Inverting the Pyramid: The History of Football Tactics

by Jonathan Wilson

  Nelsinho Baptista, the experienced Brazilian coach who took charge of Corinthians in 2007, has developed software to explore the weaknesses of one system when matched against another. ‘Imagine Team A is playing 3-5-2 against Team B with a 4-5-1 that becomes 4-3-3,’ he said. ‘So Team A has to commit the wing-backs to deal with Team B’s wingers. That means Team A is using five men to deal with three forwards. In midfield Team A has three central midfielders against three, so the usual advantage of 3-5-2 against 4-4-2 is lost. Then at the front it is two forwards against four defenders, but the spare defenders are full-backs. One can push into midfield to create an extra man there, while still leaving three v two at the back. So Team B can dominate possession, and also has greater width.’

  One of Team A’s central defenders could, of course, himself step up into midfield, but the problem then is that Team A has four central midfielders, and still lacks width. And anyway, if a defender is going to step into midfield, why not simply play a defensive midfielder in that role anyway?

  Egypt won the African Cup of Nations with a 3-4-1-2 in both 2006 and 2008, but that was probably largely because straight 4-4-2 still tends to dominate the thinking in Africa. In fact, aside from Egypt - and at times Cameroon - only Guinea and Morocco, both of whom used a 4-2-3-1, did not set up in some form of 4-4-2. Significantly, the majority of the genuine contenders had strong spines and deficiencies wide and, in a generally excellent tournament, the one consistent disappointment was the standard of crossing. That could be a generational freak, or it could be related to the fact that when European clubs are looking to sign African talent, they tend to have what Manchester United’s Africa scout Tom Vernon calls ‘the Papa Bouba Diop template’ in mind. The African players who have succeeded in Europe in the past have usually been big and robust, and so clubs look only for something similar. Players called up by European clubs at a young age develop faster and have a higher profile, and so it is they who make it into the national team.

  Decline of 3-5-2

  Egypt 4 Cameroon 2, African Cup of Nations group, Baba Yara Stadium, Kumasi, 22 January 2008

  Egypt 1 Cameroon 0, African Cup of Nations final, Ohene Djan Stadium, Accra, 10 February 2008

  Vernon, who runs an academy in the hills above Accra, also believes the way the game is first experienced by children - at least in Ghana - has a tendency to shape them as central midfielders. ‘Look at how kids play,’ he said. ‘They have a pitch maybe twenty or thirty yards long, and set up two stones a couple of feet apart at either end, often with gutters or ditches marking the boundaries at the sides. So it’s a tiny area. The game becomes all about receiving the ball, turning and driving through the middle.’ The result is that most West African teams - and this was particularly true of Côte d’Ivoire - have at least two good central strikers, so tend to play them, with little width that could have unsettled Egypt’s two excellent wing-backs, Ahmed Fathy and Sayed Moewad.

  In their first game in the 2008 Cup of Nations, Egypt hammered Cameroon 4-2. They went on to add a further ten goals in disposing of Sudan, Zambia, Angola and Côte d’Ivoire before meeting Cameroon again in the final. In that first game, Cameroon’s coach Otto Pfister had his players a 4-4-2; in the final, he opted for a 4-2-3-1 and for the first time in the tournament, Egypt struggled for fluency. The defender Wael Gomaa looked like a spare part, anxiously and uncertainly wandering into midfield, and, although Egypt bossed possession, they ended up beating a limited side only because of a terrible individual error from Rigobert Song.

  Even Steve McClaren acknowledged that three at the back is only effective if the opposition play with two out-and-out centre-forwards. Given Bilić’s Croatia are one of the few sides left who do still play with two strikers - Eduardo da Silva drifting and sniffing, with either Mladen Petrić or Ivica Olić providing a more physical foil - McClaren’s decision to adopt a 3-5-2 for England’s Euro 2008 qualifier away in Zagreb actually, for all the scorn subsequently directed at him, at least made theoretical sense. The problem was that England are so unused to playing with anything other than a back four that they played it badly - and, moreover, against a team whose players were just as adept at picking apart an inadequate 3- 5-2 as England’s would have been against an inadequate 4-4-2.

