Monday 10 June 2013

The Problem With Gamebooks Trilogy - Part 5

Fixing Gamebook Plots

In the previous post in this series I suggested a potential solution to some of the relational issues between reader and character in interactive fiction.  But fixing characters is only half a solution. The other half requires us to address another gamebook problem area, plot.

I will avoid talking about plot in terms of creative writing, there are far better sources for that. Rather I will focus on how the gamebook's strict branching tree format needs to evolve to facilitate better interactive storytelling.

Gamebooks plots are a string of set-pieces.

Simple gamebook plots are often a sequence of interconnected set-pieces. The player moves from scene to scene, the designers assuming that ‘kill the wizard’  or ‘progress through the map’ is a robust enough premise to glue the narrative together.

Although fun, these simplistic romps lack genuine causality. They defy the definition of plot: a causal sequence of events that make up a story. As E.M. Forster explains instead of "The king died and then the queen died" it's better to create a causal relationship between the two events, for example "The king died and then the queen died of grief." Events need to unfold in a logical chain, to create narrative momentum, and pull the reader along with them.

Gamebooks and causality

Gamebook ‘causality’ is often confined to a single choice or small scene. There is no sense that your actions in one scene have any causal relationship with subsequent scenes and events.

The limitations of the print medium have made far-reaching causal effects in gamebooks impractical. To do ‘causal’ properly a gamebook would need store a player’s previous choices and apply them to the following scenes. With the existing format this would clearly be an onerous task for both the player (who would have to record or remember his previous choices) and for the designer (who would need to write a multitude of scene variations to accommodate the players previous choices). But without tying scenes together we don't have plot, just a sequence of incidents.

Causality is required to make a series of scenes a plot. 

But this is the internet age, the constraints of analog gamebooks no longer concern us. Digital gamebooks can save our progress, remember our choices and record variables beyond the usual stamina and luck. However, the problem remains that although a gamebook can remember, it would still be left to the designer to generate the broad range of narrative possibilities. Although we are no longer restricted by page or word counts, it's not a practical solution, it just makes the writing of a gamebook an increasingly colossal task.

With a simple game engine we could let the processor do some of the work for us.

Liquid narrative

Current gamebooks, with their print origins, are in a solid narrative state. That is, all the text exists in pre-written chunks. What is written is set in stone, it's solid. This requires the game designer to anticipate and pre-generate all the content.

Solid narrative. Scenes are self-contained, predefined statements. 

Gamebooks in a liquid narrative state, a possibility for digital gamebooks, would allow the narrative to be generated from a set of smaller story fragments. Gamebooks are already broken into paragraphs - why not break them down further? By increasing the granularity of our story fragments, reducing them to sentences and words, the gamebook could be a more personalised experience and enable the potential to engineer causal links for each play through.

A liquid narrative would not store scenes in a rigid finished form, but assemble them based on a simple set of rules (I'm not talking about a complex IF engine here!) Although the designer would still need to consider the permutations resulting from a players choice history, he would not have to construct complete paragraphs, the engine would do this bit.

With liquid narrative a user's previous decisions would 'colour' subsequent scenes and available choices.

Procedural scene generation

So our player's prior choices have been recorded, variables and flags are stored. The data is then parsed and a scene is assembled by the story engine. But how does this engine decide which fragments to use and in what order?

Well, I'm not quite sure yet... but here are some initial thoughts:

1. Physical fragments
These are used to describe the physical setting, characters and objects.

2. Emotional fragments
These are used to manage rising/falling tension. To control the push and pull between hope and fear (see Robin D Laws Hamlet's Hit Points).

3. Relational fragments
These are used to reflect your previous interactions with NPCs.

4. Thematic fragments
These are used to present alternative viewpoints to the one being expressed by the player (though their choices) to create a tension or uncertainty.

This all needs much more cooking!

Physical, relational, emotional and thematic fragments come together to form a completed description.

Writing a story in this object orientated way will be challenge. Can these fragments flow seamlessly together so they that they are a pleasure to read? How will you track all the permutations and sense check them? Methinks some authoring tools will be required.


To clarify, I'm not advocating dumping the branching tree structure at this stage, just enhancing it. I believe gamebooks are a progressive, essentially linear, experience and controlling the core story flow should remain in the hands of the author, to let him deliver his message intact. What liquid narrative allows is: less binary outcomes to choices, the ability to have far reaching consequences to decisions and the ability to engineer causal links between separated scenes.

++UPDATE++ Both Twine and Inkle provide the functionality to create the concepts described in this post. The question now is how do we use this power to create more engaging gamebook experiences? 

Next some thoughts on minimizing the harm interaction has on the suspension of disbelief...

