By the late 1930s, the prevailing tactical orthodoxy in Brazilian football was still the 2-3-5, a system that elsewhere had been either tweaked or jettisoned in favour of the W-M, the metodo or even the Swiss verrou by the middle of the decade.

This was rooted in a Brazilian preoccupation with preserving the sanctity of the centre-half's midfield organisational role. Even those Brazilians who strayed from 2-3-5 were loath to copy Herbert Chapman's vision of a 'third-back' as an outright stopper, having seen how it led to the wing-halves retreating. Instead they advocated something closer to Vittorio Pozzo's innovation of a deep-lying centre-half, only with a difference; whereas Pozzo's Italian centre-half was a midfield screen (often a technically limited one at that), the Brazilian centre-back was a disciplined holder who was also free to advance with the ball when in possession — a libero, in the most creative sense of the word.

To the south, Uruguay had long been linked with the fortunes of Italian football, partly because of its substantial immigrant population from Italy, some of them players who would seek their fortune in the Italian league before returning to Montevideo as coaches. Later waves of Italian immigration included those already steeped in the values of the game as interpreted in Italy, amounting to a substantial and sustained flow of footballing ideas across the Atlantic.

Little wonder then that in a world in which many dutifully imitated Herbert Chapman's W-M, Uruguayan teams rushed to adopt the Italian metodo or variants thereof. For the Uruguayans, the metodo offered a framework for the associative, possession-based football that dominated its public's affections insofar as it retained the No. 5, the classical centre-half. Nevertheless, and in adherence to Pozzo's blueprint, the Uruguayan centre-half became very much a wing-half in assignment — that is, his game became increasingly one of vigilance and enforcement. And so the creative void left in the midfield was filled — to an extent— by the retreat of the inside-forwards, who became advanced midfielders rather than support-strikers as they had been before. 

Uruguayan ideas would find a receptive audience in Brazil's southernmost state, Rio Grande do Sul, which just over a century previously had formed a fluctuating border between the Spanish and, Portuguese empires and in which cultural traits were imbued with the gaucho tradition of the plains. These cultural affinities transferred also to football, as the local journalist Fabrício Carpinejar acknowledged. "Gaucho (southern Brazilian) football has always had a platense temperament," he said, something that extended to the terraces. Fans there adopted Argentinian-style chants and supported the team for 90 minutes, demanding players should "sweat blood for the shirt"; more mainstream Brazilian fans demanded primarily to be entertained. A southern Brazilian club, 14 de Julho, from the frontier city of Livramento, was even one of the founder members of the first national Uruguayan football league in 1902. 

That influence coloured the whole perception of football in the south and by the late 1930s, southern teams such as Juventude, Guarany and Riograndense were fielding metodo - style centre -halves as a defensive midfield base. Whatever the direct Uruguayan influence on southern Brazilian football, though, its scope for exercising influence on the rest of the country was limited by the geographic isolation of the south from the national heartland around Rio de Janeiro and São Paulo. The national team rarely included players who represented clubs outside the Carioca and Paulista regions.

Uruguay had arguably been the best side in the world in the twenties and thirties, winning two Olympic golds and the first World Cup. They didn't travel to Europe for the 1934 and 1938 World Cups — both won by Italy — and the Second World War denied them the opportunity to re-establish their status. Wartime neutrality brought Uruguay and Argentina domestic prosperity propitiated by healthy exports, while the prospect of conscription into Mussolini's army was a powerful factor in dissuading ambitious men who would previously have emigrated to Europe from abandoning the River Plate estuary. 

Those conditions soon led to a thriving domestic game, and pushed the Uruguay national team to take its ebullient game on tour, principally through South America. Crowds were enticed by the allure of the Olympic celeste and the chance to view a side that, with Argentina, had dominated the Copa América. 

Brazilians were particularly taken by the opportunity to upstage the Uruguayans in the Rio Branco Cup, which had got off the ground in 1931 as one of a series of small-format competitions among South American nations. It would gain some degree of prestige as the listlessness of neutrality in the South Atlantic wore on during the 1940s. Two editions, both single-game finals and hosted in Montevideo, had been held by the time it was hosted in Rio de Janeiro in 1940, this time in a two-legged format. Further editions would be held in 1946 and 1947, the latter contested over a four-game schedule.

Those excursions across the Brazilian border represented a vital injection of footballing novelty at a time when most in Brazil still took their lead from the British, or British-inspired coaches. It was indicative of the British influence at the time that most technical terms tended to be approximations of the English rather than coinages with a Portuguese root: a full-back was a beque, for instance, and a left-half a half-esquerdo. But with the advent of the Second World War and the Empire's maritime and engineering manpower channelled into more pressing matters overseas, an opportunity was created for other influences to take hold of the Brazilian imagination.

