Most seasons when we get to this point in the Champions League, I say the same thing: Manchester City are the most likely team to win the Champions League, but the most likely outcome is that a team other than Manchester City wins the Champions League.

Well, not this year! Because Manchester City have already been eliminated from the Champions League.

City, though, were the betting favorites to win the previous seven Champions League titles at this stage in the competition. With City gone and a lopsided draw that pits four of the top five teams in the Club Elo ratings on one side of the bracket, this season’s edition of the tournament feels more wide-open than it has in a really long time.

So, to break it all down, I’m going to do what I do every year: subjectively select a handful of statistical benchmarks, compare those numbers to the previous 14 winners of the competition, and progressively eliminate teams that don’t match up until we’re left with only one team still standing.

All data, unless otherwise noted, is for domestic play only and comes via Opta and Stats Perform. With that, let’s predict the winner of the Champions League!

Predictive measurement No. 1: Scoring goals

It’s a new year, so I bring with me a new, slightly more sophisticated statistical measurement.

Well, it’s not that much more sophisticated, but it’s better than what I’ve done in the past: count the number of goals per game a team has scored. And it’s also better than looking at how many goals a team should have scored per game: expected goals.

No, at this point in the season a better representation of team strength is a combination of a team’s ability to create chances and then convert those chances: 70% expected goals and 30% actual goals.

As I write each year, the baseline for a lot of these measurements was established by the 2011-12 Chelsea team that finished sixth in its own league and won the Champions League in a game that looked like this:

Based on our adjusted goals model, Chelsea averaged a mark of 1.61 per game. That’s only slightly worse than … the other Chelsea team that won the Champions League, Thomas Tuchel’s 2020-21 side that averaged 1.62 adjusted goals per game. Every other past winner in our dataset averaged at least 1.97 in this metric — the peak was Bayern Munich‘s mark of 2.76 in 2019-20.

The only team that doesn’t break the Chelsea threshold in this year’s field is Aston Villa, who are averaging 1.5 xG, or expected goals, and 1.4 actual goals per game. Luckily for them, they play the only other club we’re eliminating at this stage, Club Brugge, who get filtered out as a team that’s not in our statistical database and hasn’t won the Champions League in any of the previous 14 seasons.

Teams eliminated:Aston Villa, Club Brugge

Teams remaining:Real Madrid, Barcelona, Arsenal, Liverpool, Bayern Munich, Inter Milan, PSG, Atletico Madrid, Bayer Leverkusen, Borussia Dortmund, Benfica, Lille, PSV, Feyenoord

Predictive measurement No. 2: Preventing goals

When it comes to conceding goals, the 2011-12 Chelsea side once again brings up the rear, but it’s not as far off from previous winners. Roberto Di Matteo’s side allowed 1.22 adjusted goals per game, while the average among the past 14 winners is 0.93, with the best being Bayern Munich’s mark of 0.65 in 2012-13.

One of the eliminated teams this season is the team that Bayern Munich beat in the final that year: Borussia Dortmund. They’re allowing 1.39 adjusted goals per game this season, which ties them with Manchester City for the 69th-best mark among teams in Europe’s Big Five top leagues, Portugal’s first division and the Netherlands’ Eredivisie. Our other casualty is Lille, who are conceding 1.4 xG per game and 1.0 goals per game in Ligue 1.

Funnily enough, Lille and Dortmund are playing each other in the round of 16. Neither team has much of a chance of winning the whole thing, but if you want to see some goals, this is the tie to check out.

Before we move on: Barcelona, currently the betting market’s second favorites (after Real Madrid) to win the Champions League, do have a better defense than 2011-12 Chelsea this season, but that’s not saying a ton. Their adjusted goals conceded (1.1 per game) is the fourth worst among the 15 teams we have data for, and it ranks just 28th among that subset of Big Five plus Portugal and the Netherlands.

Teams eliminated:Borussia Dortmund, Lille

Teams remaining:Real Madrid, Barcelona, Arsenal, Liverpool, Bayern Munich, Inter Milan, PSG, Atletico Madrid, Bayer Leverkusen, Benfica, PSV, Feyenoord

Predictive measurement No. 3: Controlling field tilt

Teams have certainly won the Champions League final itself by playing on the counter. Real Madrid three times, Chelsea twice and Liverpool all lifted the trophy after winning a match when they had less than half of the final-third possession (or “field tilt”).

