The great answer

(TL;DR : In most games you open a door with a key ; in this game you open a door with any tiny metallic object no matter what the actual object is.)

During the first months of 2017 at Swing Swing Submarine we have been working on a few prototypes for possible games.
In short, we couldn't fund any of them and had to stop, but it still remains a few prototypes that I think carries their lot of interesting concepts.
Here I'm going to talk about a project called "The Great Answer" (aka TGA).
The team included William David, Clément Duquesne, Simon Trousselier and myself Yann van der Cruyssen.
The prototype we released wasn't really the typical "vertical slice" with a precise challenge but rather a sandbox for us to try out random ideas and check how fun or technically feasible they were.
In this document I'm going to focus on what I personally worked and cared the most about, while other members of the team worked on other aspects of the project.

Narrative background

The game takes place in a out-of-time universe that mixes sci-fi and mythical elements.
Each game session starts with archeologists (including historians, scientists, etc) gathered around a table, they have found a number of ancient runes that seem to tell a story.
A sure thing is that the story is about a traveler called Adama, but no ones agree on the exact details of the translation. One of the archeologists speaks to give his or her own version of the story.
In each game session (of about one hour) the player will embody a specific "version" of Adama as told by one of the archeologists.
So, the narrative takes place on two layers : the "intra" part that is about Adama and the "meta" part that is about the archeologists arguing with each other.


The players plays as Adama, the controls use only a mouse and the gameplay is mainly about :
- traveling on a map
- reading text making text-based choices
- choosing items in an inventory
- occasionally drag-and-dropping words out of texts

Generative design

We wanted each player to feel like playing their own way, experiencing very unique game sessions, and try things that potentially no other player had tried before.
That led us to use traditional rogue-like techniques (procedural map, encounters, random items etc) but also use some kind of generative storytelling.

Although the general rules of the game say consistent, each session starts by picking random starting conditions (i.e. items at start) goal (i.e. crafting a certain item or finding a specific character) and modifiers (i.e. in this session, the weather will always be extremely cold).
We wanted each game session to last about one hour (at some point the translators would run out of runs and the game would stop at the end of the day).

Defining items by their attributes

Suppose this situation : somebody requires the character's help to cure a disease, the player can help by giving a medicine and get a reward in return.
In many adventure games the "medicine" will be an item with a specific ID and if the players has or selects it, the "puzzle" is solved.
In our case the game almost never checks the actual ID of an object but rather it's attributes, for instance : "is it edible and healthy enough ? Then the puzzle is solved".
This allows for multiple solutions and outcomes out of a single game situation (for instance if the player gives a medicine that also have some magical properties, the disease could become an entity on their own and speak, so that the story goes on in an unexpected way).

The prototype about 150 items and we manually filled forms about their attributes (there were about 50 usable attributes at the time of the prototype).
We've been using ink with a lot oh additional code to express sub-collections of items by conditions.
Attributes include boolean check-boxes (is it magical ? alive ? made of wood ? beautiful ?) numbers (price, weight, size) and more complex values (color, name).
Yes the beauty in this game isn't subjective but made of arbitrary hidden hard-coded values based on my own taste.

Still we didn't want to flood the player with many useless possibilities in the case of a large inventory.
For a given situation, the game would first check if anything interesting from the inventory could be used, if not it wouldn't even let the player try.
But if something could be use, then the game would only show a sub-selection from the inventory including useful items but also a few random intruders.
For instance even if the inventory contains 20 items, the game will say : "Do you want to cure the disease with incense, magic potion, potatoes, or a screwdriver ?".

Another good example of the use of attributes is when the weather gets too cold.
The player would either risk catching a disease or sacrifice something made of wood to build a fire.
Typical wooden items include : a shovel, material to repair the carriage, pencil... Burning an useful item isn't a good idea, but burning a small item would not make a good fire.

It was possible for the player to modify the attribute of an existing object mainly through shops (specialized characters could be asked to modify liquids, metals, etc by giving them flavor, making them magnetic etc).
We designed and experimented with item attributes for a while, but the plan was to also use attributes to define places and characters.

