Chelsea Might Be in First, But They're Not the Best Team in the Premier League
The table is lying to you once again
Let’s say you don’t watch soccer, don’t know any players or teams, maybe check in on a game or two every year, but generally know how the sport works. You were sitting around on Saturday afternoon with nothing to do -- maybe you just had lunch, maybe you’re waiting to have lunch. Maybe you were in the middle of your weekly if not daily lunch crisis? No, just me? Anyway, you were on the couch and you flipped on NBC, and it was the second half of a Premier League soccer game. The crowd was going wild, one of the coaches on the sideline had this beautiful head of hair and this literary, anguished look in his eyes. You hear the announcers call him “Thomas Frank”, and you briefly allow yourself to wonder how he had time to found The Baffler, write What’s the Matter With Kansas?, and coach a soccer team.
As the game continued, it was clear that Thomas Frank’s team -- the one in red-and-white stripes, not the all-blue -- was the better team. It also became clear to you that Thomas Frank has Thomas Frank’s jawline thanks to a lifetime of aggressively chewing on gum. At one point, after his team’s 10th or so shot in a row, Frank looked up to the sky, as if he were searching for answers from what he clearly believed to be a vengeful god. His team had been losing, 1-0, since you turned the game on, and they lost, 1-0, despite attempting the final 14 shots of the match. You briefly marveled at how people could care so much about a sport in which the clearly worse team could win a game. And not in the any-given-Sunday sense, either; at least underdogs in football or basketball typically have to play well in order to upset a favorite. This? This was just dumb-luck.
Maybe a day later you told one of your soccer-watching friends that you, in fact, watched a game on Saturday. They mentioned something about a team that had never been in the Premier League before this season playing against the defending European champs. Sounded about right based on what you saw -- until your friend told you that the blue team was the defending champ and the blue team was in first place in the Premier League.
Through eight matches, Chelsea have the most points in the Premier League: 19, one more than second-place Liverpool, two more than Manchester City. But much like on Saturday, they’ve rarely looked like anything close to the best team in England.
Let’s start here: expected goals, the best publicly available predictor of a team’s future performance.
Now, there’s still plenty of noise in numbers like this -- especially since teams have played uneven schedules so far. But there’s really not enough noise to explain such a massive gap from Liverpool and Manchester City to Chelsea. Both teams have been creating and conceding chances at a rate that is ... 650-percent better than Thomas Tuchel’s side. You can tell a story or two that shrinks the gap a bit -- Chelsea played down a man for a half against Liverpool, City played up a man against Arsenal for most of the match, stuff like that -- but nothing brings Chelsea anywhere near the levels of the other two sides. In terms of doing the repeatable things you can truly control on a soccer field -- creating good chances and not conceding them -- Chelsea haven’t been close to the best team in the Premier League so far this season.
Why do they have the most points, then? Mostly a combo of good goalkeeping, bad goalkeeping, and good luck. Based on Statsbomb’s post-shot xG1 model, which accounts for not only the location of where the shot was taken but also where it ends up on the goal frame, Edouard Mendy has saved 2.3 goals above average so far this season -- best in the league. Kepa has even added some value, too, saving four-tenths of a goal above average. However, Chelsea have allowed 10.1 xG so far this season, but only three goals. Their keeping accounts for 2.7 goals worth of that gap, but the rest of it is mainly due to something outside of Chelsea’s control: opponents misplacing their shots. On the other end, it’s the opposite story. Chelsea’s post-shot xG is exactly the same as their pre-shot xG: 11.9. Yet, they’ve scored 16 goals. In other words, against Chelsea this season, opposing keepers have made saves at a below average rate and opposing attackers have finished at a below-average rate -- less signs of a great team, and more something that’s unlikely to continue.
