Another game decided by an attacking full-back (and failure to learn lessons)

Case study: Arsenal 0-3 Chelsea, 29th November 2009

It’s hard to believe you’ll see two such similar goals in such quick succession all this season. First, an Ashley Cole cross from the left was turned into the opposite corner by Didier Drogba. Three minutes later, an Ashley Cole cross from the left was turned into the opposite top corner by Thomas Vermaelen, who unfortunately got there ahead of Drogba. If relatively basic goals are conceded twice in one game – never mind four minutes apart – it’s fair to say there’s something quite basically wrong with how the conceding side set out defensively.

Skip to 5:05 here, and watch for 90 seconds to see two identical goals.

Arsenal v Chelsea
by ninomerengez

I don’t think Arsenal’s performance on Sunday was as disastrous as could be assumed from either the scoreline, an embarrassing 0-3, or by the post-match reaction that implied Arsenal couldn’t win the title almost solely because of this result (actually, Arsenal can’t win the title because of the loss of van Persie, but that’s another matter.)

But I’m only going to focus on one thing – if that isn’t too ludicrous a statement four paragraphs in – Arsenal’s lack of preparation for dealing with Ashley Cole’s surging runs from left-back. Anyone who is vaguely familiar with modern football is well aware of the importance placed on full-backs, especially when they are as good on the ball as Cole is.

Therefore, I cannot understand for a moment why Wenger decided to play Arshavin and Nasri either side of Eduardo, when one of the two (who swapped throughout the first half) would have a strong responsibility in tracking Cole’s runs. A far better option would be to have played Emmanuel Eboue – nominally a right-back – on the right-hand side, who would have coped much better with Cole, and at least has some defensive awareness, which, frankly, neither Arshavin or Nasri have whatsoever.

It’s easy to retrospectively analyse the game and criticise a manager for a single error in choosing personnel. But the reason why I found Wenger’s tactical error so disappointing is that there were two previous games this season that demonstrate why Eboue would have been so useful up against Cole.

Firstly, only three weeks ago, Manchester United demonstrated the terrific value of having a player in the side solely to pick up Ashley Cole, when Antonio Valencia was widely praised for his excellent job against him at Stamford Bridge.

Pete Gill at Football365 said ‘Having confessed to his tactical culpability for his side being ‘too open’ against CSKA Moscow in midweek, Sir Ferguson redeemed himself in west London. With Valencia effectively man-marking Ashley Cole and Ryan Giggs pushing in from the left to nullify the anonymous Michael Ballack, Chelsea struggled to exert any authority.’ Sam Lyon for BBC Sport said, ‘With Rooney floating and Antonio Valencia pushed up against left back Ashley Cole, United stifled their opponents' rhythm.’ Simon Burnton for the Guardian said, ‘Valencia v Cole on the right flank seems the key battle. At the moment it's looking easier for Valencia to escape from Cole than the other way round (the Ecuadorian is doing a pretty good marking job when Chelsea attack).’ Most concisely, I said, ‘Valencia's advanced, wide positioning was particularly frustrating for Ashley Cole.'

Do you see? This wasn’t some obscure feature of the game picked up by me and me alone, this was the key feature of the game – in the only game all season Chelsea were outplayed. They play a diamond midfield, with two upfront. They’re desperate for width from their full-backs, and with Ivanovic a complete donkey on the ball, Cole is the only outlet. You have to close him down, and if you don’t, you’re asking for trouble.

Secondly, Wenger had already done pretty much the same thing away at Manchester United earlier in the season, when Eboue was pushed up high against Patrice Evra (the only other left-back in the world on the same planet as Ashley Cole), and it worked pretty well as Arsenal conceded from stupid mistakes rather than because they were outplayed. For the record, Eboue has also played in the front three this season against Celtic, AZ and West Ham, so it’s hardly as if Wenger thinks he’s completely out of position there. So why, in this, the one opponent you really need a defensive-minded player on the right, did Wenger go with Arshavin and Nasri, both of whom did absolutely nothing for the two goals that were conceded whilst the game was still a contest?

This blog slightly pretentiously highlights ‘lessons’ at the end of the match reports, but this episode really sums up why I do it, because it was Wenger’s failure to learn the lessons outlined above that really cost Arsenal, in what was an absolutely crucial game in their season.

