Full time — eQt evolved across the whole 90.
loading…
How to read the stars
Bands are fixed, so four stars means the same thing on every match in every league. A draw seldom clears three — that is the scale being honest, not broken: the draw really is the weaker call in most matches. The highlighted row is the likeliest single outcome.
What these numbers are
The 1X2 probabilities are the market's own view, de-vigged — Pinnacle where the tape holds it, otherwise the median across every book in the snapshot. Our engine does not out-predict that, and this page makes no claim that it does; the 19-league null test says public-data prediction is closed across the European pyramid.
What the engine adds is what no book posts here: goal intensities fitted by the boundary law (the difference from the 1X2, the sum from the totals market), then evolved as a 10×10 scoreline state vector by eQt — giving the scoreline distribution and the half-time split. Prices shown are the best in the snapshot across all books; exchange prices carry commission that is not subtracted here.
The three in-play windows are priced conditional on the scoreline, so the page still needs no live feed: at the break you read off the score in front of you. Because the chase multipliers key off the sign of the goal difference, only the margin moves the price — 1–0 and 2–1 are one row. Rows marked † sit at a non-level score and therefore depend on the unfitted CHASE assumption; level-score rows are assumption-free.
Frozen parameters: proportional de-vig, G = 10, KMAX = 400. Pre-match pricing uses constant intensities; the CHASE and RED_CARD multipliers are named unfitted assumptions and are not applied on this page.