How to Predict Champions League Matches: Comparing Clubs From Different Leagues

In a domestic league every team plays every other team, so their results line up on one scale. A club's goal difference against twenty league rivals tells you a lot about how it compares with them. In the Champions League and the Europa League most pairs of clubs have never met, and their leagues have no direct link. A team that dominates its own league may simply be the best side in a weaker league.

So the first question in any European match is not 'who is in better form?' but 'how strong is each league?'. Getting that wrong moves every probability in the match, often by more than form or injuries do.

Start with what is measurable. Our model (Dixon-Coles) rates each club's attack and defence from roughly the last two seasons of its own league, weighting recent matches more. That gives an expected-goals rating relative to the league average: a club scoring 40% more than an average side in its league gets a clear attack bonus, whatever league that is.

This step is pure data and works the same in the Premier League, the Eredivisie or the Norwegian Eliteserien.

Next, each league gets one strength level, with the Premier League at zero and the others below or near it. A gap between two leagues raises the expected goals of the club from the stronger league and lowers those of its opponent by the same factor.

Where does that level come from? Results between leagues are rare, so we start from our own estimate of each league's level and then nudge it, gently, towards what the sharpest betting market implies across all the European fixtures we have seen. The difference between two clubs of the same league stays purely our model; only the gap between leagues leans partly on the market. We say this openly on every European page.

Home advantage still matters in Europe: travel, crowds and familiar pitches all count. Our model fits a separate home advantage for European games instead of reusing each league's own number.

The league phase also changes motivation. Every club plays eight games against different opponents, the top eight go straight to the last 16 and places nine to twenty-four go to a play-off round. In the final rounds some clubs are already through or out, which can mean rotated line-ups. A goal model does not see that, so treat late league-phase probabilities with a little extra caution.

On each Champions League and Europa League match page you see the win, draw and loss probabilities, over 2.5 goals, both teams to score and the most likely scores, next to the market's margin-free price. Each club's form comes from its own league, with that league named next to it.

When our number and the market differ a lot, the market is usually closer: it reacts to team news and is sharper than any public model. We show both so you can see the difference, and every prediction is locked with a hash before kick-off and graded afterwards.