How we analyse a match
What the model looks at, what it cannot know, and what it is actually worth.
This figure counts the matches whose projected outcome proved correct, across every match we analysed and then compared to the real result. It is recalculated continuously. We publish it including when it drops.
What the model looks at
Five families of signals, cross-referenced for every match. None is decisive on its own: it is where they agree — or disagree — that the read comes from.
Each team's last ten matches, weighted: a result from three days ago counts for more than one from two months ago. Attacking and defensive momentum are read separately — a team can string together wins while conceding heavily.
Past meetings between the two sides, with their context. A 4-0 in a friendly is not a 1-0 in a last-16 tie.
Days since the last match, distance travelled, cup-and-league congestion. A team that played 72 hours earlier 2,000 km away does not arrive in the same state.
Injuries and suspensions known at the time of analysis, weighted by how central the player is to the setup — not a simple headcount.
What is actually at stake: survival, qualification already secured, squad rotation before a bigger fixture. A team with nothing left to play for does not play like a team playing its season.
What the model does not know
An injury in the warm-up. A refereeing decision. A state of form no data captures yet. We give probabilities, not certainties — and we measure the gap between the two.
A starter pulling out forty minutes before kick-off changes the match. Our analysis is already fixed.
A red card in the twelfth minute, a contested penalty. Nobody models these events, and neither do we.
A waterlogged pitch or a crosswind affects some teams far more than others. We do not factor it in yet.
A fallout between a player and a coach, a bonus on the line, a transfer announcement. No public data captures it reliably.
Stating our limits increases trust, it does not reduce it. A tool that claims to know everything does not deserve to be believed.
Where the data comes from
Our data comes from professional football statistics providers (line-ups, live scores, head-to-head history, advanced statistics), cross-referenced with more than 1,500 news sources to capture injuries, suspensions and club context.
From the most complete plan onwards, every analysis cites the web sources it actually consulted: you can open them and check for yourself.
We use neither bookmaker odds nor betting market data, at any stage of the model.
What this is not
SportsIqo produces a read of a match: a breakdown of outcomes, indicators and scenarios, each with its confidence level. We recommend no option, we highlight none, and we display no odds.
The numbers are there to be read and compared, not followed. That is also why, on a match page, all three outcomes are shown with the same visual weight.
Today's matches, with the detail of the read.