Yet, while certain matchups demand being comfortable without the ball, it’s really hard to win enough matches against the best teams in the world if you’re unable to keep the ball inside your attacking third for extended periods of matches.

Even our threshold creators, 2011-12 Chelsea, maintained a field tilt of 56.1% that season. The average among the past 14 winners was 64.26%, and the ceiling was, unsurprisingly, created by the famed Messi-Suarez-Neymar Barcelona team that controlled 74% of final-third possession in its matches.

The only remaining team that doesn’t meet the 56.1% mark this season is Diego Simeone’s Atletico Madrid. Of course they don’t, you might be thinking. And well, you’re right. They’ve never been above 54.9% in any of the seasons since 2010-11.

But this version of Atleti might be the least proactive one we’ve seen. They’re allowing their opponents to complete a higher percentage of their passes than ever before (85.4%), and their field tilt is the fourth lowest they’ve recorded since 2010.

Teams eliminated:Atletico Madrid

Teams remaining:Real Madrid, Barcelona, Arsenal, Liverpool, Bayern Munich, Inter Milan, PSG, Bayer Leverkusen, Benfica, PSV, Feyenoord

Predictive measurement No. 4: Pressing

Last year, I eliminated Real Madrid from contention because they allowed one of the highest pass completion percentages in all of Europe … and I don’t mean among just teams in the Champions League. I mean among all teams in the Big Five leagues. Instead, they won it all and they lowered our threshold for that metric all the way down to 84.5%.

In fact, among the 14 previous winners, only five have allowed opponents to complete their passes at an 80% clip or higher: Real Madrid last year, Real Madrid in 2022, Real Madrid in 2018, Real Madrid in 2017 and Real Madrid in 2016.

That’s, perhaps, a warning sign for the likes of Bayer Leverkusen, Arsenal, Liverpool and Inter Milan — any of whom would become the first winner not named Real Madrid to allow opponents to complete more than 80% of their passes. However, they’re all well below what Madrid allowed last year, so they get to stay — for now.

With pressing, there’s something of a sweet spot. While scoring more goals or conceding fewer goals is never a bad thing, extreme pressing teams have rarely won the Champions League. I think that’s because teams that post outlier pressing numbers are teams that (A) dominate leagues with a playing style that can’t just then be applied to matchups against the best teams in the world, and/or (B) are prone to stylistic mismatches.

To that end, we must say goodbye to the first-place team in Germany and the joint-first-place team in Portugal: Bayern Munich and Benfica. The average Champions League winner has started its average possession 36 meters from goal. At the top end sits 2022-23 Manchester City, who started their possessions 39.6 meters from their goal. Bayern’s average possession this season begins 41.2 meters away, while Benfica’s begins 41.7 meters away.

While Bayern have put up incredible numbers in the Bundesliga, I do think this metric shines a light on their potential weakness in this tournament. In Germany, they can just win the ball high and attack, win the ball high and attack, and on and on. But as their struggles against the likes of Bayer Leverkusen and Barcelona indicate, that won’t be the case against the higher level of competition in this tournament.

Heck, this just happened against their round-of-16 opponent:

It’s, unfortunately, a different side of the same coin for Leverkusen. None of the 14 previous winners allowed more than 13.26 opponent passes outside of their own defensive third before attempting a defensive action (PPDA). While Xabi Alonso’s team checks almost every other statistical box of previous champs, it falls short here, with a PPDA of 13.73 — higher than any remaining team other than Atletico Madrid.

Teams eliminated:Bayern Munich, Bayer Leverkusen, Benfica

Teams remaining:Real Madrid, Barcelona, Arsenal, Liverpool, Inter Milan, PSG, PSV, Feyenoord

Predictive measurement No. 5: Pace and possession

The current state of European soccer is one of anti-chaos. In the Premier League, the number of possessions per game is on a steady decline, while the average duration of each possession is on the rise.

Teams are holding on to the ball for longer and turning it over way less often. Among the 14 previous winners, only two of them played games that averaged fewer than 80 possessions per team: Manchester City in 2022-23 and Real Madrid last season.

