GoalSoon · 30 July 2026
West Brom played like a top-six side and finished with bottom-half points. Over three seasons that gap has predicted what happens next.
A team's results and a team's performances are not the same thing. Expected goals (xG) measure the chances a side created and allowed; goals measure what went in. The difference between the two is mostly noise — finishing streaks, goalkeeping runs, deflections — and noise does not repeat.
So before forecasting anything, I tested whether it actually holds here. Taking every club that played consecutive seasons in the Championship and the Premier League with full xG coverage (70 club-seasons, 2023→2024 and 2024→2025):
Luck in season T → change in points per match in season T+1:
r = -0.485, slope = -0.82,
t = -4.57
Control — underlying play (xG difference) in season T → points per match in T+1: r = +0.655, t = +7.15. Performances carry over. Luck does not.
| club | matches | pts/match | xG diff | luck | projected change |
|---|---|---|---|---|---|
| West Brom | 44 | 1.14 | +0.23 | -0.46 | +0.38 |
| Watford | 43 | 1.33 | +0.14 | -0.35 | +0.29 |
| Blackburn | 44 | 1.09 | -0.01 | -0.33 | +0.28 |
| Birmingham | 45 | 1.40 | +0.27 | -0.31 | +0.26 |
| Portsmouth | 44 | 1.16 | -0.07 | -0.29 | +0.24 |
| Sheffield Utd | 43 | 1.37 | +0.31 | -0.24 | +0.20 |
| QPR | 44 | 1.27 | -0.06 | -0.22 | +0.18 |
| Middlesbrough | 47 | 1.70 | +0.60 | -0.09 | +0.07 |
| Charlton | 43 | 1.07 | -0.43 | +0.03 | -0.03 |
| Bristol City | 45 | 1.36 | -0.12 | +0.10 | -0.08 |
| Stoke | 44 | 1.18 | -0.26 | +0.10 | -0.08 |
| Swansea | 46 | 1.44 | -0.10 | +0.14 | -0.12 |
| Derby | 42 | 1.43 | -0.00 | +0.15 | -0.12 |
| Southampton | 48 | 1.73 | +0.40 | +0.16 | -0.13 |
| Preston | 45 | 1.27 | -0.34 | +0.17 | -0.14 |
| Wrexham | 45 | 1.56 | -0.16 | +0.25 | -0.21 |
| Millwall | 44 | 1.86 | +0.21 | +0.27 | -0.22 |
| Norwich | 43 | 1.44 | -0.10 | +0.29 | -0.24 |
Projected change = luck × -0.82, the slope measured above. Over a 46-game season, 0.10 points per match is about 4.6 points.
These clubs arrive from other divisions, so their numbers are not directly comparable with the table above — a −0.33 xG difference in the Premier League is a far better team than a −0.33 in the Championship. They are shown separately for that reason.
| club | from | matches | pts/match | xG diff |
|---|---|---|---|---|
| West Ham | Premier League | 37 | 1.05 | -0.33 |
| Wolves | Premier League | 33 | 0.46 | -0.72 |
| Burnley | Premier League | 34 | 0.56 | -1.16 |
| Bolton | League One | 25 | 1.96 | +0.54 |
| Cardiff | League One | 20 | 2.05 | +0.38 |
| Lincoln | League One | 22 | 2.04 | +0.26 |