different circumstances. In a single-story model, you can selectively detail an element because you know that element's purpose in the story. Within this more chaotic model, there is no singular purpose. For example, an NPC who was at first a negotiator turns out to be a romantic interest of a PC. Now, how does courtship work within her culture and social class? Because of the need for detail, it can be easier to design from a modern-day, historical, or alternate-history setting. These typically have a wealth of reference material for detail.

         The familiarity of the elements is also important. For example, the Holy Roman Empire of 1802 is reasonably well documented by historians, but it will be difficult for players to internalize the details. Familiar elements take on an iconic value, which is a useful shorthand. In my Water-Uphill World campaign, this was actually a difficulty in getting started. The game was in the style of children's fantasy stories such as Narnia, but the setting was quite alien. In contrast, my following campaign mixed two familiar historical elements -- the Vikings of the Icelandic sagas, and Northeastern Native Americans of early contact. Although it was an alternate history with many altered details, the use of familiar, iconic elements made the latter campaign easier to get a handle on for the players.

         In order for the scope to work, play needs to actually stay within it. Over time, the scope may widen or shift. But that should be a slow and gradual process. Thus, you need to design in strong reasons that keep the PCs' interests within the scope. For example, you might have a campaign based around two warring cities -- with PCs in the military of one city. You need to detail the two cities, focusing on their military culture and capabilities.

         In Water-Uphill World, the scope of the game, at least initially, was the Palace. It was a big place with over a dozen buildings and lots of factions. As a floating island, though, the PCs had no immediate means to leave it, which inherently prevented the scope from creeping too much. After six or seven sessions, the PCs left and went to an underground city. I tried to detail that as well, but it was sketchier in the end.


         I generally divide the scope into Factions -- these can be formal organizations, clans, or just ideological groupings. The factions might be as simple as the two opposing sides in a war. There should probably be at least four factions, but no more than ten. These may include sub-groupings within a larger unit (like vying political blocks within one side of a war). Factions are necessary because if you are going to keep a handle on ongoing events, you need to have some abstraction larger than individual characters. You need to be able to generalize with things like "This PC action will piss off members of Faction X", rather than relating to each individual NPC.

         You should be careful about developing individual characters within the factions, though. Characters in practice often take on a life of their own, especially as they are influenced by the surrounding players. It is more stable to prepare in advance the broad tendencies and relations of the factions, and fill in more details about the individuals only once they have entered play. By having the framework, it becomes easier to write up individuals. So when I write up an individual NPC, I am not just creating out of nothing. I am creating them within the known framework of relationships. This tends to give the NPCs a sense of grounding in the fictional reality -- i.e. they are not created to be opponents or allies to the PCs. Instead, they are created based on "What would the second-in-command of X faction have to be like?"

         You should make sure that the factional conflicts run deep and are difficult to resolve. For potentially violent campaigns, this means coming up with reasons to restrict open killing. In short, you don't want a little push to bring the conflict to armed battle that results in the immediate elimination or subjugation of one side. However, as long as the factions don't violently eliminate each other, interaction is fine and helpful. Two factions may negotiate to resolve a particular disagreement, but there will remain underlying differences that divide them. Restricted violence also gives some stability to both PCs and NPCs. The PCs as a whole should not constantly be in danger of dying. Similarly, important NPCs should not frequently be killed off. Such characters are difficult and time-consuming to develop. They should die under the right circumstances, but the context should be designed to make it less likely. Such stabilizing forces can also mitigate the effect of a "loose cannon" among the PCs.

         More generally, there is a balance between the setup being too rigid and too unstable. If there is too much stability, then the PCs will be unable to effect change, which can be frustrating. On the other hand, if it is too unstable then you may have a bunch of carefully prepared characters quickly killed off, or scattered to different parts of the world. In short, your preparation becomes quickly invalidated.

         In the Palace of Water-Uphill World, there were three major factions and three powerful independent individuals. These were the Regent, the Nursery, and (not important politically but important for Palace life) the servants and artisans. The independent individuals were the princess, an unusual warrior, and a giant spider. These were not developed with a coherent view of factions. In practice, the players avoided the Regent and the Nursery, while hanging out in other parts of the palace interacting with the artisans and individuals. Effectively, this shrank the scope more, and they left the Palace sooner than I would have liked. The individuals were generally liked, but understanding them didn't contribute to comfort or familiarity with the Palace as a whole.

