LiverpoolvsManchester City

Manchester City arrive at Anfield as clear favourites despite Liverpool's unbeaten run, with the model assigning them a 46% win probability against Liverpool's 26%. The decisive factor is Manchester City's perfect five-win start and superior away form, offset only by the absence of Phil Foden through suspension and a lengthy injury list that has thinned their squad considerably.

Where the model stands

The model backs Manchester City at 46% to win, with a draw at 28% and Liverpool at 26%. This represents a 20-percentage-point gap in City's favour—a substantial but not overwhelming margin for a fixture of this magnitude. The model's pick is Manchester City outright. The Elo ratings underline the gap: City sit at 1660 to Liverpool's 1541, a 119-point differential that reflects their superior season performance and historical strength.

Form and the table

Manchester City's record this season is immaculate: five wins from five matches, with no draws or defeats. They have scored 13 goals and conceded five across those fixtures, averaging 2.6 goals for and 1.0 against. Their away record is particularly striking—two wins from two on the road, yielding 3.00 points per game.

Liverpool, by contrast, have won two, drawn three and lost none. They sit seventh in the table with 9 points to City's 15 at the summit. At home, Liverpool have drawn twice and lost none, but that 1.00 point per game average masks a lack of cutting edge. Over their last five matches, they have scored 1.4 goals per game and conceded 0.8—a defensive strength undermined by attacking fragility.

Both sides are unbeaten in their last five, but City's five consecutive wins dwarf Liverpool's single victory in that span.

Model vs market

The market-implied probability (after bookmaker margin removal) shows Liverpool at 32%, a draw at 26%, and Manchester City at 42%. The market is more bullish on Liverpool than the model by 6 percentage points, whilst pricing City 4 points lower. This modest divergence suggests the market is pricing in either Liverpool's home advantage or uncertainty around City's injury depth more generously than the model does.

What the AI panel thinks

The AI panel shows overwhelming consensus for Manchester City. Six independent large language models assessed match-winner probability: GPT-5.6, Claude and Qwen 2.5 all rated City at 62%, 62% and 72% respectively, whilst Grok, Gemini and Kimi K3 ranged from 52% to 59%. The Scorebase model, aligned with this publication's methodology, sits at 42%—close to the headline model.

On the handicap market (Manchester City +0.5), the panel is even more convinced, with Qwen 2.5 and GPT-5.6 reaching 72% and 76% respectively. Over 2.5 goals is favoured by five models, ranging from 55% to 65%, suggesting both sides are expected to contribute to a moderately open contest.

Verdict

Liverpool face a Manchester City side that has been flawless in the league and possesses a superior away record. City's injury toll—including the suspension of Phil Foden—is genuine, yet their squad depth has so far proved sufficient. Liverpool's home record, whilst unbeaten, lacks the goal-scoring thrust required to trouble a City side averaging 2.6 goals per game.

The model's assessment reflects City's form, table position and away prowess. Liverpool's unbeaten streak and home ground offer defensive solidity, but not enough to overturn the underlying quality gap. The market's slight shift towards Liverpool suggests some recognition of home advantage, yet the panel's broad agreement with City's superiority is telling.

Prediction: Manchester City to win, 46%