I’ve been trying to wrap my head around a practical way of engineering the process of writing a different kind of adventure for several years, and think I’m finally starting to get my head around it. There may be a technical term for this type of adventure, but if there is, I don’t know it.
I’m not even sure that I can explain it clearly, let alone offer an example. So you might need to bear with me as I plunge headlong into this exploration of the evolution of patterns in random events.
Life is full of stuff that just happens. I’ve written about that in the past (Directed Plots, Undirected Narrative, and Stuff That Just Happens), but for a long time I’ve been thinking about extending the concept but putting some structure into the structureless.
The idea is that you have a series of unrelated events that occur in a completely unstructured way that nevertheless combine to form a broader narrative in a somewhat more structured fashion. Which is like saying “blue is a color” – it doesn’t actually tell you anything useful.
So let me try again: One random event follows another, but no event happens in isolation; the first provides a context for the second, and the second provides a context for a third, and so on.
The Evolution Of Patterns
This seems like an example of the phenomenon where people impose patterns to events not because there is an actual pattern but because humans are psychologically predisposed to see patterns even when none exist. But I want to be able to impose a non-obvious pattern onto these events that only crystallizes into clarity as the final piece of the puzzle falls into place. So there is a pattern there, after all. But how to achieve that without railroading the players? Because it all sounds very contrived.
That’s the dilemma that’s been percolating away at the back of my mind ever since the aforementioned article: how to link these apparently-random events together in a structured and deliberate way that remains opaque to the players until the right time without contrivance, while maintaining player independence.
Snippets In Parallel
The solution began to present itself by thinking about the game of chess. There are a very limited range of choices for any individual move, and the compounding of past moves places a context on those choices that eliminates many of them in practical terms of achieving the goal. Nevertheless, the number of possible configurations of the game explodes into the uncounted billions or more as it progresses while at the same time funneling down into a very few outcomes in an entirely natural manner – a game can be a draw, a victory for white, or a victory for black.
Using all that as an analogy for progress within an adventure began to give me an insight into the correct way of selecting and structuring the “random” events to confer the structure upon the adventure.
For a start, the “random” events would not be random at all; they would be dictated by the overall structure, like pieces in a jigsaw. Each piece in isolation would seem to be nothing more than a random splash of color, but put them together and they form a picture, placing each of those splashes into context. Each event would be a snippet of a series of partial stories existing in parallel with each other, threads in a tapestry, corresponding to the restricted repertoire of individual moves available to the pieces on the chess board at that point in the overall game.
The Circumstances Lend Significance
The next piece of the puzzle came from contemplation of how these snippets would interact. Each would have to be designed to achieve a certain mental, emotional, or interpretive response; what the actual event was that produced the response was irrelevant to this new style of adventure narrative. So long as each “move” achieved one of a limited set of predictable outcomes in terms of the overall plotline, the events themselves could be completely random, or could even form a separate and distinct pattern of their own.
Further insight came from pattern recognition itself, when I realized that a huge amount of an image can be thrown away and because of our ability as humans to recognize patterns, we could still identify the subject of the overall image. To illustrate how powerful this ability is – and how susceptible we are to be misled by it, as a result – take a look at the series of images to the right.
- The first image is 100% complete.
- In the second, I’ve thrown away half of the visual information – and the human mind barely notices.
- For the third image, I’ve thrown away half of what was left – for every stripe of pixels with visual information, there are three black stripes of the same size. With only 1/4 of the original information, we definitely notice – but equally, we have no trouble recognizing the image despite the incompleteness of the information.
- With the fourth image, I started getting serious – blocking out 1 in four horizontal rows as well as the vertical stripes. There is less than 19% of the original left – but the lake, trees, sky, and mountains are all still readily recognizable.
- The final image really goes almost all the way. Only 6.25% of the original information is left – for every pixel of information, there are three columns and three lines in both directions that have been blacked out. Thunderbolt: The image content can still be seen!
An interesting aside: notice that color information is lost a lot more quickly than pattern recognition. The last two images are effectively black-and-white to the eye.
