Methodology

Decimal odds convert to implied probability by reciprocal: p = 1 / d. American odds convert first — a negative price of −n becomes n / (n + 100), a positive price of +n becomes 100 / (n + 100). No rounding is applied before the final display step.

Implied probability is not the market's honest estimate of an outcome. It contains the operator's margin, which must be removed before any comparison is meaningful.

Overround is the sum of implied probabilities across every outcome in a market, minus one. A two-way market summing to 1.045 carries 4.5 points of margin.

Our default de-vig is proportional: each outcome's implied probability is divided by the sum. Proportional de-vig assumes margin is applied evenly across outcomes, which is an assumption, not a fact. Where a page uses a different method — shin, logarithmic, additive — the page says which and why.

EV per unit staked is (p × profit) − ((1 − p) × 1), where p is your probability estimate, not the market's. The output is only as good as p. We publish the formula and the calculator; we do not publish a p, because we have no forecasting model and no feed to build one from.

Any future cross-venue comparison will rank on measurable, published attributes — quoted price, margin, market availability, settlement terms — with the capture timestamp visible, because a price comparison without a timestamp is meaningless.

We will not publish composite scores out of ten, star ratings or 'best' rankings. Those compress judgement into a number that looks like a measurement.

No odds, schedule, prediction-market, weather, participation or flow provider is connected to this site today. Every data module therefore renders its structure, names the fields it requires, and reports integration status instead of showing a value. The integration ledger on the Data sources page lists each feed and its state.