Sunday, June 14, 2009

3-5-2: dying in Europe, less so in Brazil

Figure 1 (left): One of Muricy Ramalho's many and myriad formations at São Paulo.

It is almost common knowledge to observers of European football, that 3-5-2 is dead. Or at least dormant.

Regardless of which nation he comes from or which one he works in, whenever some benighted coach raises the prospect of setting up his team with anything other than a back four, he is greated as a mothballed eccentric oblivious to the march of progress. Now, clearly the Dutch-orginating 3-4-3 systems are granted an exception, as they have always seemed so avant-garde, almost too futuristic to become dated, and in terms of their 'etymology', they have very little history in common with the popular interpretations of 3-5-2.

None of the above is meant to discredit popular wisdom on the issue; indeed, European football's systems of three centre-backs have all tended to share common characteristics and have by-and large been superceded by back four systems largely because their effectiveness became neutralised by the latter. On other counts, even in those departments where comparative advances were seemingly offered by back-three systems, such as allowing both full-backs to advance simulataneously, these advantages have been largely co-opted by the back four, whose modern full-backs now benefit from improved athleticism and stamina and thus do a splendid imitation job of the old wing-backs.

Already, right here and speaking of terminology alone, we are entering murky waters: is it correct to speak of wing-backs as the liberated full-backs who accompanied three central defenders (in a 5-3-2 for example)? Or, post-liberation, were they now wide midfielders in a 3-5-2?

Some will argue it makes little difference, but here again some elaboration is crucial, since there were some substantial differences, not only in terms of how they historically developed, but more broadly in the sense of how they represented a team's general approach. Was three (or five) at the back a statement of stinginess or a declaration of risk-laden attacking intent? We can enter into greater detail and pay greater heed to these nuances later on, but now let us turn the focus to Brazil and assume for simplicity sake, that a European's perception of 3-5-2 has been boiled down to two major strands.

These two varieties are not, as one might presume, 3-5-2 and 5-3-2: rather, they have more to do with the composition of the back three. Whether it is three defenders marking zonally, or two man-markers and a sweeper. Yes, the sweeper himself can be anything from another centre-back with good ball control and distribution to an erstwhile midfield playmaker (Beckenbauer, Schuster and Scirea).

In countries where the back four was firmly engrained, there were -largely confused- attempts to introduce 3-5-2 and wed it to the prevailing characteristics in place. Thus, across much of northern Europe, the tendency was to maintain the existing back four and simply add an extra centre-back. This was fairly perceived as a more cautious approach, as players who had been accumstomed to marking zonally found it difficult to reconcile themselves to man-marking, despite the cover of the spare man, and still the attacking width was being provided by those same full-backs from the preceding back four, which was not often given their more conservative nature.

Why not so often? Because under a back four system, and until the Brazilian-inspired trend would reach European shores a decade later, both full-backs did not attack simultaneously. While one advanced, the other would move closer to the centre-backs to form a back three strung across most of the width of the pitch. This was intended to be reciprocal, with each full-back deputising for the other. Moreover, a keen defensive positional sense was a must of both players most notably when both were to align themselves tightly on either side of their centre-backs when around their own 18-yard area. In other words, neither full-back was autonomous of the defensive line. None of these requirements apply to a wing-back. Indeed, it is a common criticism modern attacking full-backs in a line of four that they are postionally suspect. Equally, the charge levelled at more conservative full-backs such as Gary Neville is that while they are terrifically solid players and great defenders of wide areas they are insuffuciently thrusting in attack for the requirements of today's game (-my own personal solution for the supposed obsolecence of such players is that they necessarily become natural candidates for wide-defender roles in a back three, but that's a matter for another day).

Suffice to say, that a plausible defintion of a pure wing-back is that of an wide player equally disposed to defence and attack (and who is not a winger, let that be clear!) and who lacks the positional nous of a full-back. In other words, he is a pure wide-midfielder. An attacking full-back is a wing-back with the positional nous (and usually the height) required to play in a back four.

In many British teams who converted to 3-5-2, another, different kind of difficulty arose; the existing inclination towards zonal marking extended to the newly constituted back three. Even the Balkan and later Italian variants (particularly 3-4-1-2), in which the system was zonal, were still predicated on the opponent using two strikers, and so redundancy was one tactical adjustment away if the opposition simply used one less forward than there were spare defenders. Hence 4-3-3, 4-1-4-1, 4-2-3-1 and the like could render the 3-5-2 unbalanced, overmanned (in defence) and even undermanned (in midfield) to varying degrees.

