J3 League · Wednesday, 4 November 2026 · 05:00 UK time
CompactStats rates Osaka vs Roasso Kumamoto (J3 League, Wednesday, 4 November 2026) 47% for over 2.5 goals and 51% for both teams to score, with Osaka 38%, the draw 29% and Roasso Kumamoto 33%.
| Market | Our chance | Call |
|---|---|---|
| Over 2.5 goals | 47% | Under |
| Over 3.5 goals | 25% | – |
| Both teams to score | 51% | – |
| Osaka win | 38% | Pick |
| Draw | 29% | |
| Roasso Kumamoto win | 33% | |
| Goal in the 1st half | 65% | |
| Goal in the 2nd half | 75% | |
| Goal in both halves | 48% |
In the app: live scores, every market including half-time goals, sortable tables and a slip you can share.
| Team | Games | O2.5 | O3.5 | BTTS |
|---|---|---|---|---|
| Osaka | 13 | 54% | 38% | 62% |
| Roasso Kumamoto | 13 | 38% | 8% | 23% |
| This match | League | |
|---|---|---|
| Goal in the 1st half | 65% | 66% |
| 2+ goals in the 1st half | 32% | 34% |
| Goal in the 2nd half | 75% | 76% |
| 2+ goals in the 2nd half | 39% | 42% |
| Goal in both halves | 48% | 49% |
| Half with most goals | 2nd 46% | 2nd 47% |
| Team (own games) | 1st half goal | 2nd half goal | Games |
|---|---|---|---|
| Osaka | 58% | 80% | 50 |
| Roasso Kumamoto | 66% | 66% | 50 |
League = last season and this season combined (this season alone 64% in the 1st half). Every match ranked: first half goals today · second half · both halves · tomorrow.
More: Osaka season stats · Roasso Kumamoto season stats · J3 League table and form
League. J3 League has had over 2.5 goals in 49% of its games. That is the starting point.
Each side. Osaka 7 of 13 (5 of 7 at home), Roasso Kumamoto 5 of 13 (2 of 6 away). Steadied toward the league rate, so a few odd games can't swing it: 50% and 45%.
Last season. Still blended in until game 40: 48% and 47%.
Mix. 40% home side, 40% away side, 20% league: 48%.
Goal model. Expected goals 1.22 (Osaka) v 1.12 (Roasso Kumamoto) give 42%; 30% weight: 46%.
Calibration. Low ratings in this range have run a little pessimistic in testing, so they are lifted: 47%.
Worked from the same inputs as the rating. The full model is in the methodology.
They are calibrated chances: a 60% call should land about six times in ten and miss about four. Ratings freeze before kick-off and are graded afterwards. The model reads goals, not team news, so it knows nothing about injuries, rotation or weather. Stats, not advice · 18+.