Championship predictions: the league where the table lies
Published · Updated 2026-08-29
The Championship is the largest single competition in our dataset and the most compressed. Those two facts are why it is simultaneously the league we have the most evidence about and the one where a confident forecast is hardest to justify.
The tightest margins we record
From 2,785 Championship matches — more than any other competition we cover:
- Only 12.2% of matches finish with a margin of three goals or more. That is the lowest rate on our board, against 18.2% in the Premier League one division above and 21.6% in the Eredivisie.
- 41.8% are decided by a single goal, and 26.1% are drawn.
- 2.54 goals per match, with 47.9% going over 2.5 — the under is the favourite more often than not.
Add those together and roughly two Championship matches in three end level or by one goal. The distance between the best and worst side in the division, measured in the way that actually decides matches, is smaller than in any competition we model.
What compression does to a season
The consequence people notice is that the table moves violently. The consequence that matters for a forecast is subtler: when teams are close together in true strength, the observed table is a much noisier estimate of that strength than it looks.
A club can take a run of narrow wins and climb six places without anything real having changed, because in a league where 41.8% of matches turn on one goal, the difference between a win and a draw is frequently one deflection. Points totals in the Championship therefore contain a higher proportion of luck than the same totals in a more spread-out division — and any model that reads the table as evidence of quality will overreact to it.
Our ratings are built from scorelines rather than from league position, weighted by recency, which mutes but does not remove the problem. The honest position is that Championship probabilities sit closer together than most readers expect, and that this is the data being accurately represented rather than the model hedging.
552 matches, and what the volume buys
Twenty-four clubs playing each other home and away is 552 league matches a season, against 380 in England’s top flight and 306 in the Bundesliga. Each club plays 46 times before any play-off.
That volume is genuinely valuable to a rating model: more matches per club per season means the fit has more evidence and team ratings stabilise sooner in the campaign than they do in a shorter league. It is the one respect in which the Championship is easier to model than the Premier League.
It cuts the other way for the clubs. A 46-match season plus cup competitions and, for four of them, a play-off is a schedule with real fixture congestion — and congestion is a factor our model does not observe. It sees results, not rest days, which is a documented limitation rather than an oversight.
Promotion, and where the two divisions differ
Three clubs go up each season into a materially different competition: the Premier League averages 2.93 goals to the Championship’s 2.54, and produces a three-goal margin 18.2% of the time against 12.2%. A promoted side carries a rating estimated in a slow, compressed league into a faster, more spread-out one, which is why our early-season top-flight numbers for those clubs deserve the caution set out in what our model knows about promoted clubs.
Every Championship fixture gets 1X2, Over/Under 2.5 and both-teams-to-score probabilities, refreshed through the day, alongside form, head-to-head and the model’s verdict, on the Championship hub.
No accuracy percentage appears here by design — static text cannot be audited and would be stale within weeks. Every Championship prediction is settled after full time and published per league and per market on the track record, with misses given equal billing.