MATCH ANALYSIS: Chelsea 1-3 Manchester United

Chelsea played a very odd game over the weekend, losing 3-1 to Manchester United at Old Trafford despite generating enough chances to score three or four goals. Furthermore, Andre Villas-Boas opted to go to the seat of United's power and attack, which rather threw everyone off. While it turns out that that might not have been the best idea in the world, it sure made for an interesting match. Formation chart after the jump.

Formation

Figure 1: Manchester Unitted vs. Chelsea starting formations, 9/18/2011. Mouseover players for stats. Data: Guardian Chalkboards. Powered by Tableau.

From the tactical standpoint, there wasn't much unusual going on, so far as Manchester United are concerned. Sir Alex Ferguson has bucked the tactical trend of late, focusing on wide play and using two banks of four in defence. His 4-4-1-1 has nonetheless been very effective, mostly because the wide players are very good and Wayne Rooney's habit of roaming into the midfield while ostensibly playing as second striker mitigates many of the problems that a two-striker system presents.

Many analysts have pointed to United's shape as the key factor behind their unconvincing performance in Europe, especially against Barcelona, but I would disagree. Going toe to toe vs. Barcelona in central midfield is hubris in the extreme, so one should focus elsewhere. In Ferguson's case, he's targetting the wings whilst unleashing Rooney in the centre for maximum chaos and mayhem.

Unfortunately for United, that choice plus several injury problems have meant that their centre was very vulnerable indeed. When Darren Fletcher, Anderson, Phil Jones and Johnny Evans are essentially the core of a side, that side will be vulnerable down the middle. It was there than Andre Villas-Boas chose to target.

A Chelsea staff member told me [Sunday] morning they'd play open as they had nothing to lose accepting United superior but hoping chances would be created and taken.

-Guillem Balague. Source: Twitter.

It's unclear just how much truth is in that tweet, but what is certain is that Villas-Boas opted for an extremely open match. Raul Meireles was once more summoned into the holding role, flanked by Ramires and Frank Lampard. Fernando Torres was the man to lead the line, with Juan Mata and Daniel Sturridge on the left and right wing. The only other surprise was that David Luiz was benched in favour of Branislav Ivanovic, although that particular move was almost necessitated by choosing Meireles over Mikel. It was, more or less, the same 4-3-3 that we saw at the Stadium of Light against Sunderland, although Torres was playing higher up the pitch than Nicolas Anelka did against Sunderland.

The formation was designed for a slugging match and to ask questions of a shaky United defence. It did the attacking part of its job, but suffered badly whenever the hosts were on the ball. Meireles did not do an adequate job tracking Rooney, and while John Terry and Ivanovic were nearly perfect in the way they dealt with Javier Hernandez, the wingers gave Chelsea problems throughout, because the fullbacks couldn't support the attack without opening up channels for United to launch long diagonals to their wide players. Nani in particular tore Chelsea apart, dominating a lackadaisical Ashley Cole on our left flank.

Nevertheless, the basic thrust of Villas-Boas' strategy was a success. The fact that Chelsea were down 3-0 at halftime was probably not a fair reflection on the scoreline - the Blues had generated just as many chances as United (although, I suspect, talk of Chelsea outplaying their hosts is exaggerating the matter somewhat). Although the midfield never looked fluid, chances were being generated, and the team looked menacing on the attack.

Obviously, it all went wrong - we'll get into the whys later, but that three goal deficit wasn't entirely luck - and the manager was forced to make changes in the second half, taking off Lampard for Anelka. Lampard had been the most effective midfield player in terms of retaining possession, but hadn't really helped the Chelsea cause on the attack. With a forward coming on for a midfielder, the shape had to change, and the team became a 4-2-1-3, with Ramires dropping back and Mata playing as a trequartista and Anelka taking his spot on the left wing.

