This is how the Elo model predicts who will advance to the semifinals

The quarterfinals of the Champions League 2025-26 They already have a first reading. And it’s not football, but mathematics. From system Elo —a model taken from chess that measures the real ‘strength’ of teams based on their results—can be estimate the probabilities of classification in each round. The result: four clear favorites, although none with a completely clear path.

The starting point of the entire model is the ClubElo Elo ranking. It is not a traditional classification: it does not count titles or accumulated points, but performance. Teams, selections and even analysis houses use it because it allows level differences to be translated into probabilities of victory. In football, for example, FIFA uses a modified version of the Elo system (called “SUM” algorithm) for its World Ranking since August 2018.

Each match modifies the score that the teams can obtain depending on the rival, the result and the context. Winning against a great team adds much more than doing so against a lesser team, and losing against an inferior team penalizes more. Therefore, the Elo is a fairly accurate photograph of the current competitive level of each club.

With that basis, the first thing we do is translate that difference in level into match probabilities. That is, if a team has more Elo than another, it becomes a favorite, but not absolutely: the system converts it into a percentage of chances of winning. This results in a realistic estimate, where even the lower estimate always has a margin.

Our predictive adaptation

From there, taking the ranking at the time the pairings were established (March 18), in MARCA we have adapted the model to real playoffs. First, incorporating he factor campo —key in Champions—. Playing at home is not the same as playing away, and the model incorporates it as a small “bonus” for the home teaman average advantage of +65 Elo points. In practical terms, this makes many matches balance: a team may be inferior overall, but favorite in its stadium. This explains why several ties appear so close when analyzed match by match.

The next step is adapt those probabilities to what really happens in football: not everything is victories or defeats, there are also ties. For this, a average rate of a tie based on historical data from the Champions League that is around 27%, which allows the options of each match to be divided into three possible results. It is an important detail, because ties have a lot of weight in two-legged qualifying rounds.

Finally, We crossed all possible round-trip scenarios, including overtime and penalties (assumed as a 50-50)to obtain the final classification probability. The final result is not a closed prediction, but a sum of scenarios: a percentage that reflects how many times each team would pass if that tie were played many times. It is, in short, a way to convert the ‘strength’ of each team into real probabilities of being in the semifinals.

PSG – Liverpool: a coin toss with a slight French advantage

With that process, The most even match is that of Paris Saint-Germain against Liverpool. He PSG part with a 54% sorting optionsa minimal advantage that is explained more by the context than by a clear superiority. Both teams are favorites when they play at homewhich generates a crossing that is very sensitive to details: an extra or minus goal can completely change the scenario.

On the way, In Paris, the French have a 46.4% probability of victory, for a 26.6% from Liverpool (remember that we established the average tie at 27%), while in the return than Anfield the scenario is reversed: 40.1% for the English y 32.9% for PSG. It is, in terms of probabilities, the closest matchup to a heads or tails.

Real Madrid – Bayern: Munich tips the balance

The balance appears more inclined between Bayern Munich and Real Madridalthough without becoming definitive. The German team reaches a 55.5% probability of classification for him 44.5% of Madrid. The key is in the return in Munich. The model reflects that the German team has a greater margin in its stadium, which compensates for the equality of the first leg.

On the way At the Bernabéu, Madrid is a slight favorite (38.8% versus 34.2%)but in the vuelta the model is a lot more convincing: 47.5% victory for Bayern compared to only 25.5% for Madrid. That imbalance in the second game is what ends up tilting the tie. Even so, the overall difference remains short, which places the crossing in that intermediate terrain where there is a favorite, but not a dominator.

Barcelona – Atlético: Barça superiority even in the Metropolitan

The Spanish duel between Barcelona and Atlético de Madrid does show a clearer gap. Barça reaches 62.1% of classification options and, most significantly, He is a favorite in both games depending on the model, even in the Metropolitano. On the way en el Camp Nou, the azulgranas have a 52.3% probability of victory compared to Atlético’s 20.7%, and in the returnn the Metropolitan They are still favorites: 39.8% compared to 33.2%.

This double condition is what triggers its final probability: it does not depend solely on the field factor, but on a difference in sustained “force”. Even so, the weight of the draw and the tight scenarios keeps Atlético with real options.

Sporting – Arsenal: the most unbalanced duel

Finally, the most unbalanced confrontation is that of Arsenal in front of Sporting CP. The English team reaches a 73.1% classification probabilitybeing favorite even in the first leg andn Lisbon (49.6% against 23.4%). On the way back in London further widens the difference: 59.7% probability of victory for only 13.3% for Sporting.

These are figures that reflect a significant level gap and explain why it is the crossing with the least uncertainty of all.

In summary, the Elo system draws quarters with two very open ties (PSG-Liverpool and Bayern-Madrid),one with a clear favorite (Barça-Atlético) and another with an obvious dominator (Arsenal-Sporting). ‘Strength’ leads the way… but as always in the Champions League, reality still has a lot to say.