The Order Gives Meaning
In part, this is a result of looking at something very recognizable. If the image was very abstract, or something we don’t know very well, we would have rather more difficulty. In part, it’s conditioning – having already seen the image in more complete form, we recognize the elements that we can see and mentally stitch together these fragments to form a more cohesive whole. But mostly, it is because we see all these separate pieces at the same time.
If these isolated pixels were presented to us in random order, we would not be able to make rational sense of what we were seeing. Even randomly ordering the vertical stripes of information from ANY of these images (including the one that is “all there” completely destroys the meaning. We rely on the relationships between the discrete elements that we perceive and fill in the gaps with ‘something in between’ to produce a synthetic composite in our heads of the information that is there and that which is not, then interpret that as a whole image.
So there are two relationships that matter between these discrete elements: the relationship with the packets around it, and the order in which they are presented. Both need to be present, and to make sense, in order to have a narrative consisting of discrete scenes form a coherent whole.
If either are missing or obscured, the overall story will make no sense until the missing/obscured elements are supplied. Once that is done, the story begins to make sense as a whole.
States Of Ignorance
How to tell a structured story with none of the elements that connect one piece to another? One way is obviously the move-and-countermove of the chess game, with one side blind to what the other was doing. Picture a situation in which white only knows the position of those black pieces that are in immediate contact with one of his own, while Black can see the whole board.
In a way, that’s an analogy for every adventure, and one of the key pieces of advice that has been offered over the years is not to confuse the omniscience of the GM with the perceptions of the enemy characters. If white were PCs, then both sides should be blind to anything they cannot directly perceive.
But in this case, White is the players, and Black is the GM, wearing his story-teller’s hat. And that’s a horse of a completely different color.
It’s entirely normal for the GM to make clear the relationship between cause and effect, linking one scene in an adventure to the scenes that come before and after, so that the overall plotline is an emerging picture in the minds of the players. Under this structure of adventure, I don’t want that to be the case; I want the events to seem random and disconnected until the whole comes together.
Another way of looking at it is the three blind men trying to describe an elephant – one finds the body, another the ears, and the third the trunk, and all get very different impressions. The goal is to be able to put the PCs into the position of the blind men and the plotline in the place of the elephant. In fact, we want to use a simulated elephant so that the three (or more) pieces don’t even seem to connect, because there are in-between parts missing.
Dismembered pieces of plot
Trying to plan something like this without a preordained outcome from each of the scenes is a total nightmare. It’s hard enough doing so when everything connects seamlessly with cause following effect in a more traditional story structure. Solving this problem was the hardest part of the whole project.
If the PCs only get half the story until the penultimate stage of the adventure, when we want it all to come together, the easiest method is to have the outcome – whatever it may be – of each step in the plot to determine not just which piece of the overall plot they are going to get to see next, but which parts they aren’t going to see.
At the same time, these plot elements are not going to be predefined. The players and GM are to free to let the plot grow organically, so that the player’s awareness of events and their responses to it appear to form a pattern that is sensible to them. Only at the penultimate stage of the adventure will the missing pieces be supplied, so that the players can discover the real story.
The solution to the problem is to define scene Templates, not scenes; then fit the choices of the PCs into those templates to produce the next step in the story as experienced by the PCs. These templates define the pieces of the puzzle – the ones that are present, and the ones that are missing. They permit players to chase after red herrings or instigate direct action aimed at solving the mysteries with which they are confronted. Nested beginning-middle-ending “loops” as Johnn used to call them, or subplot arcs, to use my preferred terminology, define a template and make each scene a complete mini-plotline unto itself.
Defining these templates in fairly generic terms based on a generalization of the disassembled “elephant” permits them to be strung together completely independently of each other. What is then needed is a scaffolding to put the real structure in place.
A structured scaffolding
It was figuring out how to construct the structured scaffolding that was the key to solving the problem, and was the item that has caused the greatest delay. Finally, though, I think I’ve come up with a way of doing so. And it was a solution that was staring me in the face the whole time.
I started out by thinking of the situation using a chess analogy, and that is where the solution lies. Picture this: Turn a chessboard to a 45-degree angle, so that one of the black squares becomes a diamond at the top. This is the initial subplot that launches the whole adventure. From this initial position, you can move to any of the adjacent “squares”, or to any “square” on the same row, defining the plot template for the next part of the adventure. If they stick to the black squares, the white squares would represent the pieces of the plotline that the PCs don’t get to discover.
