What Markets Can Tell Us About Football
A future comparison source, with its own assumptions.
A future comparison source, with its own assumptions.
An information source
A future NQ research layer may compare model forecasts with market-implied probabilities. This is a proposed analytical feature. v0.1 has no market feed, bookmaker integration or market-movement history.
An illustrative transformation
The reciprocal of decimal odds gives a raw implied probability. For example, 2.00 corresponds to 0.50. Across all outcomes, these reciprocals may sum above 1. Dividing each by their total is one simple normalization, but it is not a uniquely correct estimate of fair probabilities.
Compare timestamps
A forecast made a day before kickoff should not be treated as contemporaneous with a market observed immediately before kickoff. Any research comparison needs data provenance, a timestamp convention and a stated normalization method.
Research boundaries
A difference between a model and a market can reflect stale data, different assumptions or model error. It does not establish a profitable strategy. NinetyQuant’s focus is transparent probability research; transactional market features are outside v0.1.
Sources & further reading
Prepared by NinetyQuant. Numerical examples are illustrative. References explain general concepts and do not validate NQ’s demo forecasts.