A stalemate in the Lisbon derby that owed more to poor creativity than great defending

Case Study: Sporting 0-0 Benfica, 28th November 2009

Benfica made the bizarre decision of switching Pablo Aimar and Ramires from their usual starting positions, meaning Aimar was stranded out on the right, and Ramires’ surging runs from midfield were restricted against a Sporting side clearly looking to pack the midfield to stifle creativity.

Sporting’s side was basically a defensive 4-2-3-1, except with Miguel Veloso – nominally a holding midfielder – stationed on the left side of midfield. Fernandes, in the advanced central midfield position, was almost completely invisible thanks to the solid, reliable Javi Garcia, whilst the right-winger, Vukcevic, is so incredibly left-footed that it was possible to read his intentions before he’d even picked up the ball on most occasions.

Benfica should be applauded for attacking with five players, even in a tricky away game like this, but their problem was that Sporting basically had seven players aimed simply at stopping them. Aimar discovered there was little room up against Veloso, Ramires found the midfield claustrophic and Saviola often ran into Adrien, the young midfield destroyer. Angel di Maria was the one who found space, but his end product was more often than not absolutely shocking.

Benfica desperately needed support from their full-backs, yet Peixoto and Pereira appeared to be under strict orders, at least in the first half, not to venture forward whatsoever. They were afforded more freedom in the second period, but Peixoto found Vukcevic too much of a concern to venture forward too much.

The real problem was in the right-back position. With Sporting’s Veloso tucked in, there was a huge amount of space for Pereira to exploit, yet he is a player who seems completely lacking in any pace or ability on the ball. With a player comfortable in possession, at least one Sporting player (probably Veloso) would have been drawn to him, opening up space in behind, which would have made it 5 v 6 from Benfica’s perspective, rather than 5 v 7.

That said, Sporting were the home side, and their lack of ambition for much of the game was shocking. Miguel Veloso is completely wasted on the left. Perhaps there’s a case for suggesting he can create from a wide area in a similar way to how Andres Iniesta does when he plays for Barcelona, but he’s not a pacey or skilful player. The way Moutinho and he command the midfield when deployed in the centre together is impressive, and if both were given the license to move forward (with the scrappy young Adrien covering behind), it would have given Javi Garcia more of a problem.

Liedson, Sporting’s sole forward, is a decent, but limited player. Since (as of last year) he’s Portuguese, it’s fair to categorise him as a typical Portuguese forward. His movement is absolutely superb, but he simply offers little goal threat against top defences. There’s a perception that lone forwards have to be strong and tall – they don’t, they can play the role effectively by constantly showing good movement, as Wayne Rooney does for Manchester United. Liedson does that well, but when his nine outfield teammates are so far from him, it’s impossible for the space he creates to be exploited effectively.

I haven’t seen Sporting enough this season to explain why their failings currently see them closer to the relegation zone that to the top of the table (against Fiorentina in the Champions League qualifiers they seemed well-equipped), but you can’t go into a home game with such little ambition and hope to pick up wins.

Benfica disappointed yesterday, but remain one of the most tactically exciting sides in Europe.

Lessons:

• There are two options you can choose when playing a lone striker. You can have a big, strong target man (a la Heskey) or an intelligent player with good movement (a la Rooney). Both can work, but the latter will only work if you have onrushing midfielders to exploit the space the forward creates. If the midfielders are consistently thirty yards from the striker when he gets the ball, a target man would be more appropriate, to hold the ball up and wait for support.
• If you're the side on top, your full-backs must be good on the ball.

The shape of Brazil - a hard nut to crack

Team focus: Brazil

The outstanding football article of this year was written by Jonathan Wilson of the Guardian, on Brazil’s formation in the Confederations Cup. Trying to expand on Mr Wilson’s excellent prose would be suicidal, but since the piece came without diagrams – sacrilege – here is a brief summary with visual aids.

In short, European and South American football journalists interpreted the system differently. Europeans saw it as a 4-2-3-1, similar to the style, say, Liverpool play. South Americans believed it was diamond in midfield, with a forward (Robinho) dropping to the left, much as Thierry Henry did in his Arsenal days.