Granted, Chelsea did barely finish in fourth last season despite winning the European cup, but there’s been a significant drop-off from last year under Tuchel to this year -- on both sides of the ball:
There’s a notable difference in how the team is playing. Last season, they frequently just ground teams into dust by holding onto the ball for absurd amounts of time. They averaged 62.1-percent possession, and their average sequence (an uninterrupted possession) lasted for 12.1 seconds after Tuchel took over -- both numbers second in the league to Manchester City, arguably the greatest slow team we’ve ever seen. Only City moved the ball upfield more slowly than Chelsea’s 1.14 meters per second. At the same time, Chelsea also pressed aggressively and effectively, landing behind only Leeds and Leicester in PPDA2. They’d keep the ball forever, and then win it back whenever they lost it.
This season, it’s all slipped down. Chelsea have kept only 53.5 percent of possession in their matches this year, just the sixth-highest mark in the league. And two of the biggest reasons for the drop-off in overall possession is a decrease in their ability to keep the ball and in their ability to win it back. Their average sequence time has decreased while their PPDA has increased.
Without as much of the ball, Chelsea just haven’t been able to control the balance of chances as well. They’re allowing more shots and attempting fewer than they were last season. After taking 67 percent of the attempts in their matches last season, they’ve only attempted five more shots than their opponents this year: 103 to 98. It’s not only a pure volume game, either. They’re allowing better shots and more shots on the defensive end, which is how you get nearly double the xG allowed per game as last season. And when the team has managed to keep the ball, they just haven’t been as effective with it, either. Last season, 44 percent of their possessions reached the final third under Tuchel -- fourth-highest in the league. This year, that’s dropped to 38 percent; league-average is 39 percent.
What’s changed? Last season, 10 outfield players appeared in at least 50 percent of the available league minutes after Tuchel took over: Cesar Apilicueta (88 percent), Mason Mount (80), Antonio Rudiger (79), Jorginho (77), Timo Werner (71), Reece James (70), Mateo Kovacic (58), Andreas Christensen (56), Ben Chilwell (55), and Christian Pulisic (50). Of those 10, Azpilicueta, Rudiger, Jorginho, Kovacic, Christensen, and Werner are the only ones to play in at least half the minutes this season -- and both Werner and Jorginho have played a significantly smaller portion of the minutes this time around.
Last season, Chelsea were, essentially, an impossible team. They dominated possession, pressed high up the field, took a ton of decent shots, barely conceded any shots, and allowed only super-low quality attempts. Throughout the history of the sport, I’m not sure any team has done all of those things at once for a full season. You’re not supposed to be able to defend high up the field, keep the ball forever, shoot a lot, and then also defend in your own third better than Atletico Madrid. Being good at some of those things, by definition, precludes you from being good at some of the other things. For five months, though, Chelsea did it all well.
So, I think there are two ways to look at what’s happened this season. The first: Chelsea’s performance last season required such a unique and finely balanced collection of talent, and now that the talent has changed, it doesn’t matter that some of the talent might be better; what matters is the talent is different and that's made the team worse. Romelu Lukaku is a fantastic player, and although he’s only scored three goals, he’s still top 10 in the league for non-penalty xG+xA per 90 minutes. However, he’s essentially replaced Kai Havertz as the team’s de facto center forward. Lukaku has completed fewer than 70 percent of his passes over the last four-plus seasons, while Havertz is at 85 percent over that same stretch. You’re simply just less likely to turn the ball over when Havertz is on the field instead of Lukaku.
Then there are the other marginal changes. Mount is sort of the perfect encapsulation of last year’s team in that he did everything well: pressed from the front, could drop into the midfield to keep possession, and also did enough around the goal to prop up the attack. He’s played slightly fewer than half the minutes this year. While Werner just simply couldn’t score last year, his pace and off-ball movements seemed to open up space for his teammates, and in certain matches that allowed Chelsea to attack with fewer bodies and still score enough goals. It’s fun to make fun of the Jorginho-for-Ballon-d’Or absurdity, but he no doubt helped the team keep the ball, and he hasn’t been out there as much. Reece James and Chilwell offered some vital two-way play from the wingback roles -- or sometimes, in James’s case, the wide center back slot. Pulisic, meanwhile, gave Chelsea something different last year: a direct runner from wide who dribbled at defenders and got onto passes inside the box. He’s only made 11 percent of the minutes this season. Add all of those minor degradations/changes together -- some due to injury, some due to Tuchel’s choices -- and maybe it brings you somewhere closer to where Chelsea have been this season.
