AL- 1-0 AL- · 44% KAI 1-2 PAN · 54% AL- 1-1 ALS · 45% JUV 1-0 REN · 56% ARA 0-2 AZA · 57% DIN 1-0 NEC · 54% FER 2-0 TOR · 59% LEC 1-1 SUN · 46% OFI 0-2 BAY · 61% UNI 1-0 HAP · 54% STU 1-2 MAR · 48% FKC 1-2 FCC · 56% GET 1-2 FCL · 50% FCI 0-1 MJA · 53% AL- 1-0 AL- · 44% KAI 1-2 PAN · 54% AL- 1-1 ALS · 45% JUV 1-0 REN · 56% ARA 0-2 AZA · 57% DIN 1-0 NEC · 54% FER 2-0 TOR · 59% LEC 1-1 SUN · 46% OFI 0-2 BAY · 61% UNI 1-0 HAP · 54% STU 1-2 MAR · 48% FKC 1-2 FCC · 56% GET 1-2 FCL · 50% FCI 0-1 MJA · 53%

Can AI predict over 2.5 goals?

Updated 22 October 2026

Short answer

A little better than the base rate, and only when the models agree. Across 1,485 graded football matches, 59.5% went over 2.5 goals. When 8 to 10 of the 10 AI models predicted three or more goals, the match went over 65.3% of the time; when only 0 or 1 did, it went over 52.6% of the time.

How the number is measured

Each model predicts a score before kick-off. We count how many of the 10 predicted three or more goals, then check the final score. The record updates as matches finish.

Upcoming matches

AI over 2.5 goals predictions lists the next three days of matches with how many models expect three or more goals. See also both teams to score.

Questions

How often do football matches go over 2.5 goals?

In our sample of 1,485 finished matches across the leagues we cover, 59.5%. It varies by league and season.

Is over 2.5 easier to predict than the result?

It has only two outcomes, so the hit rates look higher, but the edge over the base rate is small. Treat strong agreement as a lean, not a lock.

Ask 10 AIs before your next match.

Free picks for every match, every model graded, and your own questions answered by the whole council.