Our W/D/L pick is simply the most probable of the three outcomes. Across 45,556 walk-forward matches, that was the home side 69.9% of the time, the away side 29.3% — and the draw just 0.8%. Users asked whether the draw was broken. It is not: it is arithmetic.
A draw competes against both other outcomes at once. Even in a perfectly balanced low-scoring fixture, the draw probability tops out in the mid-40s — the highest our model has ever produced is 46.8%. Bookmakers agree: a draw is almost never the 1X2 favourite anywhere in the world. Models that pick draws frequently are miscalibrated, not braver.
The draw probability itself is honest and on show in every fixture: games we rated 30-40% to draw actually drew 29.8% of the time. Away picks, meanwhile, are proportionate, not timid: away teams win 30.8% of matches and receive 29.3% of our picks, landing 44.2% against a claimed 45.4%. Home picks land 49.5% against a claimed 49.4%. The model beats always-picking-home by 4.2 points — and the probabilities mean what they say up and down the scale: 60%+ away calls landed 70.6%.
Method: walk-forward testing — every prediction is made using only information available before kick-off, then graded against what happened. Studies run June-July 2026 on the site's own dataset. Questions: contact us.