GoalSoon · 30 July 2026

Championship 2026/27: what the numbers say about who is about to surprise

West Brom played like a top-six side and finished with bottom-half points. Over three seasons that gap has predicted what happens next.

The idea, and the test

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.

How they actually played, 2025/26

-0.55-0.29-0.04+0.21+0.47+0.720.951.211.471.731.98West BromWatfordBlackburnBirminghamPortsmouthSheffield UtdQPRMiddlesbroughCharltonBristol CityStokeSwanseaDerbySouthamptonPrestonWrexhamMillwallNorwichxG difference per match → how well they actually playedpoints per match
underperformed — expect improvement overperformed — expect a fall roughly where they belong

The 18 clubs who stay, ranked by how unlucky they were

clubmatchespts/matchxG diffluck projected change
West Brom441.14+0.23-0.46+0.38
Watford431.33+0.14-0.35+0.29
Blackburn441.09-0.01-0.33+0.28
Birmingham451.40+0.27-0.31+0.26
Portsmouth441.16-0.07-0.29+0.24
Sheffield Utd431.37+0.31-0.24+0.20
QPR441.27-0.06-0.22+0.18
Middlesbrough471.70+0.60-0.09+0.07
Charlton431.07-0.43+0.03-0.03
Bristol City451.36-0.12+0.10-0.08
Stoke441.18-0.26+0.10-0.08
Swansea461.44-0.10+0.14-0.12
Derby421.43-0.00+0.15-0.12
Southampton481.73+0.40+0.16-0.13
Preston451.27-0.34+0.17-0.14
Wrexham451.56-0.16+0.25-0.21
Millwall441.86+0.21+0.27-0.22
Norwich431.44-0.10+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.

The six newcomers

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.

clubfrommatchespts/matchxG diff
West HamPremier League371.05-0.33
WolvesPremier League330.46-0.72
BurnleyPremier League340.56-1.16
BoltonLeague One251.96+0.54
CardiffLeague One202.05+0.38
LincolnLeague One222.04+0.26

What this does not know

Transfers. Nothing here accounts for this summer's squad changes. A club that sells its best two players will not improve because its xG says so.

Incomplete season. Clubs have between 42 and 48 matches on record rather than a clean 46 plus play-offs, so everything is expressed per match rather than as a total.

Small samples for the promoted clubs. League One xG coverage runs at roughly 58 per cent.

Limited power. The relationship is real but modest — luck explains about a quarter of the variance in what happens next. This narrows the range of likely outcomes; it does not name a champion.
ChampionshipxGseason preview
Follow this
New pieces go out by RSS — paste that into any reader and you get them as they land. Everything here is free, including the full season packs the numbers come from.