Alternatively, the PCs can make a different choice and move from a black square onto one of the white squares. That means that they then start discovering the white-squares part of the plot while missing out on the black squares. Changing colors is effectively the same as a plot twist.
In terms of narrative, it’s all move-and-countermove. The opposition or overall problem makes the first move; the PCs respond, dictating which course of action they will be following when the next piece of the plotline happens, and placing the template of the situation into context. The enemy then makes a countermove – this can either be obvious to the PCs (an event on the same color as the square they currently occupy) or can be hidden from them (an event on a square of the opposite color). A player can even have a flash of insight or make a correct assumption and reposition themselves on the board on an entirely different row, from which events then develop. For a while, the number of ways the plot can shape itself grow exponentially. But, as you proceed down the board toward the far corner, your range of options narrows until the PCs are either on one of the two bottom-most black squares and need a plot twist to get them into the final white square, or there has already been a plot twist, and they will be on one of the three white squares that lead to the ultimate conclusion of the plot.
A three-square-sided illustration
This illustration should make what I’m talking about a little clearer. The possible outcomes from the starting encounter are b, c, or e. From b, the characters can advance directly to encounters d, e, or g, or indirectly to c or f. Note that boards with an odd number of squares do not require the plot twist that changing colors are specified to entail, but if there is a plot twist with an odd number of squares, there has to be a second plot twist to get to the final solution, i.
With that clarified, I can get back to explaining the structure that I had in mind. If characters move to the right as shown (from a to c, for example), it represents a Red Herring (1 space), a diversion or mistake (2-3 spaces), or a false trail (3+ spaces). (From the direction of movement, this signifies something from “out of left field”). The most direct line, a to e to i, will not normally be allowed, because going directly to e means that the characters won’t have the necessary information to enable them to get to i.
So, let’s list the valid solutions:
- a – b (plot twist) – d – g (2nd plot twist, red herring) – i (3rd plot twist)
- a – b (plot twist) – d – g (2nd plot twist, red herring) – h – i (3rd plot twist)
- a – b (plot twist) – d – e (red herring) – g (2nd plot twist) – i (3rd plot twist)
- a – b (plot twist) – d – e (red herring) – h (2nd plot twist, red herring) – i (3rd plot twist)
- a – b (plot twist) – d – f (diversion/mistake) – h (2nd plot twist) – i (3rd plot twist)
- a – b (plot twist) – e – g (2nd plot twist) – i (3rd plot twist)
- a – b (plot twist) – e – h (2nd plot twist, red herring) – i (3rd plot twist)
- a – b (plot twist) – c (red herring) – e – i (2nd plot twist)
- a – b (plot twist) – c (red herring) – h – i (2nd plot twist)
- a – e – g (plot twist) – i (2nd plot twist)
- a – e – h (plot twist, red herring) – i (2nd plot twist)
- a – c (plot twist, red herring) – e – g (2nd plot twist) – i (3rd plot twist)
- a – c (plot twist, red herring) – e – h (red herring, 2nd plot twist) – i (3rd plot twist)
- a – c (plot twist, red herring) – e – i (2nd plot twist)
- a – c (plot twist, red herring) – f – g (2nd plot twist) – i (3rd plot twist)
- a – c (plot twist, red herring) – f – h (2nd plot twist) – i (3rd plot twist)
There are undoubtedly others as well. These don’t cover the innumerable “flash of insight” options discussed a moment ago, for example.
Every 2nd step, it should also be remembered, represents a countermove or response by the villain or opponent. a-b, a-c, or a-e are all moves by the players in response to the initial move from a.
The Initial Encounter
“Thief & known pickpocket bumps into PC Alpha while running from the watch” sounds like a fun opening encounter. So we would have the ‘where are you’ establishing narrative, the collision, and an encounter with the watch. The thief may get away, he may not. The watch will tell the PC who the NPC is, and probably advise him to check that he still has all his valuables about his person. He may do, he may be missing something. That’s the sum total of our opening sequence, space ‘a’ on the board.
