Midhun P M — Case studyWC Predict '26
- Next.js
- Appwrite
The problem
World Cup season at college means every group chat turns into a prediction contest, scored by someone in a spreadsheet who inevitably gives up by the quarterfinals. As IEEE Sahrdaya's Technical Coordinator, I wanted the branch to run something real instead.
What I built
WC Predict '26 is a prediction platform hosted on the branch site. Players sign in, bet on match outcomes, and climb a live leaderboard. Settlement is automated — when a match ends, scores update and the leaderboard reshuffles without anyone touching it.
The numbers: 58+ players, 737+ bets placed over the tournament.
Two market types kept it interesting:
- Pool markets — everyone's stake goes into a shared pot, winners split it. Rewards being right when others are wrong.
- Fixed-odds markets — you know your payout when you bet. Rewards calling upsets early.
Tech decisions
Same stack as the branch site — Next.js and Appwrite. One auth system, one deployment, no new infrastructure for a month-long event. Players were already signed into ieeesahrdaya.com; making them create another account would have halved participation.
Automated settlement over admin panels. Any system that needs a volunteer to manually score 64 matches will fail by the round of 16. Settlement runs itself when results come in.
What I learned
Betting mechanics are a UX problem before they're a math problem. The fixed-odds vs pool distinction had to be obvious from the bet slip, not explained in a rules page — most players never read one.
Also, leaderboards are engagement rocket fuel. Most of the 737 bets came from people checking their rank and deciding they needed one more prediction to catch a friend.