How often are we right?
Every forecast is locked in before kick-off and scored afterwards. Here is the honest summary — including the 3,862 matches we got wrong.
What that comparison means
Always betting on the home team would have been right 43.5% of the time. We call the winner correctly 50.9% of the time. That gap is the model actually adding something — but it is a modest gap, not a magic trick.
Our last 12 calls
✓ right · ✗ wrongWhen we are confident, are we better?
Yes — and that is the only reason the confidence label exists.
This is how a tipster gets to advertise "78% winners": by only counting the easy ones. We publish every match, which is why our headline is 50.9% and not 78.4%.
What this does not mean
Being right more often than "always pick the home team" is not the same as beating the bookmakers. Our probabilities are anchored to the prices, and where the model actively disagrees with the market it has not been profitable. We publish that number too.
How the model works →