Thunderclouds on the Horizon: The initial situation
Having decided the first encounter, with some details still to be decided, we can now establish what the overall plotline is going to be about. The only restriction is that it seem to have absolutely nothing to do with the initial encounter.
“Necromancer/Villain steals magic item to use to twist the spirits of Elysium/Heaven to the cause of Evil.” Not bad. “Businessman attempting to unseat the head of the tradesman’s guild, PCs get caught in the crossfire” probably works better at our example 9-squares scale, though. The first idea would have required too much discovery to fit within the 9-boxes constraint – even though it’s the idea that was lurking in the back of my mind as an example throughout this article!
Little plot threads, all in a row
Next, we count the number of rows of diamonds, both black and white, ignoring box a. There’s the b-c row; there’s the d-e-f line; the g-h row; and finally, of course, the i square. That’s four. (I’d need to check more examples, but I immediately find myself wondering if this count is always one more than the number of squares being used as counted on the side of the board? Never mind, it’s easy enough to count each time).
To each of these, we assign a generic label that describes a step in the plot:
- b-c: PC attacked
- d-e-f: PCs investigate a mystery
- g-h: One enemy neutralized
- i: Revelation/Discovery, Confrontation, Aftermath
Notice that these are quite independent of each other. We know that i is the conclusion of the adventure, but everything else could happen in any order, and any of these items could be left out without compromising the final plotline.
taking a compass bearing
On any given line, there are only a few possible combinations of entry specification (plot twist, red herring, etc) and event label (PC attacked, etc). What’s more, which of those comes into play depends entirely on the choices the PCs make; each of the above sequences is independent of each other, but they can combine to tell a coherent story. It sounds like we’re all set.
Or are we? It’s all well and good to say “from a you can go to b, c, or d” – but how is the GM to choose? There are no labels to these choices. We know that two of them involve plot twists but that’s all about the plot after the reaction to what the PCs do in scene A.
Actually, there’s one thing more that we know, and it’s that information – in combination with those other factors – that defines the solution to this remaining problem. We know that each square consists of both a PC reaction and an NPC action. We don’t know which one goes first, because that depends on how we get into that square, something that we can’t predict, as the number of valid combinations listed makes clear. In general terms, the antagonist is either responding to what the PCs have just done (a ‘predictable’ choice) or is responding out of ignorance of what the PCs were going to do (a ‘wild’ choice).
In theory, the plot can proceed from its current square to any of the adjacent squares that we haven’t been to before. To try and give the plot impetus and move it toward a solution, we have stated that the preferred direction of travel is downwards, and the second choice is side-to-side, but that’s an ideal.
Expanding the plot
What needs to be done is to further detail the cell contents based on the generic label and the things that we DO know about the cells. That then permits us to choose the action that most makes sense under the circumstances of what the PCs have done and what the antagonist knows about that action. (This is where the true limitation of this small example layout becomes apparent – we don’t have very many choices to fill. It would normally be the case that we would have a lot more choices to fill – as many as eight in a single line if we were using a full 8×8 chessboard). So let’s do that now (my sidenotes on the process are in italics):
- a: Initial scene. The thief wasn’t trying to steal from the PCs, he wanted to plant something amongst the PCs effects to use the PC as an unwitting mule to carry the goods away from the scene of the crime. This is the opening move of a gambit by one of the antagonists (call him enemy #1) to get the PCs to eliminate his rival (enemy #2) for him.
- a: After the ‘pickpocket strike’ the PC discovers a small handbook that he has never seen before, filled with numbers, possibly a code of some sort, hidden amongst his possessions. This is actually accounts information, evidence of bribery and corruption on the part of enemy #2.
- b-c: PC attacked. Which PC is attacked, why, and by whom? Those are the critical questions. In our example, we have two defined factions working at cross-purposes, so this is easy. When you don’t have a predefined solution, or you need more than two answers, you may need to get more creative.
- b: Enemy #1 arranges for Enemy #2 to learn that the PCs have the stolen incriminating evidence. Enemy #2 sends forces to attack them and get it back. The plot twist is that a potential ally is being cast as an enemy.