So, here is exactly the same formation in both diagrams, with black lines to signify the different interpretations. The European 4-2-3-1 on the left, the South American 4-4-2 diamond on the right. Which notation is correct? They’re probably as right and wrong as each other in their separate ways, but the important thing is to understand the role of each player, which is hopefully articulated through the arrows depicting movement.

So whilst most spent today's game disappointed at the fact the England team featured just two first-choice players (the reason I've ignored England’s tactics), the Brazilian formation provided some entertainment.

Pleasingly, the formation was the same as in the Confederations Cup. There were slight differences in personnel – Elano, who provided the exquisite pass for the goal, would not have played had it been for the energetic Ramires being injured. The scorer, Nilmar, featured because Robinho excused himself, and Lyon’s Michel Bastos, nominally a left-sided midfielder, has come in at left-back and should make the position his own in the next few months.

I have two points to make in addition to Mr Wilson’s:

Both Gilberto Silva and Felipe Melo come deep to pick up the ball from the centre-backs (a), and yet both are also given the license to go forward when necessary. Of course, never do the same at the same time. They are able to dovetail in a way, for example, the Lampard-Gerrard combination never could for England when they played 4-4-2. The advantage of having two holding midfield players who can also pass the ball and attack is huge, and I can’t help wondering, and hoping, that Capello will have seen this and considered that playing both Gareth Barry and a fit Owen Hargreaves would be appropriate next summer. (In my book, those two with Steven Gerrard, Frank Lampard, Joe Cole and Wayne Rooney ahead of them would be perfect, but that’s a debate for another day.)

Nilmar’s position today (b) – occupied by Robinho in the summer – is extremely difficult to pick up, especially for a defence like England’s that are used to playing against flat, boxey formations. (1) Brazil’s goal (2) the penalty incident and (3) the closest Brazil came to scoring again (when Brown blocked Nilmar’s header), were because of diagonal balls played in behind the defence towards Nilmar’s run from the left, and in that respect, Nilmar is effectively worrying both the right-back and right-sided centre-back in equal measure.

It’s been said that this Brazil side isn’t as attacking, open or exciting than in previous years, but the formation they play compared to 2002 features one fewer centre-back, and one extra attacking(ish) midfield player. The negativity is probably more down to the fact that these players are not household names and the clubs represented in Brazil’s line-up today include relative minnows Sevilla, Villarreal, Panathinaikos and Galatasaray. Nevertheless, this is an excellent Brazil side, and the likes of Bastos, Ramires and Maicon are superb footballers entirely as able as their equivalents from 2002.

The argument that Brazil’s style of play is not free-flowing is negated by the fact that their formation cannot be described adequately with mere numbers. It will be interesting how the Sunday papers depict their shape – the only thing we can be sure of is that none of them will get it right.

United win the tactical battle; Chelsea win the points

Case study: Chelsea 1-0 Man Utd, Sunday 8th November 2009

Football is not chess.

Formations and tactics don’t always win games, which is probably the first of many limitations of this blog. Today was one of those days – the only goal of the game came from a refereeing decision that was, frankly, incorrect. However, since this is a blog focussing upon tactics, we shall try to ignore that error of judgement, and instead look at the game as a whole.

It’s hard to find fault with Sir Alex Ferguson today. Look at the match statistics and it’s difficult to say that United did not have the better of the game. 61% of the possession in the second half (the first half was 50-50), 12 attempts on goal to Chelsea’s eight, seven corners to Chelsea’s none. Away from home. Yes, these facts mean nothing if not converted into goals, but a manager can only hope to dominate possession and create more chances – both of which United did. The poor shooting from United today was not Ferguson’s fault, and from a purely tactical point of view, United dominated.

Chelsea got lucky today – if I were a Chelsea fan, I would be hoping Ancelotti does not see this as a victory for a tactical shape that didn’t work.

1) The diamond formation is obvious in its weakness – width in midfield (a). This can be overcome if the midfield players are comfortable drifting to wide areas – as happened when Ancelotti played a similar style at Milan – Seedorf naturally drifted to the left side, and Gatusso had a tendency to come to the right. Even Kaka could float to the right hand side, especially when he became used to the narrow 4-2-2-2 Brazil were playing, where he was inclined to provide right-sided width.

