DUST - Truth Under Decay

Inspiration

Misinformation spreads rapidly when people are overwhelmed with content and short on time. Most media-literacy tools are static or lecture-based, while misinformation itself is dynamic, emotional, and urgent.

We wanted to build something that mirrors the real problem: time pressure, uncertainty, and imperfect decision-making. That idea became DUST.


What It Does

DUST is a real-time fact-checking game where players inspect decaying web pages and archive what they believe is true before the page collapses.

Game Modes

Solo Mode

  • Progressive difficulty
  • Faster decay over time
  • Increasingly subtle misinformation

Multiplayer Race

  • Head-to-head timed rounds
  • Competitive scoring

Co-op Mode

  • Shared time pressure
  • Combined team score
  • Separate leaderboard from solo play

Scoped Leaderboards

  • Solo leaderboards are separated by level.
  • Co-op has its own leaderboard.

Core Scoring System

** Score = ∑ (correct picks × combo multiplier) + clutch bonuses − penalties **

  • Combo multipliers reward accuracy streaks.
  • Clutch bonuses reward strong late-game decisions.
  • Penalties discourage reckless or random selections.

Core Game Systems

Decay Engine
Implements time pressure and progressive difficulty scaling.

Archive / Reveal Loop
Players select content, reveal correctness, and receive scoring in a tight loop.

Multiplayer Presence

  • Reconnect handling
  • Default wins on disconnect
  • Rematch logic

Challenges

  • Keeping state transitions deterministic across playing, revealing, and gameover phases
  • Preventing leaderboard inconsistencies across game modes
  • Handling multiplayer edge cases such as disconnects and rematches
  • Balancing difficulty so later levels feel challenging but fair

Accomplishments

  • Delivered a polished real-time full-stack game under hackathon constraints
  • Implemented mode-aware leaderboard architecture
  • Built a save-status UX that reflects true backend completion
  • Successfully shipped both competitive and collaborative multiplayer modes

Lessons Learned

  • Data modeling decisions (such as leaderboard scope keys) directly impact correctness.
  • Explicit and idempotent mutations simplify real-time system evolution.
  • Spec-driven development reduces regressions and improves shipping speed.
  • Small UX details around persistence and feedback build user trust.

What’s Next

  • Seasonal ladders and weekly challenge boards
  • Improved anti-cheat validation for multiplayer scoring
  • Expanded content generation and moderation tools
  • Team mode, tournaments, and richer post-match analytics

Built With

  • clerk
  • convex
  • nextjs
  • shadcnui
  • tailwindcss
  • turborepo
  • zustand
Share this project:

Updates