Stockport County vs Doncaster Rovers Preview: Carabao Cup Opener Exposes a Telling xG Anomaly

Last updated 1 August 2026. With seven days to go until this Carabao Cup first-round tie at Stockport County, the predictions data is now in and it is worth taking a careful look at what it actually tells us, because on the surface this fixture looks straightforward and below the surface it is considerably less so.

What the Form Data Actually Shows

Stockport's home numbers over the last ten matches are genuinely impressive. Eight wins, two losses, 21 goals scored and only six conceded, which gives you a goals-per-game average at Edgeley Park that most Championship clubs would accept without hesitation. The over 2.5 rate at home sits at 70 per cent across that sample, which means this is a venue where goals tend to happen, and where Stockport impose themselves structurally in a way that produces sustained attacking output.

The interesting thing is what happens when you look at the xG data sitting beneath those results. Stockport's overall form record shows just six expected goals for against thirteen expected goals against across their last ten matches, yet they have scored 18 and conceded 14 in reality. That is a team running well ahead of where their underlying chances suggest they should be, which is a classic regression warning. It does not mean Stockport are about to fall off a cliff, but it does mean you should apply a discount to the idea that their goal production is as sustainable as the raw results imply. The market, pricing Stockport at 1.65 to win this, is largely working from outcomes rather than process, and that is where value questions start to emerge.

Doncaster Away From Home: The Numbers That Complicate This

Doncaster's sample size for the current season is genuinely thin. One match, one win, four goals scored, none conceded. You cannot build a model on that. What you can do is look at their previous season's away data in the same league tier, where across nine matches they returned five wins, two draws and two losses, with eight expected goals for and only five expected goals against. That xG split away from home is actually very healthy. It suggests a side that, when travelling, creates more than they concede in terms of genuine chance quality, even if the raw goal numbers look tighter.

Their away shots-per-game figure from that sample sits at 45.5, and possession averaged 13.5 per cent, which tells you something important about their shape. Doncaster away are not a possession-based build-up side. They are almost certainly set up to be compact, transition-focused, and efficient in transition. A low possession share combined with a positive xG differential is the fingerprint of a well-organised defensive structure with genuine counter-attacking threat. That is a profile which can cause real problems for home sides who dominate the ball but become stretched in the process.

The Model's Probability and Where the Edge Sits

The model gives Doncaster a 31.1 per cent probability of winning this match. The market at bwin is pricing that at 4.75, which implies roughly 21 per cent. That gap of ten percentage points is the kind of discrepancy that is worth examining seriously rather than dismissing. The Stockport win at 1.65 implies about 61 per cent, and the draw at 3.75 implies around 27 per cent. When you add those up you get to roughly 100 per cent after the overround is stripped out, but the distribution of probability is doing most of its work in the home win column in a way that may not be fully justified by the process data.

What the data actually shows is a home side with strong results but fragile underlying numbers, facing a visiting side who are structurally sound away from home and whose xG record in that context is notably better than you would expect from a team priced as heavy underdogs. The Carabao Cup context matters too, because both managers are likely to rotate, which disrupts the structural patterns we have been analysing and introduces variance that the model may not fully capture.

Goals Markets: A Clear Steer

On the totals, the model gives Over 2.5 a 53.5 per cent probability against a market implied probability of roughly 60 per cent at 1.66. That is a negative edge of about six percentage points, which means the market is priced tighter than the model's expectation. I would not recommend the Over 2.5 here on those numbers alone.

The interesting thing is that the Asian totals market at bet365 has Over 2.5 at 1.87 and Under at 1.97, which is much closer to a 50/50 split and actually converts to a slightly better implied probability for the Over than the standard totals market suggests. That narrower spread on the Asian line is worth noting if you are drawn to the goals angle, because 1.87 for Over 2.5 on the Asian handicap represents meaningfully better value than 1.66 on the standard market.

On BTTS, the model gives Yes a 55.5 per cent chance against the market's implied 58.8 per cent at 1.70. Again, negative edge. The model is not backing either goals market with conviction, which in itself is a signal. Stockport at home clean-sheeted in 40 per cent of their last five home matches and only 30 per cent overall in the last ten, so Doncaster scoring is plausible, but not sufficiently supported by the model to justify chasing the BTTS Yes at current prices.

Injuries and Team News

The data sheet currently carries no injury information for either side, which is not unusual seven days out from a Carabao Cup first-round fixture. Both clubs are likely to confirm their squads closer to kick-off, and given the early season timing, pre-season fitness concerns and squad depth decisions will shape the line-ups significantly. This is a match where the team news, when it arrives, could shift the structural analysis considerably, particularly if either side's primary build-up players are rested in favour of fringe squad members.

The Pick

The Doncaster away win at 4.75 with bwin is the only signal in this fixture that shows a genuine positive edge, sitting at ten percentage points above the market's implied probability. A 31 per cent model probability against a 21 per cent implied probability is a meaningful gap. The confidence rating is only 31, which reflects the thin current-season sample and the rotation uncertainty that comes with an early cup tie. This is not a high-conviction bet. It is a value bet in a specific sense: the price is long enough relative to the probability that it clears the edge threshold even at moderate confidence. My approach here would be a small stake, no more than one unit, with the explicit acknowledgement that the sample size limitations on Doncaster's current season data make this a position you hold loosely rather than one you build around. The market's number on Stockport looks generous to the home side given what the underlying numbers actually say.