Unfortunately for Chelsea, Ballack, Lampard and Deco are all very much central midfielders– and all three have spent a fair amount of time in their career with their club’s (Chelsea; Leverkusen/Bayern; Porto) midfield built around them. None are used to drifting to the flanks to keep the shape in the side. Meanwhile Essien, in the holding role, must stay central.

Chelsea do have two players who naturally come wide – Joe Cole and Florent Malouda, and Malouda’s excellent displays this season have provided Chelsea with that missing quality. It was disappointing to see him benched today, and if Chelsea start a home Premiership game again with Essien-Lampard-Ballack-Deco it will be very surprising, especially with Cole fit again and pushing for a starting place.

2) The first half was characterized by the battle on United’s left, Chelsea’s right (b). The fact that Valencia remained wide (c) whilst Giggs tucked in (d) meant the obvious area of space for both sides to attack was down the same side.

(Valencia's advanced, wide positioning was particularly frustrating for Ashley Cole, who got forward a lot less than his opposite number Patrice Evra, as the chalkboard on the left demonstrates.Evra attempted 20 more passes than Cole throughout the game (and in more advanced positions) a significant difference and all the more surprising considering considering Evra was the away player, and Cole the home player.)

Whereas United’s attacks often seemed to be primarily constructed to shift the ball out to Evra on the left, Ivanovic on Chelsea’s right was very much a last resort – one could almost hear the groan around Stamford Bridge every time Chelsea’s creative players were forced to come deep and shift another ball out to the Serbian, who is a good player, but not a good attacking player. It is incredible that Chelsea have four players in their squad who primarily play at right back (Bosingwa, Ivanovic, Belletti, Ferreira) and yet as soon as the first-choice becomes injured, they are suddenly stuck for any inspiration in an attacking sense from that position.

This in mind, it would have been entirely sensible to bring on Jon Obi Mikel for Ivanovic, and shift Michael Essien to right-back, where he excelled a couple of seasons ago (remember, Chelsea started him in their Champions League final against United there). Mikel would have done a decent job in the holding role, and Essien, an excellent attacking force, would have seen more of the ball than any other player on the pitch and would have been a genuine attacking threat. In the end, Ivanovic constantly underhit crosses and United were happy to let him have the ball.

It was also surprising that Nicolas Anelka wasn't instructed to move to the right when Chelsea didn't have the ball (e) which would have made it more difficult for Evra to go forward.

3) It's always surprising how rarely managers change things at half-time if their side aren’t playing well. It’s the only time they can inform and instruct their whole side to adapt to the changes, and with two clear problems (width in midfield and the lack of a right-back comfortable on the ball) that could have been rectified by a substitution, it was strange to see no changes. Yes, they went on to win the game, but unless Ancelotti was relying on a goal which came because of two separate incorrect refereeing decisions, it is hard to argue why he was right to keep things as they were. I’m always tempted to believe that managers are reluctant to change things at half-time because they don't want to appear as if they got things wrong from the outset, either in terms of tactics or personnel. Sam Allardyce, certainly not a man who cares about his public image, yesterday made two substitutions when 0-1 down at half-time, and brought on two forwards – one claimed two goals, the other claimed an assist., and Blackburn won 3-1, playing a completely different shape to their first half line-up.

4) From United’s point of view, I thought they did rather well. Their best player was Anderson, whose excellent runs only failed to result in chances/goals because of poor balls from Rooney and Valencia. Anderson was playing at the top of a midfield triangle for United, and it seems this is his best role. He came to the club with a reputation as an attacking midfielder, but all too often he has been deployed in a deeper role where he appears uncomfortable and ineffective. Today his main defensive job was to pick up Essien (much as Clarence Seedorf did to Xabi Alonso in midweek) and he did well.

This was a rare occasion where United played 4-5-1, and also a rare occasion, away at a big 4 club, where the primary role of the foremost attacking central midfield player was still a defensive one (if that makes sense), so it was perfect for Anderson. But with United (a) generally playing 4-4-2 and (b) in the absence of Cristiano Ronaldo, needing goals from wherever possible (Anderson has scored one goal in his United career) – one suspects Anderson needs either a change of shape in the United side, or a change in club, to fulfil his potential. Still only 21, he’s been written off prematurely by many, but he showed today he has the intelligence to be more than the mere run-and-foul player he often appears.

