OracleBTTS
Quantitative pricing engine for football markets.
- Role
- Founder & developer
- Timeline
- Ongoing
- Live site
- oraclebtts.com
01
The problem
Pricing football markets well requires a modelling pipeline, clean historical data and a fast feedback loop between model and market. Most people attempting it are stitching together notebooks that nobody else can run.
02
The approach
I treated the model and the product as one system: the same pipeline that backtests is the one that prices live. Everything the engine outputs is traceable back to the inputs that produced it.
03
Work to date
Automated data ingestion and cleaning across fixtures and market feeds
Probability model with versioned parameters and reproducible runs
Live pricing surface with divergence highlighting against the market
Backtesting harness with performance reporting per model version
04
Next developments
Continue reviewing data quality, model versions and reproducibility. Future changes require evaluation rather than assumptions about returns.
05
Current status
Live
A pricing engine that runs unattended and produces auditable numbers, rather than a research project that only works on one machine.
Previous project
← QI PerformaNext project
Nutrilayer →Personalised plant-based nutrition, built for practitioners.