EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·Jul 24·5 min read
My football predictor scores 0.203 vs. the bookies' 0.198 – and loses
TL;DR
A football prediction model using Elo ratings and a Poisson distribution achieved a score of 0.203, slightly outperforming bookmakers at 0.198, but ultimately failed to predict match outcomes accurately.
✦ Why It Matters
Consider refining your predictive models by integrating Elo ratings and Poisson distributions for better accuracy.
Key Takeaways
- Utilize Elo ratings to assess team strength effectively.
- Apply Poisson distribution for calculating scoreline probabilities.
- Incorporate recent form for more accurate predictions.
- Understand the model's limitations regarding external factors.
How It Works
The model starts with Elo ratings to gauge team strength, converting rating gaps into expected goals. It then uses a Poisson distribution to estimate scoreline probabilities, applying the Dixon-Coles correction for low-scoring outcomes.
This structured approach allows for a comprehensive prediction of match results.