Obaseki
Resume
Web apps

GafferScore

A Premier League prediction game. Players predict each week's scores, earn points and climb the leaderboards.

Live, with an AI agent that analyses every fixture.

My part
Founding full stack engineer: scoring logic, leaderboards and frontend.
Stack
Next.jsTypeScriptPostgreSQL

The problem

Football fans predict results every week anyway. GafferScore turns that into a game, with points for getting it right and leaderboards to settle the arguments.

What I built

GafferScore is a Premier League prediction game. Players predict the result and exact score of every fixture before kickoff, earn points for accuracy, and climb weekly and overall leaderboards. As founding full stack engineer, I built the platform: the scoring logic, the leaderboards, the prediction deadlines and the frontend.

  1. Predict result and exact score, before kickoff
  2. Lock predictions close at the first kickoff of the gameweek
  3. Score points for correct results and exact scores
  4. Leaderboards weekly and overall rankings
Parts I built

Key decisions

  • Leaderboards are computed in the database. Rankings come from one database function that takes the leaderboard type and the gameweek. The ranking logic lives in one place, and pages stay fast as players are added.
  • Deadlines are strict. All predictions lock at the first kickoff of the matchweek, so nobody can wait for early results before predicting.

Results

  • Live at gafferscore.xyz.
  • Later joined by the Gaffer, an AI agent that analyses every club each week.