Recent trends, then, suggest that slow is the way to go, but there has to be some kind of limit, right?

No team in the round of 16 exceeds Bayern Munich’s 100 possessions per game in 2019-20. In fact, no team exceeds Chelsea’s 90.6 possessions per game the season after — and that’s the fourth-lowest number in our dataset of champions.

However, the lowest is City’s 78.5 from two seasons ago. Inter Milan (74.2), Arsenal (75) and Real Madrid (75.8) are all well below that mark. Perhaps I need to tweak how I view this number in future seasons — and I will if it keeps dropping — but for now: Inter, Arsenal and Madrid are all moving at a slower pace than any team that’s won the Champions League in recent memory.

Teams eliminated:Real Madrid, Arsenal, Inter Milan

Teams remaining:Barcelona, Liverpool, PSG, PSV, Feyenoord

Predictive measurement No. 6: Frequency of crossing the ball

This is another one of those sweet-spot numbers.

If you cross the ball too much, it probably speaks to a lack of creativity and ability among your attacking-third players or your attacking-third approach. If you don’t cross the ball at all, it speaks to a one-dimensional approach in the other way: where you can score goals only from intricate passing moves.

To win enough games against the best teams in the world, I think you need to be able to do a little bit of both. And so, that means we must eliminate Luis Enrique’s Paris Saint-Germain. Among the winners, the team that used the lowest percentage of its final-third passes on crosses (8.9%) was the 2010-11 Barcelona team that had the best soccer player of all time and maybe the best midfield of all time, too.

Luis Enrique coached his own Champions League-winning Barcelona team, but even that side crossed the ball (9.3%) more than his current PSG team (8.5%) does.

Teams eliminated:PSG

Teams remaining:Barcelona, Liverpool, PSV, Feyenoord

Predictive measurement No. 7: Physicality

No one wins the Champions League by keeping 100% of the ball, creating 10 great chances every match and conceding nothing on the other end. You win the Champions League with a good chunk of luck and by being able to function in a number of different game states and game types.

At times, well, you’ve simply got to smash into your opponents to break up the flow of the match. Or maybe you have to grab someone’s shirt to shut down a potential counter.

No team that has won this tournament over the past 14 seasons has committed fewer than 9.7 fouls. While I don’t think fouling is inherently a good thing, I do think this number serves as a stand-in for the level of physicality your team is able to impose on a match.

So that means goodbye to Feyenoord (8.2) and PSV (8.0), both of whom have passed all of the prior thresholds in the Eredivisie — a much less physically taxing league than any of the Big Five or even Portugal.

Teams eliminated:PSV, Feyenoord

Teams remaining:Barcelona, Liverpool

Predictive measurement No. 8: Defensive control

If you haven’t seen the highlights from Barcelona’s 4-4 draw with Atletico Madrid in the Copa Del Rey, then please stop reading and rectify that immediately.

But if you don’t want to, here’s what Barcelona’s defensive structure looked like while they were winning the game in injury time — no pressure on the ball, attackers on one edge of the center circle, defenders on the other edge of the center circle. Just beautiful stuff:

This team is incapable of playing anything but an absurdly aggressive high line. As my colleague Bill Connelly pointed out, Barcelona are drawing teams offside more than twice as often as any team in Europe.

Barcelona manager Hansi Flick is doing what Hansi Flick has always done: betting on an aggressive attack and aggressive press outweighing the defensive vulnerabilities created in behind the back line. It worked with Bayern Munich in 2019-20, and the fact that Barcelona have lasted this far into this exercise suggests it’s working pretty well in LaLiga.

But the reason it might not work in the Champions League this year is that Barcelona are giving up higher-quality shots than anyone else left in the field and higher-quality shots than any of the previous 14 winners. Barcelona are allowing shots with an average of 0.125 non-penalty xG per shot. The prior high was the 0.113 allowed by, yes, Flick’s Bayern team.

So, yeah. That’s 2,000-plus words and an increasingly tedious, nitpicky form of statistical analysis all to tell you the following: Liverpool, the team that finished first in the league phase, are also the team that looks most like all 14 of the previous winners.

Winner:Liverpool