5) Wayne Rooney played well as a lone forward, as he always does. The only thing that let him down was his poor distribution - had he played in Anderson when the Brazilian was unmarked in the first half, United surely would have led. His movement is excellent and this compensates for his lack of height, traditionally a requirement for a lone forward. Fabio Capello and Franco Baldini were in the stand today and would both have been impressed. If Emile Heskey goes the whole season as a substitute, it will be difficult for him to remain in England's starting line-up, and with Joe Cole and Owen Hargreaves - certainly two of England's best eleven players - set to be available soon, it would not be surprising if Rooney plays this role in South Africa next summer.

Lessons:

• With a diamond midfield, at least one player in the four must be comfortable with drifting wide.
• It is imperative at home that both your full-backs are good on the ball.
• If you're playing against a side with one full-back not good on the ball, you can play a lopsided shape and allow that player the ball. United's system was similar to Tottenham's suicidal system last week at Arsenal. The difference? Arsenal had Sagna at right-back, good technically and who bagged two assists. Chelsea had Ivanovic, who is poor on the ball, and caused no danger.

Two similar sides cancel each other out

Case study: Milan 1-1 Real Madrid, Tuesday 3rd November

Strange one this. Cracking first half, very mediocre second – the exact opposite of the game at the Bernabeu 13 days ago. Few lessons for tactical battles as a whole, but certainly some interesting things to note from both sides.

The teams broadly played the same shape – Real with a slightly lopsided 4-2-3-1 (a left-back on the left wing, and a striker on the right wing.) Kaka tended to search for space by drifting to the left, but with Marcelo so deep, Oddo was relatively free to pick him up without leaving the right side empty.

On Football Weekly this week James Richardson described Milan’s shape at the weekend as a 4-1-2-3 – today it was a 4-2-1-3, with a slightly defensive shift in the midfield triangle, and Seedorf clearly more advanced than Pirlo and Ambrosini.

Milan’s wide players, Pato and Ronaldinho, showed no interest in tracking back whatsoever (making it a 4-2-1-3 rather than a 4-2-3-1), meaning that Real’s main threat came when Benzema moved to his traditional forward role ((a), where he pounced upon a rebound to put Real ahead) and Sergio Ramos bombed forward (b) to support the attack.

One interesting aspect was how closely Seedorf marked Xabi Alonso whenever Real had the ball (d) – even if Seedorf was further up the pitch than the ball. In other words, even if the ball was with Benzema or Marcelo, Seedorf wouldn’t track back, and would instead prevent the backwards ball to Alonso. This was excellent at stopping Real from creating from deep, but had the inevitable knock-on effect that whenever Milan won the ball back, Seedorf wasn’t in space.

Therefore, the two spare players in midfield in the first half were Diarra for Madrid, and Pirlo for Milan. Clearly, Pirlo is the more creative player, and in the San Siro a point was a better result for Real than for Milan – therefore for the second half, Pellegrini shifted Diarra forward onto Pirlo (c), making Real’s shape a 4-1-4-1 but cancelling out both midfields.

This meant that pretty much every midfield player was nullified, and the second half meandered towards the inevitable 1-1.

Lessons:

• Real look for Xabi Alonso as their outball at every opportunity, but he lacks the positional freedom in this system to go looking for the ball, so near-man-marking him is rather effective.
• Benzema is currently not intelligent enough in terms of positioning and movement to play in a wide role.
• When Milan play Pato, Ronaldinho and Inzaghi/Boriello in a forward three, all look to make forward runs whenever Seedorf or Pirlo get the ball, which is quite easy to defend against. Ronaldinho, in particular, needs to be more intelligent and move laterally into the centre to cause teams damage, especially to open up space on the left, which Zambrotta can exploit. This is precisely how Milan won the penalty they scored their only goal from.

Why is no-one mentioning the obvious reason why Man City have gone crap?

Case Study: Birmingham 0-0 Man City, Sunday 1st November 2009

Manchester City won their first four Premiership games of the season, playing a 4-3-3 system. Now, they have now failed to win any of their last four games, playing a 4-4-2 system. No-one is mentioning the shift in shape – one that appears to have weakened the side, and doesn’t bring the best out of any of their players. Why?