Chelsea were slightly (but only slightly) more defensively capable in this new shape, and even more of a threat to score. Torres put the Blues on the board within a minute of the restart and should have scored twice more in the half. United, of course, hit the woodwork twice and missed a penalty, so they certainly weren't having any problems cutting Chelsea's defence open - the formation change didn't improve Cole's or Bosingwa's fortunes one bit.

At the end of the day, the chosen shape represented a very bold gamble. The attackers, by and large, did their jobs. The defence did not, and that is why Chelsea lost.

On Chances

Figure 2: Manchester United and Chelsea shooting, 9/28/2011.

The narrative that emerged from the game is that Chelsea were the better side. This is, so far as I can tell, utter nonsense. Yes, the Blues had more shots. Opta might tell you they generated more chances, but that's a matter of believing whatever intern they had on this game is infallible when it comes to deciding just what is a good chance and what isn't, and considering they haven't even noticed Nani's hitting the crossbar immediately before winning a penalty, I'm going to have to go with the data on this one.

Chelsea did have more shots, but the volume of shots doesn't matter - the likelihood of scoring from said shot does. The chalkboard to the right shows where the teams were shooting from (again, it's missing the Nani crossbar incident for reasons unknown).

You'll probably notice the relative lack of United shots from outside the area. Excluding attempts that were blocked, Chelsea outshot United from range by seven to two. Inside the area is a different story entirely - the difference between the two sides in terms of shooting chances is mainly concentrated in long-range shots. And long range shooting is a bad thing.

We can get an idea of how many goals any given shot is worth based solely on field position, using historical data. Obviously, this isn't ideal, because football is about player position as much as ball position, but one still suspects it's a more useful tool than simply looking at raw shot counts or the somewhat more nebulous 'chances created'. From that, and including United's penalty, we have Chelsea 'earning' 4.03 goals and United 'earning' 3.97.

In other words there was very little to separate the sides when you look at chance quality, and there was poor finishing from both parties, something entirely obvious if you watched the match. Chelsea cannot use the excuse that they were better than their opponents at generating goalscoring opportunities, because they weren't. They were, however, quite good at it, if not the finishing aspect.

Speaking of long-range shooting being, generally, a bad idea, Daniel Sturridge was a major culprit for Chelsea. The 22-year-old appeared to adopt a shoot-on-sight policy that led to a lot of efforts that flew wide or straight at David de Gea in the centre of goal. While it's fine to take risks, the long-shot is, in general, not a particularly useful one, because it usually turns easy possession into goal kicks. That's particularly dangerous when both sides are so good at attacking - because United and Chelsea were both having difficulty extracting the ball from one another, simply giving it to them was an incredibly bad idea.

This continues a trend of Sturridge playing a very odd role with the club. Sturridge never seems to be playing quite the same game as the rest of the team. That was true of Sunderland, it was true of the Bayer Leverkusen match and it was definitely true at Old Trafford - Sturridge's approach to the game was immensely English, while the rest of the team was trying to play a possession game. I'm not sure whether this is to do with him being played out of position on the flank or by design (although I don't know why it would be) but currently he's both isolated on the right wing and wasteful when he does get the ball. That needs to change if he's to retain his starting spot.

A Certain Striker

Figure 3: Chelsea radial passing maps at Manchester United, 9/18/11. Powered by Tableau.

Let's talk about Fernando Torres. He'll be remembered for that awful miss after rounding David de Gea in the second half, but he put in what was probably his strongest shift in a Chelsea shirt, scoring a goal, playing a pass that only failed to be an assist on a technicality (blame Ramires!) and generally playing incisive football throughout the match.

Torres' problems in a Chelsea jersey have been well documented. The most obvious issue the striker has is a complete lack of decisiveness. I don't want to speculate as to his mental state, but when Torres is going bad, he dwells on the ball for too long and allows himself to be crowded out rather than take a shot or play a pass. That's obviously not an ideal way for a striker to behave.