In Brazil, these deficiencies have yet to be faced, mainly because most teams play with two forwards of some description and wingers have yet to make a reappearance. But another explanation for this persistence lies in the increasing complexity of some Brazilian formations, particularly 3-5-2. Asymmetry is common, often intentionally so (as was the case under the famously lopsided zona mista which was both a back four and a back three at the same time) with many coaches simply laying out the formation in such a way as to account for their players' natural tendencies. In attempting to redress inherited defensive problems, Brazilian coaches who adapted 3-5-2 conversely found new opportunites for offensive play unfold before them. An example of this is Flamengo and São Paulo sides of recent years.

Deliberate unbalance: how this can make a team more solid

As mentioned before, some Brazilian takes on 3-5-2 are asymmetrical and therein evocative of gioco all'italiana, albeit clearly without a sweeper system. Take one of the cagier Brazilian sides like Sao Paulo, for instance: coach Muricy Ramalho was forever tweaking his side and so well rehearsed were they, that the same starting 11 could play as a back four or a back three without any recourse to for substitutions. They would typically play with three centre-backs, a full-back, a wing-back, two holding midfielders, an attacking midfielder (stationed wide), and two forwards.

In 2006, Ramalho inherited a solid side at a club already reknowned for its institutional stability behind the scenes. He was loathe to risk changing the team's style: defending deeply and springing counter-attacks, and also using a target man upfield as an "out-ball". They would go on to win three consecutive championships. But where Muricy did lend a distinctinctive touch was in the arrangement of his players on the field.

The whole system was one of weights and balances so that hardly any defensive position would be vacated without another player filling it in. The midfield behaved almost like a chain - much the same way a classic back four tends to. So, for example, how to explain the gaps behind the wing-back on the left - who would cover for him? As shown below in Figure 1, that would be one of the volantes (defensive midfielders) who would then move out to the flank. But in doing so, he would abandon his post just in front of the back three and so the right-wing back (who was often a full-back or even a volante by trade) would tuck into the centre to become the screening presence there.

Normallly the attacking midfielder (meia), who was stationed to the left would move infield thus leaving room for the left wing-back to overlap. But equally, the meia could stay relatively wide and thus allow the wing-back to launch a diagonal run infield. So we are left with a two man left-sided midfield who alternated their inside and outside movements. Who would help the advancing right-back when the weight of the side was tilted towards the left? That would be the job of the second striker who would drop out wide to assist. As can be seen, there was no symmetry to Ramalho's system but there was an intensely-laboured synchronisation about the collective movement.

Freeing the wing-backs to advance infield

In the case of a more attack-minded team, we have the example of Flamengo, and we see a more interesting use of the the third centre-back, in this case a conservative full-back or a tenacious volante. This formation was not borne out of a desire to shore up the centre-backs, but rather to maximise the dynamism of the wing-backs.

Below left: Flamengo's more adventurous 3-5-2 (or 3-1-4-2).

In Leo Moura and Juan, coach Cuca had at his disposal two bonafide wing-backs whom he did not trust to perform in a back four. The solution lay in an interesting back three consisting of two centre-backs plus a conservative full-back, with a defensive midfielder just ahead plugging whichever spaces were most vulnerable. The central midfield was occupied by the dynamic box-to- box player Kleberson with a genuine playmaker Ibson playing in a slightly more advanced position. Cuca's system gave both wing-backs (now defacto wide-midfielders) plenty of licence to cut inside, not just in rapid counter-attacks, but also to participate in passages of patient build up play. Both are expected to double back and assist the quasi-full backs (the centre backs who push out to defend the flanks).

In this aspect at least, the system echoes England's variation at Euro '96 whereby Terry Venables played Neville and Pearce as wide defenders in a three, with Ince dropping in from midfield and with wide midfielders ( two from Anderton and McMannaman or Le Saux ) pushing on. In effect we are left with a back three that can behave like a back four without resorting to a designated sweeper. In Flamengo's set-up, we see a novel use of the third centre-back. By sucking the opposing markers into the centre, Leo Moura leaves space for the pseudo centre-back behind him to bomb ahead as a surprise element to the Flamengo attack as illustrated in Figure 2.

Emphasis on defense: wing-backs not trusted to defend in a back four.

Below: Palmeiras line-up in a 3-5-2 (which leans towards 3-4-1-2 when in possession and 3-4-2-1 when not) despite the coach's misgivings about the system.

