During the first few weeks, I created a document enabling us to follow the GAG (Game Accessibility Guidelines) by job role, allowing each team member to know in real time what remained to be done in terms of accessibility. I took the Must Haves and a few Could Haves and applied MOsCOW logic, holding individual interviews with each member of my team to discuss their approach to accessibility on the project.

This initial document was followed by task tracking on ClickUp by our producer, with more specific tasks for each GAG, allowing accessibility to be included throughout the project during each sprint for everyone.

Also, I created a document very early on in pre-production to bring together several resources that list methods, guidelines, and tools used for accessibility. I then proposed a V1 of dock to the other projects!

During the project, Ubisoft Chroma was made accessible to everyone, so I quickly learned how to use it to communicate with the two artists on the project, which allowed us to check color usage during the final weeks of the project and even include it on OBS during my playtests.

All the settings menu in DEN

Resources on general accessibility

During the production of DEN, I organized and planned three playtest sessions to verify several hypotheses. I organized everything from recruiting participants to planning and conducting the playtests in the school lab to analyzing the data and delivering the results to the team.

> The SoundTest was held in April with only the first versions of the animal sounds in the game. The purpose of this SoundTest was to check the emotion understood by our future players for each of the 8 sounds we had already produced. These sounds were divided between the 3 branches + the neutral state. In other words, four of them were concerned about their passive state and the other four were reactions to player movements. We organized two different questionnaires.
The first was for people within the school, as they had already heard about the game, the possible emotions present, etc. This could potentially create bias, so they could only respond by checking the emotion.
Another test was open to people outside the school, where participants could describe the sounds in their own words.
With nearly 40 participants combining qualitative and quantitative data, we concluded that two of the eight sounds were very well understood, which was less the case for the others. The verbatim comments helped our sound designer modify the animal sounds for the final game.

> This second session mainly allowed us to balance the appearance of the animals behaviors based on how players played the game. At the beginning, the values for the “fear” branch were nearly twice as high as the other two behaviors over a period of about 10 minutes. We therefore reduced the value that player movement adds to this behavior branch and slightly increased the other two branches. We also evaluated players' emotions using two tests that I learned in psychology: the PANAS (Positive and Negative Affect Schedule) in its shortened form and the SAM (Self-Assessment Manikin) scale. These scores showed slightly less positive emotions than in neutral conditions for 4/7 players, and 5/7 players tended to feel slightly more negative emotions compared to a neutral situation, according to the literature.
Finally, it was through this playtest that we switched to a trackball to prevent any desire to click with a mouse (clicking did nothing during the game). This also reinforces the physical installation effect and DEN experience.

> The last and third session took place at the end of June, shortly before our final presentation. It was planned quickly in order to check the latest sounds, animations, and the three added endings to verify whether or not our players understood all of these new elements. It allowed us to polish several elements, such as the duration of the new animations and the repetition of certain sounds, but also several things we hadn't anticipated with the AI's behavior, such as the fact that it “gets stuck” in a corner of the screen, blocked by the player's cursor. It was also an opportunity to test our “automatic mode” to make the game as accessible as possible and avoid the frustration of leaving the game without understanding "anything".

> One of the failures of DEN is its feedback. We tried as much as possible to make the experience as simple as possible, but from the very first playtests, the lack of feedback was noticeable. The team nevertheless wanted to persevere and try to add as little as possible. The jury's feedback was almost unanimous in terms of the incomprehension they sometimes felt while playing. However, others were very impressed by the game and its originality, while understanding its message, generally already sensitive to the subject.

> One solution that we added at the end of production was an “automatic mode" that allows players to force the appearance of a branch of their choice (curiosity, agonisty, or fear) regardless of the movements they make. It is worth confirming a pop-up explaining that this is not how the game was designed, as it explicitly reveals behaviors (something we wanted to avoid). This mode has allowed more people to understand the game and see its complexity through the animals!