Here, we saw none of that. There was never any dawdling on the ball. Torres was (presumably) identifying what he wanted to do and getting it done as quickly as possible rather than waiting for something to happen to him. While the three most memorable moments of the match, as far as he's concerned, were probably the pass to Ramires, the goal and the rather epic miss late on, I don't think that any of them are as instructive as two other incidents - the attempted bicycle kick from a Danny Sturridge cross in the first half and the weaving run through the United defence that nearly resulted in his second goal.

Those two moments speak to a Torres Liverpool fans know very well and that Chelsea ones don't (disregarding the matches he played against us, that is). Call the bicycle kick reflexive, call it confidence, call it whatever - it still required a decision to be taken and implemented instantly, leaving no room for self doubt. The dribble was similarly impressive, taking two defenders out of the game and ending with a vicious shot de Gea barely beat away. Have we ever seen Torres take a player out like that, let alone two? I'm going with no.

So, Torres was extraordinarily good, despite the miss to end all misses. That's not really useful, though. Why was he good?

I can name, off the top of my head, four Torres performances that stand out as 'good': The away match in Copenhagen, the cameo against West Ham United in which he notched his first Chelsea goal, the Stoke City draw (although this is questionable, since everyone around him was poor) and the United game. Throwing out Stoke, where Torres got a grand total of zero good chances, the common thread in all of this is the presence of Nicolas Anelka, who has now provided both of the Spaniard's assists in a blue shirt.

Why does the Torres-Anelka pairing seem to work so well? Let's let Anelka tell us:

[The assist came from] a good move from Fernando, I saw him and tried to put him in the space for him to score and it worked, it was good. It was a clever pass but he made the good move, I just had to adjust my power for the pass and everything came perfectly. These kind of runs I do when I am a striker so I know what Fernando wants and how he wants to score goals because I am almost the same. It makes it easier to give him the right pass.

-Nicolas Anelka. Source: ChelseaFC.com.

With Daniel Sturridge's struggles laid out earlier, it's probably time to run with a Mata-Torres-Anelka forward three, at least for league play. Those three plus Raul Meireles appear to have a very good understanding, and Chelsea's season really hinges on getting their £50M man back onto top form.

Hardly Dominance

Figure 4: Manchester United vs. Chelsea team/individual passing, 9/18/2011. Powered by Tableau.

Chelsea's previous four opponents have had nothing like a possession advantage against the Blues - Chelsea had outpassed Stoke, West Bromwich Albion, Norwich City and Sunderland by a combined 1104 attempts, which is rather a lot for four matches. Against Manchester United, there was never going to be a clear-cut advantage in terms of possession, and the hosts unsurprisingly had the better of things for the vast majority of the match.

Chelsea had season-low passs attempted and season-high conceded, hardly a surprise considering where they were and who they were playing against. A comparison between the same fixture last year, while probably not particularly useful, is still interesting - in the 2010/11 season, Chelsea attempted 463 passes (80 percent completion) at Old Trafford to United's 408 (73 percent), while this year the equivalent numbers were 447 (83 percent) and 575 (83 percent).

Passes don't tell the whole story, of course, but Carlo Ancelotti fielded a far more defensively robust midfield against United last season that Villas-Boas attempted this season, and they were aggressive in pressing their hosts, vastly reducing their passing tally. This time around, Chelsea refused to press the midfielders until they crossed the halfway line, leading to Anderson and Darren Fletcher having it relatively easy in terms of settling the ball when possession was gained and giving them plenty of time to find their wingers.

The pressing thing is interesting, because many are pointing to the midfield press as a major weapon in shutting down teams, and are baffled by its absence. While it is true that one can put immense pressure on the ball holder with a very aggressive press, it's also an inherently dangerous technique. Andre Villas-Boas actually discussed very early on when he described his tactical philosophy - in order to break a defence, you have to lure them to you with the ball. Once broken, they're far more easily circumvented, either with good one-on-one play or by exploiting the holes that appear in the shape.