- c: Enemy #1’s forces, disguised as belonging to a third party (enemy #3), attack the party to retrieve the planted evidence. This attack is intended to fail. The fact that the attackers are not who they seem to be is intended to be easily discovered; enemy #3 is someone with the ability/resources to decode the notebook and the authority to remove enemy #2 from power. The plot twist is either that the PCs will fail to penetrate the deception and make enemy #3 an actual enemy, or that they have been successfully manipulated by enemy #1 into acting against enemy #2, a potential ally.
- d-e-f: PCs investigate a mystery. We now have three factions other than the PCs embroiled in this little plot. The PCs have one three choices to pursue: they can try to decode the book, they can try to chase down the thief (if he wasn’t taken away by the watch), or they can pursue whoever they think attacked them either by going to enemy #3 (requires c) or learning about the rivalry between enemies #1 and #2, then approaching enemy #1, thinking him a possible ally (requires b).
- d: The PCs attempt to decode the book’s contents either themselves or by seeking out an expert. If this leads to a plot twist, the expert will discover that the book is a forgery. This is the most direct route to the conclusion, confronting enemy #1. …which is why it is the left-most of the options. Either way, the nature of the book will be revealed, telling the PCs part of what is going on, but not who is doing what to whom.
- e: The PCs pursue whoever they think attacked them either by going to enemy #3 (requires c) or learning about the rivalry between enemies #1 and #2, then approaching enemy #1, thinking him a possible ally (requires b). Ultimately, this boils down to the continued success of enemy #1 at manipulating the PCs, because they are continuing to follow the sequence of events he has mapped out for them. Enemy #3 will be able to identify what the book purports to be, but will not discover that it is a forgery.
- f: The PCs chase down the thief. Having fulfilled his role in the plot, the Thief is now a red herring. When they catch him, they can interrogate him. He knows that he was paid to plant the forged book on them by enemy #1, but he is scared of enemy #1, and at the same time, sees an opportunity to profit from the situation; he will name enemy #3 as responsible, thinking that he can extort enemy #2 for not sending the PCs after him. If the PCs are any good at their jobs, they will immediately find the Thief to be an unreliable source of information, and not fall for his attempted deception. There are four possible outcomes from this encounter: the PCs can be steered back to enemy #3 (square e), they can be misled by the thief into holding enemy #2 responsible (square h), they can decide to get the book deciphered independently because they mistrust the thief (square d) or they can force the thief to disclose enemy #1 as the manipulator (square g).
- g-h: One enemy neutralized. There are two antagonists, and we need two entries, so this is pretty straightforward. These squares almost always require a plot twist, unless the PCs are arriving at one of these squares directly from squares b or c. Coming from b means that the PCs have either fallen for the manipulations by enemy #1 hook line and sinker (h) and attacked what is ultimately the wrong individual from their point of view, or they have figured out that they are being manipulated and decide to confront their seeming-enemy (h) – so the nature of the confrontation will change, but not the identity of the person being confronted. Coming from (c) means that the party discovered enemy #1’s attempted deception and have decided to confront him directly without investigating via (e). That in turn means that he would be able to attempt to bluff his way out of it by claiming that enemy #2 is the one attempting to manipulate them, and that they have fallen for it! While this might seem like a plot twist, it’s actually enemy #1 doing more of the same, i.e. attempting to manipulate the PCs. But it’s far more likely that these squares will be reached by way of one of d, e, or f.
- g: The PCs confront enemy #2. The nature of the confrontation depends on which square they are coming from, as described above. In the course of the encounter, the hostility between enemy #1 and #2 will be revealed/recapitulated. Either the PCs will succeed in having enemy #2 removed from power (just as enemy #1 wanted) or in recruiting him as an ally against enemy #1. If the former, they will discover enemy #2’s real “little black book”, revealing that the one they have is a forgery, and pointing the finger at their having been used by enemy #1. Either way, this leads to the conclusion.
- h: The PCs confront enemy #1. Enemy #1 can attempt to bluff, and may or may not succeed, depending on how the PCs reached this position. If the bluff fails, enemy #1 will come clean, while still pointing the finger at enemy #2 as being corrupt.
