We only call it when we're sure.
One line per match, published only when the model is at least 80% confident. Most matches get nothing — that is the point.
86.5%
of our calls turned out right.
4,463 calls across 1,834 matches, last 90 days.
4,463 calls across 1,834 matches, last 90 days.
3,859 right
604 wrong
604 wrong
00:00
Mon 28/09
Curaçao — Nicaragua
Over 1.5 goals
· +1 more
85%
sure
01:00
Mon 28/09
Trinidad and Tobago — Dominican Republic
Over 1.5 goals
87%
sure
01:00
Mon 28/09
Club León — FC Juárez
Club León to win or draw
82%
sure
10:25
Mon 28/09
Japan — Venezuela
Japan to avoid defeat
· +1 more
93%
sure
11:00
Mon 28/09
South Korea — Uruguay
South Korea to avoid defeat
83%
sure
12:00
Mon 28/09
Kyrgyzstan — Lebanon
Under 3.5 goals
81%
sure
12:00
Mon 28/09
Tajikistan — Palestine
Tajikistan to win or draw
· +1 more
86%
sure
15:00
Mon 28/09
Ranchers Bees — Rivers United
Ranchers Bees to win or draw
84%
sure
18:45
Mon 28/09
Belgium — France
Over 1.5 goals
82%
sure
18:45
Mon 28/09
Northern Ireland — Hungary
Under 3.5 goals
80%
sure
19:00
Mon 28/09
Guadeloupe — Barbados
Over 1.5 goals
86%
sure
21:00
Mon 28/09
Cuba — Bonaire
Over 1.5 goals
· +1 more
84%
sure
22:00
Mon 28/09
Suriname — Martinique
Suriname to win or draw
83%
sure
What this is. A confidence number, not a tip. When we say 85%, it happens about 85% of the time — that is the only thing we claim, and the record above is the whole record, misses included.
What this is not. We do not claim to beat the bookmakers' prices. The odds on these calls are short precisely because they are likely. Anything you stake is your own decision.
GET /api/predictions/ — Upcoming forecasts: 1X2 / over-under / BTTS probabilities, predicted result and confidence (0-1).
GET /api/predictions/?upcoming=false — Full history with final scores embedded — check the hit rate yourself. History restarts 2026-07-23 (new engine).
GET /api/predictions/?upcoming=false — Full history with final scores embedded — check the hit rate yourself. History restarts 2026-07-23 (new engine).