  There is a theory that England had been holding Croatia before Eduardo gave them a sixty-second-minute lead, but that neglects the half-dozen decent chances they had had before then and, besides, the way Eduardo was left unmarked to head in Niko Kovač’s cross was evidence of how the discipline of their marking had disintegrated in the unfamiliar system. Gary Neville’s subsequent own-goal, as a divot confounded Paul Robinson, added a misleading element of farce to what had been a comprehensive defeat. ‘I really wanted them [England] to play with three at the back because then at each side we have one player more,’ Bilić said. ‘If we are playing slow, they have no problem because they can close you down. But sometimes we play really fast. We were very direct, very brave, and we caused them problems.’

  Bilardo suggested teams should be split with three attackers and seven defenders; Bilić opted for five and five, but the general move in the nineties was towards a middle ground, with four attacking players and six defensive. When the 4-5-1 first became popular in western Europe in the late eighties and early nineties, it was widely seen as a defensive system: ‘the right of the weak’, like early catenaccio, to be employed against stronger sides to try to frustrate them. Even today, it is not unusual to read or hear pundits complaining of sides who refuse to use ‘two up’, yet 4-5-1 was always as implicit in 4-4-2 as 4-4-2 was in 4-2-4.

  At least in a British context, strike partnerships in a 4-4-2 tend to fall into two categories: the big man-quick man (John Toshack and Kevin Keegan; Mark Hateley and Ally McCoist; Niall Quinn and Kevin Phillips) or the creator-goalscorer (Kenny Dalglish and Ian Rush; Peter Beardsley and Gary Lineker; Teddy Sheringham and Alan Shearer). In the former, there genuinely were two strikers, but in the latter, did the creator not drift deeper, linking the space between midfield and attack? The remarkable impact on English football of Eric Cantona and Gianfranco Zola was largely the result of their ability to drop off and play between the lines, confusing English centre-backs just as surely as had Matthias Sindelar and Nándor Hidegkuti. The issue, then, seems one of notation: nobody would have thought of describing, for instance, Sunderland’s promotion side of 1989-90 as playing 4-4-1-1, but with Eric Gates tucked behind Marco Gabbiadini, that is assuredly what it was. And once the instinctive recoil against 4-5-1 has been got over, it becomes apparent that it is just as flexible, just as easily recalibrated according to circumstance as the 3-5-2.

  It is arguable, in fact, that the first team to deploy a 4-5-1 to international success was the great Flamengo side of Paulo César Carpegiani, which beat Liverpool 3-0 to win the Intercontinental Cup in 1981, and could not be described as being even remotely negative. Faced with the problem of choosing between four fantasistas - Lico, Zico, Adília and Tita - Carpegiani did exactly what Brazil would do at the 1982 World Cup: he played them all. Rather than laying them out behind two strikers, though, Carpegiani opted to play with Nunes as a lone front man, with Andrade operating as a holding midfielder behind them in what would now be called a 4-1-4-1.

  Of course, five in midfield can be a defensive system. Numerous sides in the eighties used it, particularly away in Europe. Everton were particularly adept. In the first leg of the quarter-final of the Cup-Winners’ Cup away to Bayern Munich in 1985, for instance, Howard Kendall left out Andy Gray, bringing Alan Harper into midfield in his place and leaving Graeme Sharp as a lone striker. After a 0-0 draw away, Everton restored Gray in a 4-4-2 and won the home leg 3-1. Gray admits that there was nothing particularly subtle about their play that night: they realised the Bayern defence was uncomfortable against the aerial threat they posed and so set out to exploit it.

  Generally speaking, the more direct a side, the more defensive they will be in a 4-5-1. The aim for them is simply to plant
nine men between the opposition and the goal, then look for the centre-forward to battle for possession, hold the ball up and either lay it off to breaking midfielders or win a dead-ball. Ian Wright performed the role to perfection for George Graham’s Arsenal in a number of European ties in the early nineties. Once the game becomes about possession and short-passing, though, a five-man midfield becomes a far more subtle tool.

  Just who invented the 4-2-3-1 that so invigorated Europe in the late nineties, it is impossible to say. It may seem logical to place its development as happening at some time between the European Championships of 1996 and 2000, and certainly that was when it became popular, but in a sense any side whose 4-4-2 included a withdrawn centre-forward and two advanced wide players was employing it. Nobody described it as such at the time, but the Manchester United side of 1993-94, with Paul Ince and Roy Keane sitting deep in the midfield, Ryan Giggs and Andriy Kancheslskis pushing forward wide and Eric Cantona dropping off Mark Hughes used something very close to a 4-2-3-1. Arsenal did similarly in Arsène Wenger’s first full season in England, with Emmanuel Petit and Patrick Vieira deep, Marc Overmars and Ray Parlour wide and Dennis Bergkamp behind Nicolas Anelka, although Parlour could tuck in and Overmars push on to produce something more akin to an old-style 4-3-3.

