Football predictions powered by data and AI

Grais is an attempt to do what every football fan has imagined at least once — predict the outcome of matches before they happen.

We use modern machine learning models trained on tens of thousands of matches across leagues, teams, and seasons to estimate probabilities and uncover patterns that are not obvious at first glance.

This is not guesswork. It’s structured, data-driven forecasting.

What this product does

For every match, the model evaluates historical performance, team dynamics, and statistical signals to produce probabilities for different outcomes.

The goal is simple: highlight where there may be an edge — and where uncertainty is too high.

Some matches are predictable.

Some are chaos.

We try to tell the difference.

When predictions appear

Predictions are generated on the day of the match, when the freshest data is available. If you're viewing a future date, predictions may not exist yet — come back closer to kickoff.

How to read matches

Each match includes:

  • Predictions — what the model expects to happen
  • Confidence — how strong the signal is
  • Results — what actually happened

Signal strength (RWGS)

Strong
The model sees a clear edge
Good
A solid signal
Weak
A marginal edge
Risky
High uncertainty

After match results

Once the match is finished, we compare predictions with reality — not just by outcome, but by how well the probabilities held up.

This helps you understand how reliable the model actually is over time.

Almanac Moment

Inspired by “Back to the Future”, this is our signature concept.

An Almanac Moment happens when the model gets everything right — outcome, probabilities, and signal.

It’s rare.

And when it happens, it feels like we already knew the future.

Contact

Support: [email protected]

Partnerships: [email protected]