It did not escape the attention of many in Brazilian football that the Uruguayans had somehow eschewed Chapman's third-back game and simultaneously demonstrated their defensive robustness, all while upholding their commitment to elaborate football. Would it be possible, they must have pondered, to strike such a balance but along Brazilian lines?

The Hungarian Dori Kürschner arrived in Rio de Janeiro in March 1937 on the understanding that he was to be hired by Fluminense, a team that was based in the salubrious Laranjeiras district. Fluminense enjoyed the backing of the aristocratic Rodrigues family which included politicians and newspaper barons. However, for reasons still disputed today, the Flamengo president José Bastos Padilha persuaded Kürschner to sign for his club. (It's rumoured that the Rodrigues heir, Mário Filho, was sympathetic to Flamengo despite his family's footballing loyalties and, presumably scenting the winds of populism, was keen to mine Flamengo's broader support base).

If Padilha had hoped to add a gleam of European sophistication to Brazilian spontaneity, Kürschner certainly delivered; he introduced rigorous individualised training programmes and, more controversially, a version of the third-back game. That innovation did not impress the boot-room stalwart and assistant manager Flávio Costa, who himself had played as a creative centre-half.

Their partnership may have been short-lived, but the year in which Kürschner and Costa coincided on the Flamengo bench represented a synthesis of sorts, a confluence of two or even three schools of football. There was the Brazilian 2-3-5, which often featured a retreating centre-forward. There was the third-back game that Kürschner had brought from Grasshoppers of Zurich. The third strand was born of a dispute between Kürschner and Flamengo's star player, Fausto Maravilha, and it would result in a quintessentially Italo-Uruguayan style compromise. 

Kürschner insisted that the creative centre-half be converted into a third-back, but Fausto, commanding the affection of the terraces, refused to back down. He was supported by the composer and celebrated radio commentator Ary Barroso, who resented the perceived slight towards a player he admired. Barroso was an eccentric, attention-seeking character who would greet Flamengo goals in games on which he was commentating by performing ditties on his harmonica. A fixture among the Flamengo entourage, his political appetite was such that it caused Mário Filho the embarrassment of having to publish an op-ed piece in his Jornal dos Sports in which he disassociated himself from any involvement with Barroso in a plot aimed at bringing down Kürschner.

No matter how much Kürschner insisted that Fausto would still be able to bring the ball out from defence and initiate attacks, effectively as a libero, the sceptics countered that he would be dragged all over the pitch by the sort of unorthodox strikers common in Brazil. Kürschner's assistant, Costa, voiced his fears that, shorn of Fausto for prolonged periods, Flamengo's midfield would be outnumbered. His authority weakened by the Fausto dispute, Kürschner eventually lost the backing of the newly-elected club president Raul Dias Gonçalves and was sacked in November 1938 to make way for Costa.

Not that Kürschner's ideas were completely discredited in Brazil. Reporting his dismissal, the Jornal dos Sports acknowledged that "an almost ferocious campaign had been orchestrated against this great coach… with his name highly valued in Europe" and noted that Kürschner had been the author of "Flamengo's finest phase in these few years since [the advent of] professionalism".

Gentil Cardoso's Bonsucesso in 1931

Whether Kürschner was the first coach to introduce a third-back to Brazil is debatable. Gentil Cardoso had introduced an unorthodox scheme as early as 1931 with Sírio Libanês and later Bonsucesso. Having seen the W-M on his trips to Europe while working in the merchant navy, he intended to install Chapman's theory, but ultimately ended up with a much more fluid third-back game whereby the centre-back Eurigo, originally a hardy wing-half, and the right-back Heitor would alternate in stepping out to join Cláudio and Corinheiro in the half-back line. As Jonathan Wilson pointed out in Inverting the Pyramid, Cardoso's ideas were largely rejected by Rio's elite clubs, partly because he was of mixed race, but he at least began to prepare the ground.

In the end, the revolution was led by a reluctant pioneer. Having replaced Kürschner on an interim basis, Flávio Costa introduced his diagonal, an odd halfway house between 2-3-5 and W-M. He claimed in 1950 that "modern football permits no more innovations", but that was to ignore the variable of player movement. His rival Zezé Moreira, who had similarly enjoyed a playing career at centre-half, was more cautious — not necessarily in terms of his commitment to expansive football, but rather in recognising the dangers of neglecting an opponent's variations. "In football, there are no mysteries," Zezé said, "but it is necessary to plan out the game." For Zezé, it was essential to build a platform of play that would be holistic enough to adjust to the variables presented by different opponents.