  Flamengo 3 Liverpool 0, Intercontinental Cup, National Stadium, Tokyo, 13 December 1981

  There is a sense too that 4-2-3-1 is an inevitable development once sides start withdrawing one of the centre-forwards. Initially a holding midfielder would be deployed to pick him up - hence the late-nineties boom in players capable of playing ‘the Makélélé role’ - at which the trequartista would start drifting wide to find space. If the holding player followed him, that created space in the middle, so an additional player would be dropped deeper as cover, with concomitant effects for the more attacking midfielders.

  In Spain, the credit for the 4-2-3-1 tends to be given to John Toshack on his return to Real Madrid in 1999, when he used Géremi and Fernando Redondo as his holding midfielders, with Steve McManaman, Raúl and Elvir Baljić in front of them and either Anelka or Fernando Morientes as the lone striker. That said, Spanish sides have long played with split forwards - with a media punta behind the central striker - so, as in England, any team using that system with particularly advanced wingers could have been said to have been playing 4-2-3-1. Javier Irureta had been using it with Deportivo la Coruña for a couple of seasons before they won the league title in 2000, while Juan Manuel Lillo has strong claims - supported by the Spanish magazine Training Fútbol - to have invented the system while coach of the Segunda División side Cultural Leonesa in 1991-92. He had Sami and Teófilo Abajo as his two pivots, with Carlos Núñez, Ortiz and Moreno in front of them and Latapia as the lone forward. Seeing the success of the system Lillo took it to Salamanca. There, according to an editorial in Training Fútbol, the players reacted with ‘faces of incredulity because they thought it was a strange way to play; they responded to the positions they were told to adopt and the distribution of each line of the team with the same sense of strangeness and surprise as someone who had just come face to face with a dinosaur.’ Nonetheless, it took them to promotion.

  The formation’s transfer to England - at least in terms of a recognition of the formation as something distinct from 4-4-2 - came with Manchester United as an emphatic 3-2 home defeat by Real Madrid in the Champions League in 1999-2000 convinced Sir Alex Ferguson that the more orthodox 4-4-2 he had employed to win the treble the previous season had had its day in European competition (although he maintains, with some justification, that he has never played 4-4-2, but has always used split forwards).

  The 4-2-3-1, though, is just one variant of the five-man midfield. One of the attacking midfielders can be sacrificed for an additional holder, producing either a 4-3-2-1 - the Christmas tree - or the modern 4-3-3. Co Adriaanse seems to have been the first exponent of the 4-3-2-1 at Den Haag in the late eighties, and Terry Venables experimented with it with England ahead of Euro 96, but it was at the 1998 World Cup that a side using it achieved its first notable success, and it entered the mainstream.

  Aimé Jacquet’s problem was accommodating Zidane, one of the greatest playmakers the world has known, but a player of limited pace and almost no defensive instinct. His solution was to give him effectively a free role, but to do that without destabilising his team defensively, he followed the Italian convention and fielded three midfielders whose function was primarily defensive - Didier Deschamps, Emmanuel Petit and Christian Karembeu. Youri Djorkaeff was included as a further creative presence, with Stéphane Guivarc’h as the lone centre-forward. He was much derided - and it may well be that, from a technical point of view, he is the worst centre-forward ever to win a World Cup - but he performed his function, which was, broadly speaking, to provide a focal point and hold the ball up for the creators behind (once that has been accepted, the possibility opens up that Serginho’s role in Brazil’s 1982 World Cup side could be open to reinterpretation). By 2000, Jacquet had more confidence - and in Patrick Vieira a superb and mobile defensive midfielder - and felt able to align the three creators behind Henry in a 4-2-3-1.