Taking many things into account

That's one of the things I realized by playing "King of Dragon pass" : when you take enough things into account to determine the outcome of an action (the reputation of the player, beliefs of a given character, color of an item, place where something happened, historical background details...) without explicitly telling the player, they will eventually notice it, take everything into account themselves just in case the game does and the game will feel more like roleplaying.
TGA has many hidden variables (about health, reputation, karma, etc).
Every day the game will display an horoscope that is also made of hidden variables that have an influence on the outcome of actions (should you do something involving water today ? something bold ? should you sell or buy something ? etc).

Those horoscope variables vary smoothly and only the most extreme ones are told to the player in the form of a short sentence (such as "you should really eat something today").
The game also feature a number of gods (gods of burglars, love, wealth, etc) that each have their specific opinion about the play but never explicitly unveil themselves to the character.
The player can also modify those variables by themselves for instance by making specific sacrifices.
For the prototype we decided to keep all of these variables hidden which would have been too obscure for a released game.


Since I'm the kind of player who goes straight South if the character in the game says "I should go North" I wanted not to ignore players who might act like that.
Since the story of the game involves a hypothesis about the translation of the actual story (the translators sometimes put the game on hold to discuss the consistency of the translation, that is to say the players decisions), it allowed for inconsistencies and I wanted to take advantages of that.
We tried to reward the player for experimenting.
A few elements of the story are present in both the "intra" and the "meta" part and create a link between those : Adama's pet (an owl), the runes and a special hourglass.
In the future time (meta) the runes are being translated by the archeologists, and in the past time (intra) the runes are being used as sand in the hourglass.
Each time something significant happens in the game (which doesn't necessarily means "good" for the character) a needle moves to one side ; and each time something inconsistent happens, it moves the other way.

At the end of each of Adama's days one rune falls either to the "part of the story" or "out of the story" side. The position of the needle has an influence of it but since this process is based on physical simulation, there is often some uncertainty.
Enough "good" runes are necessary for the translation to be approved by the other historians (on the meta part) and for Adama to meet one god (on the intra part).
But apart from that, if one player decides to act like a complete loser within the game, he's allowed to and, with enough runes, will probably get Adama to meet some loser god in the end.
Each time the player meets and unknown god, historians gather a new unknown rune that allows for longer stories and more chance for the player to complete weird achievements.

Altering the story

Another core mechanic that the prototype doesn't show much is the possibility for the two layers of story to interfere with each other.
This is done through the use of an owl. Every time the owl shows up, the player is allowed to click on it and enter an alternative gameplay phase.
In this gameplay phase the player will only be allowed to use small stones (actually those are the runes but I'll call them stones here to avoid gameplay-wise confusion) representing concepts (death, travel, night, etc) that had been collected during the game.
By drag-and-dropping stones over a word in a text (for instance something that a character just said) the word will be replaced by another word and the story will change.
For example dropping "death" over a text saying "I want to eat" will turn it into "I want to die", and the situation changes on the fly and the game goes on. The player is welcome to experiment a lot with that.
Those stones can also be turned upside-down to change their meaning (i.e. death -> life).

Occasionally, the player will meet rune carvers that can create stones in exchange of regular items (for instance giving them a living being might result in a "life" stone).
We imagined some uses for about 15 of those stones.

Merging items

One of the random encounters I liked the most even though it was often pointless was this robot who called itself the "transformulator".
That happens once in the video around 53 minutes from start.
Basically you could give it two items and it would merge them into one and make a little comment.
The resulting item will have both of the attributes of the "parent" items or a mean.

For instance if you give it a cow and a shirt, it will create a new unseen item called "cow-shirt" that is both a piece of cloth and a living animal, you'll be able to use it in any upcoming situations just as a regular item.


The game contained many "simple" situations (encounters or places) that were designed to be replayable in many ways but each game session would also feature a "largest" narrative arc involving recurring characters, those were more linear but more detailed.
I always try to hide enough simple "secrets" everywhere in a game for any player to eventually find some without clues rather than have just a few impressive secrets and guide the players to them.
Since those secrets were mainly text-based and easy to add, I didn't care if some were never found by any player during playtests.
Each time a player tried something (i.e use an item in a specific situation) that they think is supposed to result in a meaningful outcome but doesn't, I try whenever possible to add that outcome to the game rather than discourage players to act that way.

During the beginning of 2017 we have also made two other similar projects at Swing Swing Submarine, one was like a more traditional point-and-click game but with alternative initial conditions and paths, the other was a four-players narrative game also about traveling.