Pontus Joneström

The character is an artificially intelligent washing machine which only goal is to find all dirty laundry. To add to the playfulness it was important to give it character. The controls and the mechanics play their part. We also gave it a happy yellow color and used some stretch and squeeze animations and bubbly VFX to give it more personality.

To make it feel playful, I wanted the washing machine to be shaky, clumsy and random, but still controllable. This was achieved by the programmers by adding random forces to the washing machine. In the first playable, it was very hard to control so the forces was reduced to improve mobility, but still enough to stand out visually.

We decided to go with a third person camera to allow for the player to really see all the mess they make accidentally as they move around trying to find clothes in the apartment. We also gave the player control to zoom in and out.

As the movement was made more controllable, we wanted something else that could cause more chaos in the game according to the pillars. We also needed a way to pick up clothes. We combined these requirements with the suction mechanic. The player need to use this ability to get the dirty laundry and while using it, they will also move around other items in the vicinity.

When the player uses the suction mechanic, they will also get smaller items. After a few items, they start to move slower. To get rid of the items they have to shoot them out, causing more chaos in the process. To teach this mechanic we made doors in the apartment that are only destroyed by shooting items at them. The room where the player starts acts as tutorial where they have to shoot the door to continue. To add to the playfulness, the player also gets pushed back when shooting.

We started out by brainstorming ideas. The first step was to let everybody in the team come up with as many ideas as they could. After that we talked about them and explored them.

Eventually we settled on three ideas that we explored even further and tried to figure out how they would work in practice. It was pretty clear after this that everybody liked the Washing Machine idea, so we went with that.

The first prototype showed promise and early playtesters enjoyed the concept. They felt however, that the movement was rough.

Pretty soon after the first playtest, we reduced the random forces added to the machine to make it more controllable.

By the time of Alpha we had implemented the suction mechanic and ability to pick up clothes.

The level design blockout for the alpha build.

For the Beta we now had most of the physics interactions implemented. The UI was partly implemented but we were still working on getting the HUD right. Some lighting had been put in.

Laundry under the couch pillow. At this point we tried to find as many fun locations for the clothes as possible. We wanted locations that would both provide a challenge in finding, but also cause a lot of chaos when picked up.

For the final game we added polish in the form of lighting, VFX, animations, SFX and some shader work to tie it together.

Organized the team

The team worked remotely and most of the team members had never met before. Because of this it was important to foster a team spirit and active collaboration.

The movement speed is fast and the player has a high level of control. To achieve this, we made the player fall down faster than they jump up. We also gave the player high level of air-control and the ability to jump higher by holding down the jump button.

The player has a dash ability that allows them to dash in the direction they are looking. The cooldown for this is very fast, so it feels practically unlimited. This was important to let the player move quickly between attacked hearts. To add to the game feel and clarity, I made an effect for the dash that changed the FOV of the camera and spawned speed lines during the dash.

The player can also glide by holding down the jump button. If the player is falling, they will slow down and start gliding. There is no limit on how long the player can glide, which also serves the purpose of giving the player high mobility when multiple hearts are attacked.

For the initial playtesting I prototyped a simple Wave System with Visual Scripting. This version had only spawning enemies, but no progression and very simple elements of randomness. The actual Wave System was done in C++ and was much more complicated, so we needed something quick to be able to playtest the game properly.

For the real Wave System, one of the programmers was assigned to develop it. To get started we had a meeting where we wrote down every step of a wave. This could then be used by all team members when implementing some part of the system, like dialogue triggers for example. The system was then continuously playtested and adjusted during the rest of the project.

For the real Wave System, one of the programmers was assigned to develop it. To get started we had a meeting where we wrote down every step of a wave. This could then be used by all team members when implementing some part of the system, like dialogue triggers for example. The system was then continuously playtested and adjusted during the rest of the project.

For the hearts we split it up into two stages of danger. The first being when a heart is included in a wave. A sound is played and a dialogue pop-up appears telling the player what hearts are under attack. The hearts also light up and a smoke effect is played at their base.

The second stage is when a heart is in immediate danger, which means enemies are close to it. A dialogue pop-up appears, a sound is played and the hearts starts to bleed and turn to a deeper red.

First playable

First playable showing simple AI and shooting.

Pre Alpha

Pre Alpha build where AI now targets the player when player shoots at them. Gliding and moon gravity jumping was also implemented.

Alpha

By the Alpha build, we had implemented a Blueprint version of the coming spawning system, which proved difficult to code. A health system was in place. Movement was also tweaked further and a dash was implemented.

Pre Beta

For the Pre Beta an animation system was added to the enemies. The AI pathfinding was also improved and some VFX was implemented. A score system was also in place.

Beta

For the Beta, all features was now in place. Dialogue boxes with important information was now implemented. Still, players found it difficult to know where to go and which hearts where in danger. They also found the astronauts hard to see.

Gold Master

Beyond polishing the game with VFX, lighting, post-processing and audio, we also made some adjustments. The astronaut color was changed to yellow and a location prompt was added to tell the player what heart they where close to at the moment.

The main focus was to increase efficiency by making sure visitors would easily find the way through the houses. This was done by changing the layouts in some crucial places where we had noticed previous years that visitors would often get confused or slow down. I also removed blockers like doors and drapes. In some places light was also used to guide visitors in the right direction. For quality, one of the haunted houses was also redesigned from the ground up with new rooms and scares.

Clearly defined goals

New goals for visitors per hour was set for each haunted house. These were calculated using earlier years’ data by investigating which of the haunted houses had high efficiency and identifying the reasons for this. The goals were then communicated to everybody working on the haunted houses, from management to actors.

App for real time statistics

To make sure efficiency could be measured easily in real time, I commissioned an app that was created by the IT department. The app was designed to be easy to use for staff members and management. It was also designed to remind the staff of the short term goals continuously.