  AC Milan are the best modern exponents of the 4-3-2-1, although theirs is rather more attacking than France’s had been. When they won the Champions League in 2006, Kaká and Clarence Seedorf were the advanced midfield presences, with Andrea Pirlo operating as a regista behind them, flanked by the snapping and snarling of Gennaro Gattuso and the unfussy efficiency of Massimo Ambrosini. Again, though, the key is fluidity, for both Pirlo and Ambrosini are comfortable advancing and Seedorf, equally, can play in a more defensive role.

  More common is the 4-3-3, which bears little resemblance to the 4-3-3 as practised by the Brazil of 1962. Theirs was a lopsided 4-2-4, with Mario Zagallo dropping deep from the wing to become an extra midfielder. With occasional exceptions, it remained asymmetric through to the eighties when, for instance, Newcastle would set out a 4-3-2, then deploy Chris Waddle on one or other flank according to which of the opposing full-backs looked weaker. The modern 4-3-3, as practised by José Mourinho’s Chelsea and many others, though, is really a modified 4-5-1.

  Here, perhaps, becomes clear the most significant recent shift in the understanding of tactics: the notion that there are only three bands - defence, midfield and attack - is gone. There is a general recognition now that each of those categories can be subdivided into smaller bands, although as that process goes on, it may be that the bands are so narrow as to cease to be meaningful. ‘It’s about the movement of your players, up and down, left and right,’ Bilić said. ‘There are no lines any more.’ Mourinho didn’t go quite so far, acknowledging that lines still exist, while maintaining that his players’ job, at least when going forward, was to break them.

  Under his management, Chelsea’s back four was relatively orthodox. Claude Makélélé then sat immediately in front of them, with Frank Lampard and Tiago - or, later, Michael Essien - operating effectively as carrileros ahead of him. Didier Drogba was the single centre-forward, with two wide men - some combination of Damien Duff, Joe Cole and Arjen Robben - operating both as wingers and auxiliary midfielders; not quite forwards, but not midfielders either. Sometimes it was 4-1-2-3, sometimes it was 4-1- 4-1, but it became more easily understood as 4-3-3.

  It remains debatable just how new the breaking of the lines is. There have, after all, always been defensive midfielders and attacking midfielders, while the notion of the pitch being divided into four bands was there in the W-M. Perhaps rather, as with Flavio Costa’s diagonal, what has changed is the notation. As that becomes increasingly sophisticated and more adept at representing the reality, so those realities become easier to grasp. Certainly today the term ‘4-5-1’ is so vague as to be almost useless to describe how a team lines up on the field. It is rather a generic term, describing a family of formations.

  Arrigo Sacchi is adamant that there have been no innovations since his Milan and, while there is self-interest in his cla
ims, there is also a large degree of truth to them. Perhaps more significant, though, is his attitude to the likes of Makélélé. Sacchi is sceptical about the 4-2-3-1 and the modern 4-3-3 with its designated holder because, to him, they are too restrictive. ‘Today’s football is about managing the characteristics of individuals,’ he said. ‘And that’s why you see the proliferation of specialists. The individual has trumped the collective. But it’s a sign of weakness. It’s reactive, not pro-active.’

  Mourinho’s Chelsea

  That, he believes, was the fundamental flaw in the galacticos policy at Real Madrid, where he served as director of football between December 2004 and December 2005 as the club signed a host of stars and tried to balance them with hard runners from the youth set-up. ‘There was no project,’ he explained. ‘It was about exploiting qualities. So, for example, we knew that Zidane, Raúl and Figo didn’t track back, so we had to put a guy in front of the back four who would defend. But that’s reactionary football. It doesn’t multiply the players’ qualities exponentially. Which actually is the point of tactics: to achieve this multiplying effect on the players’ abilities.

  ‘In my football, the regista - the playmaker - is whoever had the ball. But if you have Makélélé, he can’t do that. He doesn’t have the ideas to do it, although, of course, he’s great at winning the ball. It’s become all about specialists. Is football a collective and harmonious game? Or is it a question of putting x amount of talented players in and balancing them out with y amount of specialists?’

  When he returned to AC Milan in 1996 after his stint in charge of the national team, Sacchi ensured that Marcel Desailly, who had been used in midfield by Fabio Capello, was returned to the defensive line. Like Valeriy Lobanovskyi, for whom he professes a great admiration, Sacchi believes in the benefits of universality, in players who are not tied by their limitations to certain roles, but who can roam, taking as a reference for their positioning their team-mates, their opponents and the available space as much as the pitch. When that is achieved the system is truly fluid.