It must be said however, that in complete contrast to Flamengo, there are many Brazilian teams who use 3-5-2 for the purposes of playing an overtly defensive game, hoping to hit teams on the counter-attack. The most puzzling exponent of this (in that he is hardly amongst the more defensive of Brazilian coaches) the Palmeiras side coached by, until recently, Wanderley Luxemburgo. Palmeiras tend to play very deep and adopt 3-4-2-1 in defensive phase (sometimes even dropping the second-striker for another 'meia' as a more cautious approach.) Paradoxically, Luxemburgo is a big critic of three-at-the-back but offers as his defense a lack of trust in the ability of his wing-backs to defend properly- amongst the most common of the justifications cited by European coaches who are reluctant to switch to a back three. Luxemburgo laments that Brazilian defences have lost the chain effect whereby if one full-back advances, both centre-backs shift across and the other full-back tucks in tightly defend as a temporary centre-back.

Another such coach who is reticent about 3-5-2 is current Grêmio boss Paulo Autuori, for whom the 3-4-1-2, as practiced by the side he took command of in May 2009, is prejudicial to midfield creativity. This is because, Autuori claims, (and though I do not profess to understand his reasoning) one of his two meias (from his preferred 4-2-2-2 system) finds himself obliged to drop deeper into midfield, or else be replaced by a second volante. Evidently the presence of a third central defender is not enough comfort for even this more liberal of Brazilian coaches. Which again leads to that inescapable point; that there is no getting away from this.. the besottment of most modern Brazilian coaches with destructive volantes.

A future post will deal with this and analyse the possible ways out for those coaches who want to keep a solid midfield base but without inhibiting decent circulation of the ball from the back.


  1. great piece roberticus! I remember that at some time on the guardian you said something also about internacional's system but anyway, guess we leave it for another day.
    I had long discussions with people on 3 or 4 defenders. It's funny when you talk with someone who claims to be a fan of Bielsa but then disses systems with 3 defenders. Likewise, trying to explain why playing with 3 defenders à la 1990 with Kohler, Buchwald and Augenthaler against only 1 real center forward or even 0 makes no sense, because 1 player constantly sits on his ass.
    Tactical system as history teaches come up in an evolutionary process, like everything else. Given the necessities of time and all, at some point someone comes up with a system and if it is succesful, it's adapted by many more. It works the same way in nature and with replication. if there's a strategy that is succesful, then agents simply copy it. It's not a big deal and no system is old school and redundant, just not useful.
    Your article points out well the reasons and use for 3-5-2 emerging in Brazil. I personally have the belief that even in Europe we will see hopefully a reemergence of 3-4-3 systems, in the 2-3-2-2-1 form, means 2 center backs, one if you like Kaiser, the other marker, 2 wingbacks supported by a deep lying 5 or 6 like say Busquets or more a risky version with someone like Pirlo, 2 interiors, 2 mediopuntas, 1 forward. It's kind of like the Cruyff Dream Team system, or even to a degree the Van Gaal one, just that you would play mediopuntas and not wingers like Zenden. The 5 or 6 would be sort of a sweeper in front of the defense like Sammer used to be, just that he is more an advanced center back.
    That would be an interesting development, because with that system you could somehow have it easier for both wingbacks to be a bit more attacking and make life difficult for both opposing mediopuntas or wingers.
    At least I believe that if any attacking/ball controlling team likes to add any innovation to the current sets, it should go for this.It would be an interesting system for a team like Argentina, lacking these days decent fullbacks, but filled with centerbacks, 5 and mediopuntas.

  2. Nice article Roberticus, I have to absorb it and then hopefully I can make an interesting comment.

    initially i would echo Ulague's comment that I feel sometimes these formations are adopted because they are the "latest rage". A manager adopting a system and at the same time complaining about using said system strikes me as surreal.

    I personally don't care for 3-5-2, I always preferred to have 3 players naturally occupying the forward positions of attack. I'll leave it at that for now. Well done.

  3. This comment has been removed by the author.

  4. Hi Uluguer!

    thanks for adding your observations.

    I guess the key for any back three to survive in the modern game is to avoid, as you say that spare player sitting on his arse.

    I think that in the Van Gaal system this can be remedied by at least two of the three (preferably the wider ones) being quasi-full-backs, like Abidal and Puyol, and by the back line shifting across the width of the field in a chain-like motion (as does the traditional back four). So say, Abidal goes out wide to seal his left flank, then Marquez and Puyol swerve into the centre. And since they'd still have the holding mid just in front of them as a screen...this would be no different than playing an adventurous back four (say in Van Gaal's 2-3-2-3) in which both wing-backs attack and the holding mid drops closer to the two centre-backs.

