From raw odds to a clear edge.
The MetricKick engine turns market prices into no-vig probabilities, compares them to the model, and surfaces only the mispriced markets. Here is exactly what happens at each step.
Ingest odds
Collect live bookmaker prices across every major market, refreshed continuously.
We pull odds for 1X2, totals, both-teams-to-score and more from multiple books. Capturing several prices for the same outcome lets us see where the market disagrees with itself — and where the best available price sits.
Remove the vig
Strip the bookmaker margin to derive the true no-vig market probability.
Raw odds bake in the bookmaker's margin (the vig). Summing the implied probabilities of a market gives more than 100%. We normalise them back to a fair 100% to recover what the market really thinks each outcome's probability is.
Compare to model
Measure the MetricKick model probability against the fair market to compute EV.
Our model rates teams from historical results, ratings and context, then produces its own probability for each outcome. Expected value is the gap between that probability and the price on offer: EV = (model probability × decimal odds) − 1.
Surface the edge
Only markets that clear your edge and confidence thresholds reach your feed.
A small mispricing is noise; a consistent one is signal. We filter to picks that clear a minimum expected value and a confidence threshold, so your feed stays focused on the markets actually worth a look.
Confidence & risk
Not every edge is equal. We label each pick with a confidence tier so you can judge how much weight to give it. Confidence reflects the strength and stability of the signal — it is never a promise of the result.
Strong model–market disagreement with a stable price. The edge is clear and well supported by the data.
A real but smaller edge, or one where inputs are noisier. Worth attention, sized accordingly.
A marginal edge near the threshold. Informational — included for transparency, not conviction.