Forecast
What does the model think?
Every fixture gets a probability for each outcome, published before kickoff.
Football forecasts with the evidence beside them.
See what the model thinks, how much reliability evidence supports it, and what happened afterwards.
Estudiantes de Rio Cuarto v Racing Club
Liga Profesional Argentina
Racing Club to win44%
Home 26%Draw 31%Away 44%
Measured reliability
What Brainer Lab publishes
Four questions, asked of every fixture. Brainer Lab Forecast answers all four, and says so plainly when it cannot.
Evidence
How much reliability evidence supports this number?
Reliability is measured per market and per competition type, and not every forecast has the same evidence behind it. Where it has not been measured, Forecast says so instead of implying otherwise.
Provenance
What exactly was published, and when?
The forecast that stood before kickoff is recorded and kept, separately from anything computed later.
Accountability
What happened afterwards?
The result sits beside what was published — including the forecasts that were wrong.
Reliability
When Brainer Lab says 60%, outcomes given similar probabilities should happen roughly 60% of the time — where there is enough evidence to have measured it.
That evidence is not uniform. Some markets and competitions have been scored against thousands of settled matches; others have too little history to say anything yet. Forecast shows which is which on every forecast rather than averaging them into one reassuring number.
Live
See how the model changes as the score, the clock and dismissals change — and what changed to move it.
View live matches →The record
What was published before kickoff stays visible after the result, including the forecasts that were wrong.
View the archive →Methodology
How the forecasts are produced, what evidence each competition has, and what the numbers cannot tell you.
Read the methodology →Start with today’s fixtures
Every forecast carries its evidence, its publication time and, afterwards, its result.