Both men manifested their philosophies in the organising of their respective teams. Considering his willingness to tweak in the face of unexpected movement from opponents, it's perhaps curious that Zezé, ahead of his time, favoured a zonal system of marking which would later find full and triumphant expression in 1958, albeit in a 4-2-4 form. It followed for Moreira that one should not exhaust oneself chasing the elusive tail of a rival team that could present any permutation of approach play, and instead look for a proactive and not reactive mechanism that would assure the most comprehensive array of protection against all threats.

Costa, on the other hand, seemed to believe that tactical scheming had reached its zenith with the 2-3-5, and since most opponents were likely to play within that same structure, it behoved his own team to make only the slightest of adjustments to counter what would at most be minor and foreseeable alterations on the part of the opposing manager. In other words, the only way to improve on the opponents' 2-3-5 was to evolve within the 2-3-5.

In an attempt to achieve this, Costa maintained elements of man-marking, particularly with regard to the opposition centre-forward, but he also devoted torturous hours to analysis and theorising, devising ways he might accommodate "…qualities and characteristics particular to the Brazilian player". The result, developed at Flamengo in 1941, was the diagonal. "By 1945," said Carlos Alberto Parreira, who coached Brazil to World Cup success in 1994, "we apparently felt secure in the system".

For others of Costa's disposition, however, that still did not address his prime objection to the third-back; namely that the centre-half as an organisational figure was too far removed from the zone of influence in the midfield. To remedy that, Costa needed to reinforce the playmaking element of midfield but by another source.

In a concession to the advantage that the third-back system afforded the full-back line by its addition of a third body, Costa experimented with a pivoting system based on the centre-half, whereby one of the wing-halves would retreat to cover his respective flank, enabling the remaining two defenders to shift across cover centrally. This makeshift back three was still vulnerable to being turned along the other flank by a long crossfield ball, but Costa was equally concerned with the resultant benefits in midfield and attack.

The mechanism of Flávio Costa's diagonal system

The inverse of this backtracking movement was that the wing-half on the other side of the pivot would push forward and inward, taking on something of a box-to-box role. This had ramifications for the inside-forwards. To compensate for the withdrawal of the half-back who was usually behind him, the inside-forward along that side dropped closer to the midfield, as happened in the Italo-Uruguayan metodo and W-M formations, in order to assist the playmaking, while the other inside-forward was left free to continue his improvisational game closer to, or as part of, the forward line. Costa himself was adamant that this later player was much more a forward than an attacking midfielder; the 'ponta da lança' or 'spearhead', as he was known in the language of the terraces, was a "goal-man" in the words of the coach. 

Costa thus had a central pivot, a dynamic shuttler and an advanced playmaker in the midfield to support a four-man attack. His was a system of compensations and balances; it was also one of adjustments. He professed that football could not be further transformed structurally, but what he could control was to be found at a micro level within the corpus. According to Hilário Franco Junior, of the University of São Paulo, Costa's approach would see him tweak his system to meet certain opponents so that, for instance, if a rival team boasted a left-winger deemed more dangerous than his right-sided equivalent, the diagonal would pivot upon the centromédio and the remaining half-back would drop back to guard the runs of the outside-left in question. In general, however, the system was aligned in accordance with the characteristics of the players at his disposal, and by the late forties it had become the default within Brazilian football for those who did not favour Moreira's more mechanised system.

Zezé Moreira's zonally-marking 3-3-4 at Botafogo in 1949

Moreira, feeling that he needed the defensive security of Kürschner's third-back game and yet recognising that the elusive dribbling and interchange play of Brazilian forward lines threatened to pull defenders out of position, needed to reach a compromise. By 1949 at Botafogo he had devised a solution: a zonally-marking back three, with two holding midfielders screening ahead, and a withdrawn inside-forward forming a playmaking link with the front four. Curiously, it was around the same time that this 3-2-1-4 was also being cultivated in Hungary by the Honvéd team of Gustav Sebes, having followed a different developmental path.

Perhaps it is the diagonal system that helps explain divergence between Brazilian and Argentinian definitions of exactly what constitutes a No. 10, that most romanticised of positions. In the Brazilian popular imagination, there was a clear division between the duties of the two inside-forwards.