Firstly, Alec McLeish won this battle. OK, so Birmingham didn't actually win the game, but they hit the post, missed a penalty, and had more chances. Against a line-up ten times more expensive, Birmingham were the better side.

There are two things to say about Birmingham’s shape. One, McLeish played an aggressive pressing game high up the pitch, closing down City’s midfielders and centre-backs immediately, even eighty yards up the pitch. Two, they played extremely narrow.

Mark Hughes played into Birmingham’s hands in both respects by playing 4-4-2.

1) The issue of pressing.

It is much easier to press the opposition if they are playing broadly the same system as you in the centre of midfield. So Bowyer closed down de Jong, Ferguson closed down Barry. Their obvious out-ball was to the full-backs, neither particularly good on the ball, who were then closed down by Birmingham’s forwards. Had City played an extra man (most likely Stephen Ireland) in the centre of midfield (c), they would have been able to play around and tire the rather elderly Birmingham midfield pairing, and would have kept possession a lot more easily.

2) Not being able to exploit Birmingham's narrowness

If the opposition full-backs (and wide midfielders) are playing narrow, the natural instinct is to go around them and get crosses in, especially if Roque Santa Cruz, a natural in the air, is your striker. Sadly, playing right-footed Bellamy on the left meant he constantly wanted to cut inside to shoot, a la Old Trafford (b). Stephen Carr, however, was experienced enough to constantly show him down the line, where Bellamy not once got a decent cross in on his weaker left foot. See the Guardian chalkboard (right) for a brilliant representation of how inept at crossing he is. Blue arrows are completed passes, red ones are failed. His only successful ball into the box came from a free-kick, headed over by Lescott (19). Every single cross failed. That is not a criticism of him, as it is of Hughes, for playing him in the wrong role. Shaun Wright-Phillips (a) was happy to go outside but simply played poorly, and is never a good crosser at the best of times.

Mark Hughes’ substitutions were particularly amusing because it appeared he had got the idea when you saw the player ready to come on…only for him to withdraw completely the wrong player. Ireland was finally introduced…but it wasn’t as an additional midfield player, it was as a straight swap for de Jong. Martin Petrov was next – would he be the one accepting Carr’s invite to get to the byline and swing crosses in for 6’2 Santa Cruz? No, because Santa Cruz was the one to make way, leaving 5’9 Craig Bellamy as the target man.

City would have been far better off playing a midfield three of de Jong, Barry and Ireland, the latter driving forward. That would allow them three fluid forwards - Petrov on the left getting crosses in, Santa Cruz as a target man, and then a nice choice between Bellamy, Tevez and Wright-Phillips on the right.

Hughes appears to be utterly clueless about his ideal formation, let alone his ideal line-up, at the moment. The only question must be whether his bizarre team selections are because of Hughes' complete lack of tactical understanding, or because of pressure from City's owners.

Lessons:

• A striker who thrives on crosses must be combined with wingers who can actually cross the ball.
• A 4-3-3 means the wide players' job is probably 70% attacking, 30% defending. In a 4-4-2, it is 50-50. Playing a forward there generally means defensively you will be weak, and it was no coincidence that it was Sebastian Larsson, up against Bellamy, who was Birmingham's best attackin player.
• Craig Bellamy, Robinho, Carlos Tevez (all wide), Emmanuel Adebayor and Roque Santa Cruz (both central) are all more suited to a 4-3-3 than a 4-4-2. Martin Petrov is probably as effective in either. Only Shaun Wright-Phillips should prefer a 4-4-2, but is probably the weakest of these seven players, so to pander to his preferences is stupid.
• Stephen Ireland provides something no other Manchester City player can - driving runs from midfield. If Hughes wants to play two holding midfield players, the inclusion of Ireland can only mean a 4-3-3.

The problem with trying to play two formations at once? You end up playing neither

Case study: Arsenal 3-0 Tottenham, Saturday 31st October 2009

Tottenham had the dilemma faced by most 4-4-2 sides visting top four opposition this season – stay with their regular shape, or shift to a 4-5-1 more effective at winning and keeping position in the centre of midfield?