- i: Revelation/Discovery, Confrontation, Aftermath. There are three ways to get here – from e, from g, or from h. Ideally, the route will be g-i, making the confrontation with enemy #1. If the path is h-i, this will be a confrontation with enemy #2. The GM should not permit the e-i route, instead following the e-h path and reserving the conclusion to the plotline as being an encounter with enemy #2.
- i: A final plot twist is required – enemy #3 stands revealed as having manipulated enemy #1 throughout in a bid to remove both enemy #1 and enemy #2. He can either be a good guy or worse than either of them – they may have been corrupt, but they stood in his way of implementing the harsh totalitarian rule that he demanded. He will then declare the PCs persona-non-gratis throughout the Kingdom, but give them 24 hours grace before coming after them as recompense ‘for services rendered’. This either leads to the real ultimate conclusion (as the PCs take down enemy #3, who started gloating a little too soon, thinking that his power and position would protect him, or to a sequel adventure in which the PCs work to bring about the overthrow of enemy #3, or simply to a change of circumstances in which enemy #3 has become the power behind the throne in this Kingdom. That’s up to the PCs, but most of the time I would expect one of the first two options to be their reaction. Or, if enemy #2 became allied to the PCs, that’s an entirely different outcome – they’ve done a deal with “a” devil which will no doubt come back to haunt them at some point in the future.
Ripples In Harmony
This plotting technique is a lot more work than writing a straightforward plotline, but it preserves complete independence of choice on the part of the players. There are multiple paths through to the ultimate outcome. Each step of each path creates a knock-on effect on the next step in the path, changing its context and meaning. Each step can be said to create a ripple through the plot structure, changing the shape of events to come so that they always lead to the plot conclusion. The great advantage is that, like a choose-your-own-adventure book, the same plot could be run with the same players a second time and would reach a conclusion that could be quite different – or exactly the same, but by a very different path.
Expanding the plotline
This all stems from one seemingly-random event. Adding more cells enables a bigger tapestry to be woven. The obvious place to start is with enemy #2, who is a very passive element in this outline. That’s the inevitable result of making enemy #1 the antagonist and enemy #2 the intended victim – and the PCs, the patsies.
To give enemy #2 his own scheme to bring down enemy #1, you could eliminate square (a) (moving its content to square b) and permitting two almost-simultaneous plots to be in motion at the same time. But that would be too complex a plot structure to fit our simple 3×3 squares. The minimum would be a 4×4 structure with the uppermost cell removed. You would then have a three-corridor overall structure in which the left represented enemy #1, the right represented enemy #2, and the middle contained the red herrings and the potential for the other two enemies to interact with each other’s schemes.
The obvious plot hole
There is one obvious plot hole that needs to be addressed: what if the players decide to split up and pursue all their leads at the same time?
The answer: let them. (d) says that enemy #1 is behind events, (e) says that enemy #2 is responsible, while (f) points the finger at enemy #3 (but in an unconvincing way). The players still have to make the choice of confronting enemy #1 or enemy #2 (confronting enemy #3 leads to e, which has already been addressed).
Of course, the players might get clever and try to lure enemies #1 and #2 into a direct confrontation with each other, bypassing both g and h, under this circumstance – which is perfectly acceptable as an outcome. And it would even succeed in unmasking the real culprit – enemy #3.
The other obvious plot hole
Finally, what if the PCs decide to do something other than (d), (e), or (f)? They might choose to simply get out of town and let the mess sort itself out. They might decide to take their suspicions to the watch, or to a figure of authority. If this happens, both watch and authority would demand proof – deciphering the code book would be necessary evidence. That leaves the players back with choices d or e, and the plotline back on track. Only if the PCs choose to opt out of the adventure altogether does this plotline not reach a satisfactory conclusion.
The Pattern Of Raindrops
By isolating plot developments in the manner described while keeping a structured relationship between them that is dictated by the overall plot, we create a patternless noise akin to the sound of individual raindrops. But listen closely, and this technique will yield the underlying pattern of the plotline.
To wrap things up, here’s a rotated chessboard for you all to use. Click on the image to open it in a new window, suitable for printing.
One final suggestion: if you summarize your ideas very concisely, you could write them on post-it notes and do your plotting on an actual chessboard. This would cut out all the hassle of using code letters to link your notes with the layout and make the process a lot easier.