berliwu/world-cup-2026-match-prediction-engine — explained in plain English
Analysis updated 2026-05-18
Predict win, draw, or loss probabilities for a specific World Cup fixture from the terminal.
Simulate a full group stage or run a Monte Carlo bracket simulation for the tournament.
Predict club match outcomes, such as Arsenal vs Chelsea, using historical CSV data and a Poisson model.
| berliwu/world-cup-2026-match-prediction-engine | sunapp-ai/sun-to-spotify | 1061700625/github_vps | |
|---|---|---|---|
| Stars | 37 | 37 | 38 |
| Language | Shell | Shell | Shell |
| Setup difficulty | moderate | moderate | moderate |
| Complexity | 3/5 | 2/5 | 2/5 |
| Audience | developer | vibe coder | ops devops |
Figures from each repo's GitHub metadata at analysis time.
Requires Node.js 20+, the optional GPT-4o-mini overlay needs an OpenAI API key, and Redis caching is optional.
This is a TypeScript command-line tool that predicts match outcomes for the FIFA 2026 World Cup. You run it from the terminal and ask it to estimate win, draw, or loss probabilities for a specific fixture, simulate a group stage, or run a Monte Carlo bracket simulation to see which team is most likely to win the whole tournament. The prediction math works in layers. Four signals are blended together: Elo ratings (a measure of how strong each national team is based on historical results, weighted at 35%), a Poisson goal model based on attack and defense rates (30%), recent form from the last several matches (20%), and squad market value as a strength proxy (15%). These calculations live in separate modules with no external API calls, so they run instantly and offline. Optionally, you can enable an AI overlay that uses GPT-4o-mini to add small adjustments on top of the statistical result. The README is specific about what this means: the AI adds bounded nudges, not a replacement for the numbers. It can comment on context that raw stats miss, such as injuries or playing conditions, and nudge the probability slightly, but the statistical anchor stays fixed. There is also a separate mode for predicting club matches. Type in two club names (Arsenal vs Chelsea, for example) and the tool loads historical match data from a remote CSV file, runs a Poisson-based calculation, and optionally asks the AI for a comment. Results come out as a JSON object showing win/draw/loss probabilities, expected goals, and a confidence score. Redis caching is optional and speeds up repeated lookups. The tool requires Node.js 20 or later, uses TypeScript throughout, and runs tests with Vitest.
This TypeScript CLI predicts FIFA 2026 World Cup match outcomes by blending Elo ratings, a Poisson goal model, recent form, and squad value, with an optional GPT-4o-mini overlay for small adjustments.
Mainly Shell. The stack also includes TypeScript, Node.js, Redis.
No license is stated in the README.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.