top of page


CreepHunter is a turn-based card game with rogue-like, strategy and RPG elements. It was developed as my final graduation project in Digital Games. In this game inspired by Evil Dead, the player will face a series of enemies in 3 different regions using a deck of skill cards. In this game, the player evolves with each play, increasing his power and unlocking new cards and thus being able to defeat the game's big boss. This game was developed solo where I was responsible for all development processes, below are some of the main pillars worked on during development.

  • Programming in Unity 3D

  • Balance

  • System design

  • Level design

  • Test analysis

Article and pre-production

An article was developed that describes all project decision making, design  decisions and testing analysis and balancing. The game was initially designed to be a mobile game but the business model was changed to be a premium game.

Prototype - Board Game


To validate the game idea I designed an analogue version of the game with all the main mechanics and tested it with some players. During the test, players controlled their character following the game rules while I controlled the gameplay rules and the enemies following the AI rules that would be used later. The board was dynamic, allowing you to play different map configurations. This stage made it possible to test pacing, fun, difficulty, replayability, balancing and level design. My goal with this type of prototyping was to bring a different way of validating a game concept that was as quick and cheap as possible. And as a bonus, I was able to provide a lot of laughs at family and friend gatherings.

Level Design


The level design was divided into two sections, one of them was the world level, where the player could choose the path of his journey between the maps and the other was the level design within the combat arenas, where the player would face the enemies. Level design varied depending on the enemy class and levels were designed to support the widest variety of decks possible.
A grid system was created to assist in the construction of levels with the aim of easily changing the configuration of the maps during testing. 
The main elements that added the level were:

  • Player

  • Enemy

  • Obstacle (prevents movement and ranged attacks)

  • Water (prevents movement, but allows long-range attacks)

  • Chests (Adds new cards to the player's hand persistently)



Firebase - data base


The design of the cards was created to be easily understood by the player in a simple and informative way, using colors and a design similar to games of the same genre.


The game has a collection of 12 cards of 3 different types:
Movement: cards used to move the player around the map
Attack: cards used to attack enemies
Skilled: cards used to gain some temporary attribute such as shield or cure negative effects

A database was built to store the card data. This data is loaded from the database during the initial loading of the game. I used this strategy to make it possible to balance the cards after the build without the need to generate a new build of the game with each new update.

Balancing and tests


Using the same database, the enemy attributes and the character's base data were added to this database, with the same purpose, to be able to balance without having to do a new build of the project.
Another database collected all the decisions that players made in their matches, on which maps they died, which cards they used most and least during their matches and what level their characters reached.
This data was collected, analyzed and used to balance the cards, base character and enemies.
The game was made available on the platform for 15 days and gameplay data was collected. The data was added to a spreadsheet for analysis and balancing was carried out seeking better interaction, use of cards and adjustment of difficulty. Then the game was made available again for another 15 days. In the second round of tests, data was collected to verify the effectiveness of the changes made, which proved to be effective in addition to generating new data, enabling better balancing.

bottom of page