The journalist Marcelo Barreto has suggested that it is easier to define the No. 10 by what he is not. He is not the sort of roaming midfield general typified by Gerson in the 1970 World Cup-winning side; he occupies a more advanced area and spends considerable periods of the game with his back to goal. Gerson, conversely, sat deeper and was more of a box-to-box playmaker. In both the diagonal and later the 4-2-4 formations this cerebral figure was generally the No. 8 (the more withdrawn of the inside-forwards, which over time became the second central midfielder). Neither is the true No. 10 a support striker playing off a front man; he must be more involved with the construction of attacks. 

Alhough Pelé would become the iconic No. 10, his role was very much that of an auxiliary striker. In fact, he had caught the attention of Santos when playing as outside-left for Bauru. Of the 1970 team it has often been said that it featured five No. 10s, though in truth the most recognisable trequartista of the quintet was probably Rivelino, who in terms of average position and playing characteristics occupied a place on the spectrum somewhere between Gerson's No. 8 and Pelé's No. 10 roles: it is that blend of qualities that is distilled in the Argentinian figure of the enganche, or the Spanish mediapunta

Nowadays, even Brazilians are happy to confer No. 10 status on enganche-type players such as Paulo Henrique Ganso of Santos, or as Barreto suggests, on any attacking player whose creativity seems to flourish when given freedom of movement: Neymar, for example. Even Ronaldo, on account of his irrepressible individuality, was spoken of as a No.10.

A further consequence of the diagonal's quasi-third-back game was the emergence of the archetypal Brazilian attacking full-back. Since the third back was a wing-half in origin, he infused his new withdrawn role with a restless, locomotive quality and so players such as Nilton Santos and Djalma Santos would need little encouragement to convert the full-back positions of the 4-2-4 into a dual-function role.

In the early fifties, Martim Francisco Ribeiro de Andrade, the coach of Vila Nova from the small town of Nova Lima, about 20 miles from Belo Horizonte, formed the first recognisable 4-2-4 by withdrawing a player (usually the more defensively robust of the pair) away from the halfback line, thus gaining a fourth defender, who was in effect the second centre-back. The Paraguayan Manuel Fleitas Solich repeated the formula in 1953 with Flamengo and used it to win three consecutive Rio State titles. The Hungarian Béla Guttman, already attuned to the possibilities of a back four by developments in the Danubian game, then adopted the system after being appointed at São Paulo in 1956. There he was assisted by Vicente Feola. By the time Feola took the reins of the national team in 1958, the 4-2-4 had already overtaken the diagonal as the most widespread formation in Brazil.

One of the main obstacles to the acceptance of W-M amongst Brazilian players was its man-marking system: the reality of two teams, invariably aligned 3-2-2-3 when facing each other, meant that marking duties became symmetrical and reciprocal but not reactive; the system was not equipped to counter unorthodox positional play. In that respect it was different to the man-marking systems of the verrou and catenaccio, in which players were expected to mark certain opponents regardless of how those opponents were distributed or adjusted their collective movement on the pitch.

In the space of two decades, the tactical aspect of Brazilian football would go through remarkable transformations. If we consider the shift from 2-3-5 of the late thirties to the 4-2-4 that Brazil used in winning the 1958 World Cup, then both Costa and Zezé could be said to have devised variations on a 3-3-4, a sort of midway point on the march towards 4-2-4.

The interpretation is reductive, yet compelling. From a distant vantage point it appears, on a macro scale and conveniently overlooking the minutiae of circumstance, that football was destined to evolve from 2-3-5 through to 4-2-4 with all its attendant midway stages, its growing pains. The men who helped to get it there were as often fleeing from the confines of orthodoxy as they were moving towards a clearly stated finality. None could profess a clear idea as to what lay ahead in 1958, though countless advocates would retrospectively claim credit for having instituted a 4-2-4. 

There are those who argue the Magical Maygyars of the early fifties invented it, although the notebooks of their coach, Gusztav Sebes, show he saw József Zakariás, the putative fourth defender, as a deep-lying wing-half. Hungary's shape, if not their style, was rather the 3-2-1-4 that emerged in Brazil in the forties. Then again, could the Brazilians possibly have formalised the sum of their experiments into 4-2-4 without the vital contribution of Hungarian coaches such as Dori Kürschner and Béla Guttman, who were perhaps freed from convention by the environment in which they found themselves? Could Zezé Moreira, Flávio Costa, Martim Francisco and Fleitas Solich have imposed their respective visions had they had the opportunity to coach in Europe? 

Such mavericks and improvisers needed a meeting point at which to rally. Brazil, with its narrative of ethnic miscegenation, and for which decades idealised itself politically as an austral United States, a land of opportunity and renewal despite the constrictive legacy of its Portuguese-inherited feudalism, provided the perfect ground for experimentation.