Championship 2026/27 Predictions: The Bookmakers vs Our Model vs TigerMat

The bookmakers. Bookmakers do not publish a predicted table, so this one is built rather than quoted. We took the outright title market and the relegation market, stripped the margin out of both, and ranked every team by title chance minus relegation chance. Eight clubs have no title price quoted at all, so they are placed on the relegation market alone. The ordering is ours; the prices underneath it are the market’s.

Our model. This is a squad-quality ranking, not a match prediction. It scores each club on the output of the players who were at it last season, then normalises across divisions so that promoted and relegated clubs can be compared with everyone else — a goal in League One is not worth a goal in the Championship, and the adjustment says so. Worth being blunt about what it is not: our match-prediction models have no view on 2026/27 yet. They need five games played by both sides before they will say anything, so they will not produce a single fixture prediction until late September.

TigerMat. My son, going on football knowledge and instinct, with no access to either of the above.

Automatic promotionPlay-offs (3rd to 8th)Relegation

Note the play-off band this season runs from third to eighth, not third to sixth. Two extra clubs get a shot at Wembley, which makes a top-eight finish worth considerably more than it used to be.

Not much, as it turns out — but the agreement that does exist is striking. All three have West Ham, Middlesbrough and Burnley in the top eight, and all three have Bolton going down. West Ham are top or second on every one of the three lists.

Beyond that it thins out fast. Of the eight play-off places, only three clubs appear on all three lists. Of the three relegation places, only one club is common to all three.

• Cardiff is the widest split on the board, thirteen places. Our model has them in the play-off eight; the bookmakers have them 5/4 to go straight back down. The model is reading a promoted squad whose attacking numbers survive the division adjustment. The market is reading a promoted club.
• Norwich splits the other way. The bookmakers and TigerMat both have them in the top eight — TigerMat has them fourth — while our model puts them seventeenth, the joint-widest disagreement in the other direction.
• Sheffield United are top eight for both the market and the model, and sixteenth for TigerMat.
• Stoke are in the model’s relegation three and comfortably mid-table for the other two.
• Wolves are TigerMat’s champions and the bookmakers’ second favourites. Our model has them ninth — the price of coming down from the Premier League with numbers earned in much harder games.

One honest observation before the season buries it: on average, TigerMat’s table sits closer to the bookmakers’ than our model’s does. Our model is the contrarian of the three, not the human.

Headline measure — average position error. For each of the 24 clubs we take how many places the prediction missed by, and average it across the table. Lowest number wins.

Band hits. Champion, the two automatic promotions, the play-off eight, and the relegated three. A straight count of how many each got right.

Both measures are set now, before anyone knows anything, precisely so that nobody gets to pick a flattering yardstick in May.

• The transfer window is open. Our model scores last season’s squads — players who have since left are still counted, and players who have arrived are not. Every one of these tables will look dated by September, which is part of the fun.
• The market ordering is derived, not quoted. No bookmaker published this table. We built it from two of their markets, as described above.
• The odds are a snapshot. Title prices are as at 7 August. The relegation market had not been re-published since 21 July, so we checked it against a bookmaker’s own list from 7 August: five of the six overlapping prices were identical, and the one that had moved (Watford) does not change the order of the table by a single place. The market’s view of who goes down has been remarkably still all summer.
• This is not betting advice, and no part of it is a suggestion that any price is wrong. It is three predictions, recorded, to be marked in May.