         In retrospect, this was not well planned out. In my next campaign, I ended up having very strong and well-defined factions -- there were a small number of tribes and clans. Each clan had a family tree, and the players all had a sheet of the family trees. This meant that an individual had a distinct and easily displayed position within the faction. I felt this worked much better for giving both the players and myself an understanding of the scope.

The Player Characters

         The PCs, then, are created by the players and set loose within this scope to interact. The PCs are by far the most important ingredient. Each other PC is generally far more prominent to a player than any tabletop NPC or short-term character. There are a few things to consider for this. The PCs should be powerful relative to the scope. As a rule of thumb, at least one PC should probably be on the top ten list of most powerful individuals in the scope; and the PCs should be a major concern for the most powerful. The PCs may or may not be an integrated part of the scope -- i.e. they might fit within the existing factions, or they might be external to them. If they do not start out part of the factions, however, they should develop relations. The factions should eventually seek alliances or make enemies of them.

         You should collectively design the PCs to have an overall bond which keeps them together. Given the chaotic approach, it is important to have ties that are difficult to break -- i.e. more than simple friendship or common interest. Relationships can and should take unexpected turns in play. You might intend during design that two PCs are friends, but they soon have a terrible rift over some incident. Particularly in pre-modern societies, I am fond of using blood relations. Friends can split, but a brother is always a brother even if you don't like him. Another bond could be that the PCs are all officers assigned to a particular ship. In Water-Uphill World, the PCs were the only people they knew from their world, which forced them to depend on each other.

         For individual PCs, the most important issue is that they be proactive. This involves attitude of both the players and the PCs. Good PCs should have ambition. This need not be selfish ambition -- it could be ambition for a cause. However, what this amounts to is an unwillingness to settle for the status quo. (Kim 2004)

Relationship Charts

         Given a set of individual PCs, the most vital part is the relationships among the PCs. However, the relationships in practice are outside of the GM's control. The GM can plan and suggest relationships, but the results from play can and will vary from this. One tool for charting out the relationships is to create a chart. The idea of the chart is that it cross-indexes characters, so if there are 8 characters it is 8 x 8 for 64 boxes. Each row represents a character, and specifically that character's story. Each column represents a supporting character in that story -- i.e. one of the other PCs. What is being documented here is what potential impact the column character has on the row character's story.

         So, for Water-Uphill World, we might have:

Protagonist Story Supporting Cast
Noriko Kate Martin Steve
Noriko taking responsibility X impractical bookworm irresponsible adventurer
Kate learning an adult world irritating X
Martin finding passion lacks vision fellow explorer X
Steve getting power goody two-shoes X

         Now, the important part for me was filling in the blank spots. The ideal here would be that each PC has an important supporting role in each story (except, of course, his own). But this isn't an exact guide. All the stories and roles going on are bound to clash with each other -- so you're just trying to maximize your chances rather than nailing each one.

Externalization & Symbolism

         One criticism of immersive story is that without clear cues from traditional single-protagonist story structure, the players' attention can become lost in minutiae. The multiplicity of conflicts and themes can be overwhelming. To offset this, you can try to build in clarifying elements -- symbols that externalize the conflicts within the setting. In keeping with the principles of the style, a symbol should be:

  1. Intra-diagetic. The object needs to speak for itself, not be the result of external description or emphasis.
  2. Discrete. The symbol should not be construed as leading in a particular direction, or requiring a particular arrangement to look right.
The point here is that symbols should not be guides. They are directionless, part of the soup that the players manipulate. When the players pick a direction, they take on new meaning by proxy.

         Within Water-Uphill World, Magic served this function. I had conceived of Magic as a place. The PCs would concentrate deeply, and then find themselves in a maze where there were certain unusual objects that they could carry. Each unusual object represented a magical power that the character would have in the normal world. So, for example, Kate in Magic found that she had a bag, which contained things that she touched in the real world and came to Magic with. This was a sort of eidetic memory. At any time she could go to Magic and read a book which she touched.

