01 / Match shape
Duration, teams & difficulty
I balanced match length, player count and word difficulty so a session stayed competitive without becoming exhausting or ending before its tactical systems mattered.
Case study / 05
Typing Soccer
For Apple Developer Academy Challenge 4, my team and I built a macOS game where typing controls the match. I prototyped the idea, refined its mechanics and balanced the many systems that turn a typing duel into football.
- Context
- Challenge 4
- Platform
- macOS
- Modes
- 1v1 + 2v2
- Gameplay
- SpriteKit
- Multiplayer
- GameKit
- My role
- Design + AI
01 / My contribution
Prototype, test, balance, repeat.
My role was to establish the game idea and refine how it felt over a full match. That meant treating every rule as part of one connected pacing system rather than tuning features in isolation.
02 / Movement economy
Running, passing & stamina
I refined movement speed, stamina pressure and passing behavior to make possession feel earned and to prevent one fast typist from flattening the team game.
03 / Match rules
Extra time & penalties
I worked through edge cases such as tied scores, extra time and penalty shootouts so the match had a complete competitive arc.
04 / Authenticity
Player attributes
I collected football player data from FC Ratings and used names, speed and stamina attributes to influence how different players behave on the pitch.
02 / Core loop
Type fast. Win the ball. Score.
The design translates familiar football moments into short competitive typing decisions. Every duel changes possession, position or scoring opportunity, so typing performance always has a visible consequence on the pitch.
01 / Kickoff
Typing duel
Every match begins with a word challenge. The fastest player wins possession and starts moving toward the opponent’s goal.
02 / Pressure
Intercept
Defenders can trigger another duel to challenge the ball carrier and reverse momentum.
03 / Tactics
Build the attack
Passing, formations, positioning and stamina create decisions beyond raw words-per-minute speed.
04 / Finish
Beat the keeper
A final goalkeeper challenge converts the attack into a shot, a save or a goal.
03 / AI opponent
An opponent that makes human mistakes.
I developed the AI opponent around believable variation rather than perfect execution. Its timing and decisions create competitive pressure while leaving room for recovery, surprise and different skill levels.
01 / Typing model
Skill, length & variation
The AI’s typing time is calculated from word length and skill level, then varied with randomness, simulated fumbles and occasional missed shots.
02 / Football behavior
Rule-based positioning
Rule-based algorithms control movement, defending, passing and positioning, connecting the typing layer to recognizable football behavior.
03 / Resilience
Disconnect takeover
If someone disconnects during multiplayer, the AI can take over that role so the remaining players can finish the match instead of losing the entire session.
04 / On-device coaching
Feedback after the final whistle.
I explored Apple’s Foundation Models for a post-match coach. The on-device model analyzes WPM, accuracy, mistakes, duel performance, goals and fastest words to explain strengths and identify what the player can improve.
A rule-based fallback produces feedback when the model is unavailable. The feature was designed so the learning experience remains dependable instead of making generative AI a single point of failure.