As for the second way to look at it, there’s a reason we’ve never really seen a team do what Chelsea did last season for a full year: it’s not possible. Your opponents eventually figure out what you’re doing, or a couple of your players stop performing at the absurdly high two-way level required for it to happen, or your opponents begin to execute on non-shot things they weren’t executing on before or the shape of the game subtly changes in ways that make your style no longer as effective -- or all of the above. Perhaps the slower-paced empty-stadium matches helped Chelsea do it, too.
So, maybe the better way to look at Chelsea right now is to look at all of Tuchel’s Premier League games in charge -- rather than to chop it up between this season and last. Over those 27 matches, they’ve averaged 1.68 xG per game and allowed 0.83 xG for a per-game differential of plus-0.85. That’s great, but still a ways away from Liverpool and City. Don’t believe me. Just listen to Tuchel himself.
“In general, Liverpool and Man City have proven in the last years they are the benchmark in terms of quality and consistency”, he said on Sunday. “They have shown what it takes to become champions. We have to improve in every aspect of the game, offensively and defensively, and we will not stop trying to create more for our strikers”.
He’s right -- about all of it.
Per FBref: “ Regular xG, or what can be considered ‘Pre-Shot xG’, is calculated considering all shots at the time of the shot without knowing the quality of the shot attempt. It not only includes shots that are on target, but also shots that are deflected or off target. Post-Shot xG is calculated after the shot has been taken, once it is known that the shot is on-target, taking into account the quality of the shot”.
Per Stats Perform: “Opponent passes allowed per defensive action, in the opponent’s defensive 3/5ths of the pitch”. The lower the number, the more aggressive the press.
So far this season Chelsea have lost to City, then to Juventus, and then were pinned back by “lowly” Brentford for the last 30 minutes of the game last weekend. Last season, Chelsea had a relatively easy path to the Champions League final - it didn’t face Bayern or PSG, then got the better of Pep in the CL final.
Last year Chelsea’s toughest opposition in the CL was Atletico. Otherwise, the opponents were mediocre: Porto’s whole team cost as much as Chelsea’s annual transfer budget and Real Madrid is nearly geriatric at this point. Chelsea will need the luck of the draw to avoid facing Bayern (the best team in Europe right now), City, and Liverpool after the group stage. Even PSG could pose a problem if it figures out how to use all its talent coherently.
I am going to say that Chelsea are my favourites to win the EPL this season, regardless of stats.. it’s too early to rely heavily on xG at this point in proceedings, and their young defenders plus the trio of Werner, Lukaku & Havertz all have a big margin of improvement to come - they are nowhere near their potential yet. Helluva squad, I think, & reigning European champs.
Liverpool’s squad is still thin for the 38 games, and Man City might’ve dropped £100m on Grealish, have up to £500m of talent in reserve for any given game, but are they on an upward curve? I don’t really think so, and might still yearn after an Aguero type goal scorer
There’s a cliché about form/stats going out of the window in local derbies, and while there’s always lots of London sides, Brentford and Thomas Frank are still honeymooners in the EPL. Chelsea ‘winning ugly’ isn’t a surprise though. How do we quantify the capacity for ‘winning ugly’?
Trends & momentum are very important to players confidence and team performances. Being CL champs & then showing up the following season with a title isn’t quite a pattern but Liverpool & Man Utd have done it before. I’m betting Chelsea do it next May.