In the end, Harry Redknapp did neither, and settled for a compromise. Wilson Palacios, the left-sided midfielder in a four, tucked in to form a midfield three alongside Tom Huddlestone and Jermaine Jenas, with Robbie Keane dropping to the left to cover the gap left by Palacios’ drift inside. Meanwhile, David Bentley kept the width on the right-hand side. Ideally, it was a 4-5-1 when Tottenham didn’t have the ball, and a 4-4-2 when they did, with Keane joining Crouch upfront.

Clearly, the key player here was Keane. Forced to play in two positions at once, his workload was huge. Keane is not a left-sided player, nor is he blessed with impressive defensive skills. Indeed, Redknapp’s decision to play him on the left against Chelsea earlier in the season was one of the worst tactical decisions of the season, according to the ever-excellent Sarah Winterburn on Football365.

Today, the first part of the plan actually worked very well – Tottenham’s three in midfield matched Arsenal’s three (a), and Arsenal’s passing in the first half was as bad as it has been all season – short passes through the centre were constantly intercepted – Cesc Fabregas in particular found the midfield claustrophic and struggled to link up with Robin van Persie.

The compromise brought two problems, however. Firstly, it meant Tottenham played with Peter Crouch upfront alone. Crouch’s lack of pace meant Arsenal could afford an extremely high defensive line (b), meaning Crouch generally received the ball forty yards from goal. Whilst many a joke is made about Crouch’s relative lack of prowess in the air for someone at 6’7, today was the day where his height could have made a difference in the box. Not only was he up against the league’s smallest defensive partnership in Gallas and Vermaelen, he has a very good record against Arsenal, and has scored headers against them for both Liverpool (left) and Southampton. Arsenal are scared of Crouch in the penalty box, but not 40 yards from goal – as the only person who can score headers from 40 yards is Martin Palermo. This also negated David Bentley’s main talent– his crossing ability – and he was largely ineffective.

The second problem was on the defensive side of things. Bentley’s natural width meant Clichy found little opportunity to go forward (c), but Keane’s dual role meant he had to act as a left-sided midfielder in defensive situations, something he clearly struggled with, as Bacary Sagna constantly found space (d) and drove the Arsenal attack forward.

Two of Arsenal's three goals came as a direct result of the lack of pressure upon Sagna.

Keane’s lack of defensive awareness caused Arsenal’s first goal (above). A fairly innocuous situation on the right-hand side from the Arsenal throw-in, Keane failed at the fairly basic defensive task of picking up either Cesc Fabregas, or the throw-in taker Sagna (most clear on the replay). As a result, with no pressure on the ball, it was easy for Fabregas and Sagna to swap passes, Sagna had all the space he liked to whip a ball in, and Arsenal were in front. It’s hard to imagine a genuine left-sided midfield player would have afforded Fabregas and Sagna such time.

Perhaps Harry Redknapp realised this situation, because at the start of the second half he removed one of his central midfield players, and replaced him with Gareth Bale – a naturally left-sided midfield player, with Keane moving upfront. Bale’s presence largely stopped Sagna’s advance (although the withdrawal of a central midfielder meant Arsenal had a numerical advantage in the centre of midfield).

Ironically, and as if the prove the point that a left-sided player must keep their width, Arsenal’s decisive goal (above) came from Bale’s positional error.

Bale (number three) wandered infield to close down Alex Song, who gleefully knocked the ball out to now-unmarked Sagna (pause the video at 0:01 to see Sagna in twenty square yards of space), who took his opportunity to motor forward. Bale scampers back, but after a one-two and a momentary pause whilst Mark Clattenberg played an advantage, it was the same combination – a Sagna cross and van Persie finish, and Arsenal had won.

Lessons:

• Compromise formations can only work, but like any formation, only with the right players. Robbie Keane was hopelessly lost away at a big 4 club for the second time this season.
• The importance of closing down, or at least 'occupying' the opposing full-backs against top sides cannot be overstated. Often, with a congested midfield, they are the most potent creative outlet.
• Peter Crouch is not particularly impressive at holding the ball up. He is an 18-yard box player, and his lack of pace means it is easy to defend against him as a lone striker by playing a high defensive line. The stats reveal he won 14 of his 17 aerial contests - yet usually headed it down to no-one. What would have happened if those 17 had been in the penalty area?