This page explains, in plain language, how the NEXORA Sport Analyzer produces its numbers and what they can and cannot tell you.
What the analyzer does
For upcoming (pre-match) football and basketball fixtures, the analyzer estimates the probability of different outcomes — for example goals totals, both teams to score, corners, cards, half-time goals and likely scorelines in football, and moneyline, projected spread and total points in basketball.
Where the data comes from
Fixtures and recent results come from TheSportsDB, a public sports database, and are requested live in your browser when you press a scan button. The free tier limits how many results each request returns, so the analyzer scans several days ahead to find as many real upcoming matches as it can.
Football model
Each team’s attacking and defensive strength is estimated from its recent results, with the most recent matches weighted more heavily and home and away form separated where enough games exist. These strengths produce an expected-goals figure (xG) for each side, which feeds a Poisson distribution to estimate the chance of each scoreline. A Dixon-Coles correction adjusts the probabilities of low-scoring results (0-0, 1-0, 0-1, 1-1), which a plain Poisson model tends to misjudge. Market probabilities (over/under, BTTS, correct score and so on) are then summed from the scoreline grid.
Basketball model
Basketball scores are far higher than football scores, so a Poisson model does not fit. Instead each team’s expected points are estimated from recent games (again weighted toward recent form and split by home/away), and a Normal distribution is used to estimate moneyline, spread and total-points probabilities.
What the percentages mean
A percentage is the model’s estimated probability, not a guarantee. When many matches share a 70% estimate, we would expect roughly seven in ten of those events to occur — never all of them, and never a specific one.
Limitations
- The models use results only; they do not see injuries, suspensions, line-ups, weather, travel or motivation.
- Teams with few recent matches produce less reliable estimates, especially early in a season.
- Shots, touches in the box and passes are modelled from each side’s expected goals, not taken from a measured feed.
- Data can be delayed, incomplete or occasionally wrong.
Our editorial standards
Articles on this site are written to inform readers about football, basketball and statistics. We aim to publish original work, to explain the reasoning behind any estimate, and to correct clear mistakes promptly — if you spot one, please tell us through the Contact page.
Reminder: everything here is for information only. Please read our Disclaimer and Responsible Gambling page. 18+.