Read part six now

Further reading

Weapons Master

The Weapon Master has a penchant for exotic melee weapons and small furry mammals.

Follow the Mysterious Path on Twitter or Google+

What is this Mysterious Path?
Mysterious Path is half comic, half 8bit RPG, half choose-you-own-adventure. Mysterious Path will be an interactive experience playable on your phone, tablet or desktop... eventually. Imagined by the one man army that is Grey Wizard (and some occasionally helpful retainers)


  1. Now this is an exciting concept that allows players to have much more choice of their character's outlook. A lot of the time, a gamebook character's outlook is forced upon the player due to their background or how the gamebook rewards and punishes certain actions (for example, you lose luck in City of Thieves if you kill the blacksmith because he was a good man, implying that whatever choices you might make, you still have a conscience, even if you spend the rest of the book breaking into peoples' homes, killing them and stealing their stuff).

    The physical fragments can also give a lot of atmosphere - lead a successful invasion against the enemy's capital? Next time you go to the palace, it can be full of ornaments from there.

    I am also excited about the relationship fragments. For example, if you save village x from the dragon and move onto village y, you can have a fragment where the people there mention your great deed. This will make the gamebook feel more like a living, breathing, interconnected place, rather than a series of set pieces. In fact, that would do wonders for something like Fabled Lands app, although I would not like the unenviable task of thinking of 12 books worth of fragments for every encounter.

    Fragments will give gamebooks a lot of scope to make the world more alive and also increase immersion because the book will have fewer instances of 'If you have codeword/item x, turn to this page.' Instead of having items and codewords purely as markers of what you ahve done, the book will remember for you.

    I think Dave Morris's Frankenstein utilises this as previous choices affect character's actions in the future a little (although not to the extent that the monster can find friendship or stop Victor from destroying his mate.)

    1. Yeah, I'm not sure any of this is new Stuart!

      I recon Telltale's Walking Dead game must use a similar system. In that you can't change the outcome of the final scene but your participation certainly flavours the experience, increasing its potency.

      For me this is the sweet spot for gamebooks. A solid narrative spine (or message) crafted by a skilled storyteller, but with enough fluidity to allow the player to explore the themes and personalise the experience.

      PS. It's great to hear your ideas on how the 'fragments' might be employed!

  2. Great discussion, thank you for sharing.

    There's definitely a balance in print books between asking the reader to record information to allow repercussions to ripple through the rest of the book and overburdening them with recording too much.

    Digital gives you much greater tools and, perhaps more importantly, does so seamlessly. Your discussion of 'fragments' is great and much food for thought. As Stuart pointed out, Dave Morris' Frankenstein employs this, adapting the text and choices available based on your previous choices and relationship with characters.

    If you haven't done so, definitely check out the tutorials; the ability to not only flag keywords, but track counters, allows you to create 'fragments' that could relate to not just health and magic, but alignment, relationships between individuals and factions, renown, sanity, time, etc. This can then be used to open up or restrict future choices the reader can make, but also change text within a paragraph, such as changing the time of day or season - making a pass that is easy to travel by day in summer a real danger by night in winter for example.

    That is of course requires a lot of work. Closer to planning a computer game than traditional print gamebooks, but the possibilities are there.

    1. Thanks for the heads up Matt, seems like I've just described Inkle logic!

      Be interesting to see how Versu authoring tools will work too... when they are available.

  3. This is one of the issues that inspired the work I've been doing on the game engine for Dwarf King/Pirates and Traders 2 with personality traits.

    As mentioned, the PC is defined with personality traits - traits which the player can "select" in essence, through his/her actions. Do cruel deeds, and you'll pick up the cruel trait. However, not only the player has such traits - all characters are defined by personality traits.

    In any given encounter, then, there are multiple factors that can be used to affect the story. If the PC is interacting with a character with an identical personality, the results would (should) tend to be very different than if he is interacting with a character with opposite personality.

    Even more interesting - IMO - is that using personality traits to drive the story in this way makes it possible to add even more variety to the game narrative. E.g., by randomly assigning personality traits to the NPCs, it is (theoretically) possible to vary the dynamics of every playthrough. Player choices that worked perfectly to defeat the Arrogant, Monologuing evil wizard might not be as successful against the Conscientious, Careful evil wizard you meet in the next playthrough.

    This essentially is the same as the fragments approach defined above - but as you've no doubt discovered by now, I am a big believer in the idea of character-driven gameplay.

  4. With Single Deck, Double Deck Pitch and 6-Deck Shoes may be} guaranteed to discover a game of your liking. Hands down, Blackjack is the most well-liked and greatest known desk game. The object of 토토사이트 Blackjack is to get as near 21 as attainable with out going over.