  5. Hey Charlitos (alias Tego, isn't it??)

    Thanks for checking this out.

    Your point about the paradox of managers who play systems they don't even trust is one that, similarly , perplexes me.

    Luxemburgo clearly recognises what's wrong (ie that most Brazilian full-backs are effectively wing-backs nowx and so can't tuck in and defend around the bo) but instead of mentoring some young full-back and developing him into a more all-rounded defensive player (or likewise, trying to convert a mobile centre-back, who mightn't be tall enough to defend centrally, into a solid full-back) he acquiesces in the system.

    Luxemburgo's inconsistency here can be held in light of other coaches who prefer 4-2-2-2 yet with those same wing-backs; Luxemburgo might argue in his defense that as a counter-balance they end up having to place two destroyers in that deep midfield role, and that he would prefer to avoid this for the purposes of maintaining a fluid midfield. But here again he undermines his own argument by playing two relatively defensive midfielders anyway - in front of his back three. At times, and especially in Libertadores matches, he has dropped the second striker, and added an extra creative midfielder (from 3-4-1-2 to 3-4-2-1) or simply just moved his playmaker into the hole and replaced that midfield slot with another box-to-box player (ending up with 3-5-1-1).

    I'm hoping to write a future post on Luxemburgo's peculiar take on Brazilian tactical history (in which he places himself firmly in the pantheon of greats, of course!)

  6. Please note, the above deleted comment was mine.

  7. I am pretty sure I have read it somewhere that Guardiola really wants to play a back 3, with a ball playing libero a la Koeman.
    I hope to see him or somebody try it.

    My guess for his idea is like this

    Messi-------Eto'o (Villa)---Henry

    My only worry is it may be narrow down the left but Barca already were this year anyway, or maybe they are looking at options for a left wing left midfield??? I would have sworn I read something about Capel? I don't think he is of Barca quality but Silva would be fantastic, although I would like to see him at Liverpool as well, going to be an interesting season. That being said all of this Valencia raiding talk, if Valencia could hold their team together and not have so much background noise with a couple good buys they could have a very interesting and competitive team.

  8. Steve,

    You seem to have included one player too many in your line up there.

    If those rumours point to a libero being employed, then I should imagine Guardiola is looking to the 3-1-2-1-3 of the Van Gaal/Cruijff school as an option.

    But what I've heard being touted is the prospect of Pep switching to a 4-3-1-2, at least as a plan B.

  9. Roberticus,
    I did mess that up... Not sure how either!

    No excuses here!

    would definitely be narrow down the right sans Henry but I am thinking the line up I saw must have looked something like this then.
    Alves-----Xavi-----Iniesta Messi-----Etoo-----Henry

  10. Olá Roberticus, tudo bem?

    Vou escrever em português mesmo para me expressar melhor. Depois eu traduzo. Acho que o Flamengo tem o sistema mais dinâmico do futebol brasileiro atualmente. Com a saída do Ibson, o time perde muito, mas ainda mantém um meio-campo forte. Normalmente, o técnico Cuca coloca dois volantes se revezando na função de terceiro zagueiro. Ora Toró, ora Willians executam a função. Assim, o time preenche o meio-campo e, ao mesmo tempo, tem três zagueiros para defender quando é atacado. Juan e Léo Moura têm muita liberdade para entrar em diagonal pelo meio e poucas obrigações defensivas. Até que esse 3-5-2 eu defendo. Dinâmico, com volantes que sabem tocar bem a bola e sair para o jogo e com alas que vão à linha de fundo e entram em diagonal como autênticos meias. O ataque do time é que, mesmo com a presença de Adriano, ainda está devendo. O que não dá pra aguentar são alas que não apoiam, volantes limitados tecnicamente e três zagueiros parados lá atrás.

    Abraços, Marcelo Costa.

  11. Oi Marcelo,

    tudo beleza?

    Concordo com você respeito a Flamengo; pois meu argumento nesse artigo é isso; que o 3-5-2 não é para nada aquele monolito retratado por seus críticos. Acho também que o Luxemburgo é pertenece a este último grupo; para ele 3-5-2 é sempre um 5-3-2 disfarçado; mas tal como você apunta, o esquema de Cuca e bem arriscado e innovadora.

    Abraços; sigo passando pelo teu blog.