Top probability 56% is below the 65% bar — no lean. This is the same probability read against the threshold, not a separate score.
Model lean: Arsenal (56%) — higher confidence, not one-sided.
Expected goals are model output (not shot xG); halves split by the measured Premier League share (1150 matches, 44% in the first half). League averages come from all 1150 of those matches; shot numbers are each side’s real 30-match averages.
Shots and corners are real averages over the last 30 Premier League matches; possession and pass accuracy come from this season's public stats only (1-match sample). Tags derive from that data. Formation and pressing have no reliable free source and are not shown.
| Team | Form | GF | GA | GD | PPG |
|---|---|---|---|---|---|
| LDWWL | 9 | 11 | -2 | 1.40 | |
| WWWWW | 10 | 1 | +9 | 3.00 |
| Shots (avg) | 6 | 20 |
| On target (avg) | 0 | 6 |
| Possession | 27% | 65% |
| Pass accuracy | 76% | 92% |
| Shot xG (avg) | 0.28 | 1.88 |
40–99 scale from each club's relative position on real data over the last three Premier League seasons — not subjective ratings. Samples: 77 vs 77 matches.
| 2025-12-30 | Arsenal | 4–1 | Aston Villa | EPL |
| 2025-12-06 | Aston Villa | 2–1 | Arsenal | EPL |
| 2025-01-18 | Arsenal | 2–2 | Aston Villa | EPL |
| 2024-08-24 | Aston Villa | 0–2 | Arsenal | EPL |
| 2024-04-14 | Arsenal | 0–2 | Aston Villa | EPL |
The win-probability figure is an "if absent" scenario: the player's real output (xG+xA for attackers, xGA on pitch for defenders) is compared with the club's like-for-like cover, then probabilities are recomputed. It is an add-on scenario and never alters the locked forecast.
Model lean: Arsenal (56%) — higher confidence, not one-sided.
Strength ratings 1595 v 1792 (ranked 6 and 1) — a clear gap. With a fixed home-advantage term the model makes it Aston Villa 20%, draw 24%, Arsenal 56%. Last five in the league: Aston Villa L-D-W-W-L (1.40 pts/game), Arsenal W-W-W-W-W (3.00).
Sections ③–⑦ are Pass content.
Unlock with PROTop-3 scorelines are Pass content.
Unlock with PROThe per-team split is Pass content.
Unlock with PROThe 3×3 HT/FT grid is PRO content.
Unlock with PROCorner analysis is PRO content.
Unlock with PROThe five-axis matchup is PRO content.
Unlock with PROModel lean: Arsenal (56%) — higher confidence, not one-sided.
All data on this page comes from the SUKINA model and public third-party sources — for reference only; not betting advice or a guarantee of results.

