Fulham vs AFC Wimbledon: Carabao Cup Preview, Odds and Predictions (25 August 2026)

Last updated 18 August 2026. Fulham host AFC Wimbledon in the Carabao Cup on Tuesday 25 August 2026, with kick-off scheduled for 01:00 UK time. This is a fixture where the class gap between the two sides is considerable, but the interesting thing is that the model is flagging one market where the bookmakers appear to have underpriced the probability in a way that is worth examining carefully.

The State of Fulham Heading Into This Fixture

Fulham's form data shows a team that performs meaningfully better at home than away, which is relevant context for this game even though the Carabao Cup will invite rotation. In their last five home matches, Fulham have recorded three wins, no draws and two losses, scoring six goals and conceding three. The momentum slope on their home form sits at 0.3, which is the highest reading across any of their contextual windows and suggests a genuine upward trend at Craven Cottage rather than a statistical blip. Over the last ten home games, that picture extends to six wins, one draw and three losses, with fifteen goals scored against ten conceded. The over 2.5 goals percentage in that ten-game home window is 60 percent, and the both-teams-to-score rate is also 60 percent, which is a meaningful signal when we start talking about the goals markets.

The injury sheet confirms one Fulham player is currently out with a moderate injury, with a return date not expected until 9 October. The player identifier is anonymised in the available data, but the timing suggests this is a player who has been absent since early August and will not feature here. In a cup fixture where Fulham will rotate, the depth of their squad becomes the relevant variable rather than the starting eleven they would field in the Premier League.

Away from home, Fulham's last five games show zero wins, three draws and two losses, with just one goal scored and six conceded. That context is not directly relevant when they are playing at home, but it does tell you something about the team's underlying shape in games where they have to be progressive and carry the structural initiative. At home, that initiative is naturally theirs. The Carabao Cup context means they will likely set up with genuine intent to win, even with changes, because an early exit serves nobody.

AFC Wimbledon: The Data Problem

This is where the analysis requires honesty about sample size. AFC Wimbledon's form data in their current league competition covers exactly one match. One draw, one goal scored, one conceded. You cannot build a reliable picture of their defensive shape, their pressing triggers or their build-up structure from a single result, and anyone who tells you otherwise is doing you a disservice.

The more informative window comes from a separate competition dataset, which covers their last five games overall and shows one win against four losses, with two goals scored and twelve conceded. The expected goals figures here are striking in a way that cuts against simple narrative. AFC Wimbledon generated an xG of nine across those five games, which means they were creating opportunities at a reasonable rate, but their actual goals output was two. Their possession average was 13.5 percent, which means they were working almost entirely without the ball and relying on transition and set-piece moments to generate those chances. The xG against was sixteen, with an over 2.5 goals rate of 60 percent across that window.

What the data actually shows is a team that defends deep, concedes heavily in terms of volume when facing dominant opponents, and does not convert the chances their transition play creates. That is a structural problem, not a random one, and it is the kind of profile that fits a team travelling to a Premier League ground in a cup competition.

The Model Signal and Where the Value Is

The system has generated two signals for this fixture. The first is AFC Wimbledon to win at odds of 9.00 with bwin. The model assigns Wimbledon a 32.6 percent probability of winning, against a market implied probability of 11.1 percent. That is a stated edge of 21.5 percentage points, which is a very large number. The confidence rating on that signal is 33 percent.

I want to be careful here because that 32.6 percent figure requires scrutiny. Given the sample size issue I mentioned, the model is working with limited and potentially unrepresentative data for Wimbledon. The possession figure of 13.5 percent across five games is so extreme that it may be distorting the model's output in ways that inflate Wimbledon's apparent probability. I would not back the away win at these odds based on the underlying data quality available.

The second signal is both teams to score at 2.20 with bet365, with the model assigning a 54.7 percent probability against the market's implied 45.5 percent. The edge here is 9.2 percentage points and the confidence is 55 percent. This one I find more credible, and here is why.

Fulham's home ten-game BTTS rate is 60 percent. Their goals-for rate at home is high. But crucially, Wimbledon's xG data suggests they are manufacturing enough attempts to score even when they are being dominated positionally. An xG of nine across five games is not nothing. Even if they only convert a fraction of those opportunities, the probability of them finding the net in a game where Fulham will likely score at least one feels underpriced at 2.20. The market is pricing this as though Wimbledon are highly likely to be shut out, but their underlying chance creation in the data does not support that assumption as firmly as the odds imply.

Betting Summary

The signal I am treating as actionable is BTTS Yes at 2.20. The model edge is modest rather than enormous, the sample size for Wimbledon is a concern, but Fulham's home scoring record and Wimbledon's underlying xG generation across their recent games point in the same direction. The Fulham win to nil at 1.83 is the market I would fade, because that price implies a near certainty of Wimbledon failing to score and I do not think the data supports that level of confidence.

The AFC Wimbledon outright win at 9.00 carries a headline edge that looks attractive on paper. But a 33 percent confidence rating from the model combined with the structural weaknesses in the away side's data makes this a signal I am watching rather than acting on. There is a meaningful difference between a large stated edge and a reliable large edge, and with one competitive result in the book for Wimbledon this season, the former is much easier to calculate than the latter.

Prediction

Fulham to win, with both teams scoring. The structure of this fixture strongly favours the home side, the Carabao Cup context will not materially change that, and the goals markets offer more genuine information than the match result at the prices available.