Dice Adventure
HCII Research| Summer 2022
Summer 2022 I was hired as a game designer by the ARL STRONG lab at CMU, to design and develop an asymmetric co-op game for the purpose of Human-Machine Teaming studies.
Dice Adventure is the final outcome we developed, it is a turn-based game involving three characters: a dwarf, a human and a giant. They each have their own range of vision and their own set of dice, and they will use their dices to battle and break through obstacles. Their final goal is to unlock and go through the final gate, and to do that they have to find their altars and activate them.
In Dice Adventure, players need to be communicating through a pinning system, and they need to collaborate by sharing information and using their dices wisely. Such mechanic enables researcher from our collaborating research group to better create AI agents and test how they collaborate with human players.
Deliverables
Time Span
Team
A multiplayer asymmetric
co-op game designed for human-machine teaming studies
6/21/2022 - present
Yutao Huang
Pheobe Wang
Jingyuan Fang
My Impact
01
Researched over existing co-op games and board games and proposed several game concepts to the team.
02
Communicate with collaborating research team to understand their research needs and transform them into in-game features.
03
Worked closely with the artist and programmer in the team. Managed the progress of the project
04
Finished all the level designs and tutorial stages.
05
Arranged playtest, collected feedback and listed out existing bugs.
Research
We started by looking into some existing co-op games and sorted them into categories and possible directions that we can take.
Constraints
During the research phase, we also met the collaborating research team and collected some of their needs and thoughts.
In-game Communication
Since audio communication can be hard for AI player, we need some kind of built-in communication system that doesn't require audio input.
In-game Learning
Since the research team want the human-machine teaming AIs to have a different learning pattern that allows AI players to learn real-time, we need to give breaks where players can communicate and reflect.
Strategy and Decision Making
The game should be heavily based on strategy and decision making instead of something like reflects and aim.
Asymmetry in Information
We want the asymmetry of the game to be mostly in how players gather information. So that we can push them to communicate more with each other.
Turn-Based
With the feedback and ideas of the team, we found turn-based games fit the requirement the most,
We then did some more research over board games as a lot of them are turn based, and their mechanisms are normally more manageable for AIs.
After we had a clear understanding of what the team wants, we started to pitch some ideas and make paper prototypes.
Ideation
Once we had a satisfying result, we slowly moved towards one direction.
Final Concept
Characters
In the final concept, there are three characters, the dwarf, the human and the giant, the taller they are, they can see further on the map.
The three different characters also have different dices, this will give them unequal strength when they are facing a specific kind of obstacles.
Encounters
There are three different kinds of obstacles in the game, monster, trap and rocks, players have to go through them by throwing the correct number out of their dice.
Goal
The three players each have a matching altar, they each have to reach their own altar to unlock the final gate. After they each reach the final gate, they can proceed to the next stage.
Pinning System
The players need to communicate using the in-game pinning system.
08/12 Demo
Plan and Move
A new moving system to enable better learning process for the AI agents.
UX Updates
Some UX updates to enhance the player's experience.
Level Updates
New level Designs for the game including the tutorial levels.