         The maze started in the middle of a four-way intersection. Secretly, I had determined that each direction was a branch of Magic. Magic was divided by social types rather than physical. I had a four-way branching corridor. There were four signs, but rather than labeling the four directions, each sign was exactly in-between two directions. The signs were:

                             | |
                        (A)  | |  (B)
                      ______/   \______
                      ______     ______
                            \   /
                        (D)  | |  (C)
                             | |
            A) "This way to higher understanding."
            B) "This way to ultimate dominion."
            C) "This way to your heart's desire."
            D) "This way to fulfillment."
So the AB corridor represented political or social power. The BC corridor represented material wealth. The CD corridor represented love. And the DA corridor represented knowledge. In principle, along each branch of the maze, they had opportunities to get more objects, but also to lose or trade objects they had.

         In practice, all four of the kids were most interested in the knowledge direction at first. In short, they were curious rather than greedy. In that direction, there was a dead end with a hole in the wall, which gave a periodic grinding sound (about every 1.5 seconds). There were no clues here -- whoever comes seeking must put their hand inside the hole to go further. The wall turns once a hand is placed in, letting them into the room beyond where further puzzles awaited.

         One of the most interesting twists of the game, though, was when Noriko went down the AB path. That had a clear explanation of what it was. The hall ended in a stone door, beside which was an alcove where an iron rod stood on a pedestal. A sign beside it explained:

This is the First Rod of Power, which conveys to the wielder authority in arguments. With it, when you pressure them, others are influenced to concede to your arguments not only on the surface, but also in their true belief. It will also open this door. However, once grasped it may never be put down.
As it is magical, it will not encumber you physically outside of this place. However, its power will be felt and can never be turned off or discarded.
The symbolism for political power should be fairly clear. But I had not arranged at all how or even if this would work into the plot. It was just an inherent power. During the campaign, Noriko was at first a constant voice of caution to the others. At some point, when the others had split up, she impulsively went into Magic and picked up the rod. She would use it on others to protect the group. However, from that point on, she would never speak when the others discussed what they wanted to do, since she didn't want to use the power on her friends.

The Game in Play

         How the game should be run goes beyond the scope of this essay. There are enormous variations of techniques -- including systems used, means of negotiation, and so forth. However, there are a few principles that I should mention specific to these preparations. The GM should work with the players to provide starting motivations or motivating events for the PCs. However, the intent is that the preparation should be flexible enough to work even if those initial goals change. i.e. You are not preparing for particular plotlines, but rather have a collection of elements, which can be used for a variety of plots.

         So you prepare the scope and factions, and create the PCs. Within this context, it should be relatively easy for either the GM or the players to generate plots. Given sufficient information about the scope, the players can make proactive plans to do things within the scope. Alternately, the GM can consider the factions at a higher level to come up with what would reasonably happen, which may spur the PCs to activity. At the beginning, the game may be started by pre-planned plots by the GM. Over time, however, the players should become more confident of actions within the scope and create their own plotlines.

         There will still be times where the action is slow. These are often useful and good. The players should drive the action, which means that at times they will choose to sit back. Particularly for a tabletop game, it is useful to be able to fast-forward in time. So, between sessions or during a session, the group can agree to skip days, weeks, or months ahead in time. The GM should project future developments which come as consequence of prior play. So, for example, after the PCs inform one faction of a secret, you might skip to a time when the consequences of that occur. The abstraction of factions is important here to simplify such projections. Simply consider what would happen with each group.

         As GM, one should be wary of adding results or encounters based on dramatic logic. It is easily possible to inject interest by having well-timed or unexpectedly exciting outcomes. However, over time this creates a visible pattern of having exciting parts as a result of GM resolution, rather than as a result of player choices. In such cases, players often come look at events and encounters as things that the GM feels are supposed to happen -- and indeed they are correct. Within story soup play, the patterns should arise from PC choices, and not from GM-planned coincidences.


         Role-play has often been characterized as a clash, with story and drama on one hand, and game and simulation on the other. Story is thought of as requiring pre-planning of the plot, themes, or issues. However, within roleplaying games, the dramatic identification is very different than in traditional stories. I would argue that story-based games all too often naïvely imitate other narrative forms, and ignore what is different about roleplaying. They frequently dismiss methods that do not imitate static media fiction.

         The heart of roleplaying is in the identification between player and character. This is fundamentally different from static media such as movies, where every audience member has the same relation to what is portrayed.

         I have outlined one broad model for exploiting that difference. So this is a prescription, but it is not intended to be exclusive. There are and should be other models and methods that are unique to roleplaying.



John H. Kim <jhkim-at-darkshire-dot-net>
Last modified: Tue Jul 12 18:11:20 2005