Given the evidence from Sunday's match, it would appear that Chelsea made an error in not applying more pressure to Fletcher and Anderson when they were in possession - this made the sweeping pass to the flanks a more viable option for United, and that's where they hurt us. However, pressing them in the middle would have made things very dangerous should either player have circumvented the midfielder attacking them, because the Blues would have been left with a player stranded up the pitch while one of Anderson and Fletcher, with support from Wayne Rooney, raced forwards. That's not a particularly palatable option either, so in the end it'd hard to blame Chelsea for playing like they did.

And In Defence?

Apart from finishing, what went wrong for Chelsea? The midfield was good enough, the forward movement was highly encouraging and the team probably created enough chances to win the average match at Old Trafford. Javier Hernandez was marked out of the game by Branislav Ivanovic and John Terry, and despite going up against Raul Meireles, Wayne Rooney wasn't any more dangerous than his usual devastating self, which is about all you can hope for when up against Rooney when he's in the middle of one of his hot streaks.

The major problems came on the wings. There were three players in particular that Chelsea had real problems with - Ashley Young, Patrice Evra and Nani. The Young-Evra axis was dangerous because Sturridge did very little defensive work and wasn't a major threat in the attacking third, allowing United's left back to team up with Young to cause Jose Bosingwa real problems. The Portuguese fullback was by no means spectacular, but he did about as well as he could in a tricky situation.

On the other flank, Ashley Cole was absolutely dire against Nani, whose modus operandi was to pick up the ball on the right, beat a defender, and drive infield. With Juan Mata unable to provide much assistance on the defensive side of things (although he did prevent Chris Smalling going forward), Andre Villas-Boas was relying on Cole having a good day to stop Nani's runs. He wasn't, and Chelsea were badly, badly hurt by it. Cole had been showing signs of life as of late, but they certainly weren't there on Sunday, and the Blues paid the price.

Another major issue the defence did a poor job with was a simple action - ball recovery. Chelsea are increasingly having difficulty tackling opponents cleanly and are also on occasion immensely sloppy when they do recover possession. Goals number one and three were the result of an inability to transition from defence to attack effectively, with Bosingwa being unable to play the ball out in the leadup to the free kick that ended up with Smalling scoring, and number three being rather more direct when John Terry's botched clearance bounced off Nani to land at the feet of Rooney.

Conducting that transition is a hugely underapprectiated part of a defender's job, and failure to do so means having to soak up more shots and spend more time on the back foot. Without looking at any statistics, one would suspect that Chelsea's best player in this regard is one David Luiz, whose skill on the ball compared to the rest of the defenders on the squad makes him look like Lionel Messi. Granted, he makes his share of catastrophic errors, but with the Blues needing less force and more finesse at the back, it's not crazy to suggest that he should have started here.

Overall

At the end of the day, I strongly suspect claims of Chelsea superiority have been vastly overblown. The team performed well enough on the attack, but the defending was suspect throughout and you'd have to believe that the inclusion of Danny Sturridge and Branislav Ivanovic (or possibly Jose Bosingwa) over Nicolas Anelka and David Luiz were mistakes. At best, Chelsea 'deserved' a draw.

That said, very few teams could travel to Old Trafford and play well enough to even have the conversation, and despite the scoreline the Blues were hardly blown off the pitch. On top of that, Fernando Torres finally began playing like he should be and we've learned that Raul Meireles might just about be able to hold even against good teams, as long as you don't care about him being able to attack, which is... something, I guess. More than anything else, though, this game gave us an idea as to the strengths and weaknesses of the team, as well as giving us some idea of what Chelsea's best XI might look like. Since you're about to ask:

Chelsea (4-3-3): Petr Cech; Ashley Cole, John Terry, David Luiz, Jose Bosingwa; John Obi Mikel, Raul Meireles, Ramires; Juan Mata, Fernando Torres, Nicolas Anelka.