Top probability 40% is below the 65% bar — no lean. This is the same probability read against the threshold, not a separate score.
Pre-match, the model gave this outcome 40% (ranked 1 of 3).
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 |
|---|---|---|---|---|---|
| LDWLW | 8 | 9 | -1 | 1.40 | |
| LWLLW | 8 | 7 | +1 | 1.20 |
| Shots (avg) | 18 | 21 |
| On target (avg) | 6 | 6 |
| Possession | 39% | 73% |
| Pass accuracy | 81% | 93% |
| Shot xG (avg) | 1.7 | 3.77 |
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.
| 2026-04-21 | Brighton & Hove Albion | 3–0 | Chelsea | EPL |
| 2025-09-27 | Chelsea | 1–3 | Brighton & Hove Albion | EPL |
| 2025-02-14 | Brighton & Hove Albion | 3–0 | Chelsea | EPL |
| 2024-09-28 | Chelsea | 4–2 | Brighton & Hove Albion | EPL |
| 2024-05-15 | Brighton & Hove Albion | 1–2 | Chelsea | 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.
Top-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 PRONo model lean: line-ball at 40% for Chelsea, with a live draw — a match decided by details.
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.

