From 1966 to now: the evolution of World Cup football

When an estimated 400 million people tuned in to the 1966 World Cup final between England and West Germany at Wembley, many of them watching a game live on television for the first time, it changed the way we remember football.

Before the 1954 tournament, World Cups weren’t televised at all. If you wanted to know how teams played, you bought a newspaper the next day and read a column about it by some cigar-chomping sportswriter in a three-piece suit. Now fans could follow along for themselves in fuzzy black and white: And here comes Hurst, he’s got — some people are on the pitch, they think it’s all over … It is now!

But watching every game of a World Cup yourself can be a pretty hefty time commitment, as a billion exasperated spouses every four years can attest. And as for actually remembering individual matches you watched even one tournament ago, let alone four or five? Forget it.

Luckily, the arrival of football as a televised sport had another, unanticipated side effect: it made it possible to turn the game into data.

Opta, the data provider whose detailed records of on-ball events are behind most football stats you’ve seen, has gone back through decades of video and logged every World Cup game from 1966 on, touch by touch, the same way they cover a weekend’s Premier League action.

At last, we can savour the most famous moments in football history the way they were meant to be experienced… as a bunch of rows in a spreadsheet.

No, for real, this is cool.

You’ve seen clips of goals. You’ve read ghostwritten autobiographies and tactical histories the size of a team bus. But unless you’re blessed with an eidetic memory of that Yugoslavia vs United Arab Emirates match in Bologna you watched in your Ninja Turtles pyjamas at age five, there’s been no objective way to measure the many ways the game has changed.

Except now, armed with over half a century of data, we can give it a try.

Here, era by era, chart by chart, Total Football passing network by Pele possession value per 90 stat, is The Athletic’s data history of the past 14 World Cups…

---

The long ball era: 1966-70

That Wembley final people tuned in to watch didn’t look a whole lot like football as we know it today (although if you poke around in the undergrowth it’s not hard to find it flourishing in a vestigial form known to scientists by its Latin name, T. Pulis).

The passes back then were so long. A full 21 per cent of the pass attempts by England and West Germany in the 1966 final travelled at least 20 yards up the pitch. For comparison, only five games by any team hit that mark at the most recent World Cup four years ago, and three of them were Iran blazing out of the group stage.

Another thing back then was that, with the ball always sailing overhead, players weren’t always super strict about what we might call “positions.” Pop quiz — what formation did England play in that final?

OK, a quick word on some nerdery. The visualisation above is called a passing network, and it shows players’ median positions during the longest stretch of the game where their side didn’t make a substitution — which given subs weren’t introduced at the World Cup until 1970 meant all 120 minutes of this final.

The size of players’ circles corresponds to how many passes they attempted, and the thickness of a line between players indicates how often one player passed to another (at a rate of at least five passes per 90 minutes to earn a line).

See how Jack Charlton is just sort of floating around on his lonely red dot at the bottom, unmoored from the team? That was normal for a centre-back in the long-ball era. Don’t worry, the game’s going to come to you in a few decades, buddy.

The colours of the lines and circles are based on a metric called “possession value,” which really just means “goal probability”. When a player passes the ball somewhere that makes their team more likely to score, the change in scoring percentages is the value of the pass.

Possession value metrics come in a lot of flavours, some more sophisticated than others, but the one we’re using is based on a simple grid that tells you how often the average (modern) team goes on to score from a given location. It’s a ball-progression stat. Pass the ball closer to the other team’s goal and you’ll earn some possession value. Pass it backwards and you get nothing — good day, sir.

As a model of the game, that’s a little crude, right? And yet it’s more or less how football worked at the time. During the 1966 World Cup, teams kicked the ball forward seven yards for every yard they passed backwards. By 2018, that ratio had sunk to less than three to one. “A pass backwards does not mean fear but the start of another, better move,” Pep Guardiola once wrote. A generation earlier, that kind of touchy-feely talk might have got him sacked.

You can see from the pass sonar above that passes angled straight ahead have become less common (shorter slices at 11 and one o’clock) and a lot less long (bluish-green instead of bright yellow). Maybe more importantly, the sonar shows that teams in the long-ball era were reluctant to play sideways. There is a reason most completed passes in the modern game travel across the pitch, not up and down it — but we’ll get to that.

The point is, football is always changing. Sometimes those changes are driven by tactical trends, sometimes by the laws of the game, and sometimes by other, weirder stuff, like, say, choosing to stage a World Cup in the desert outside Doha.

One of the stories we football analytics dorks like to tell about how we’ve changed the game goes something like this: around 15 years ago, Opta and other companies began collecting event data, and in the last decade analysts have used it to popularise a metric called expected goals that said, basically, “Shoot from closer to goal, dummies”. Gradually, in league after league, Philistines heard the good news of statistics and lo, it came to pass that players did shoot from closer to goal.

That may all be true, but World Cup data makes clear that the great shot shift was going on a long time before anybody had heard of xG. In 1970, 62 per cent of shots came from outside the penalty area. By 2006, that was down to 54 per cent. Fifty years ago, it took around 15 non-penalty shots to score a goal — nowadays it’s closer to one in 10.

Shooting was especially selective at the last couple of tournaments, it’s true, but long-term trends make it fair to wonder how much data really had to do with it. What was it stats people are always saying about correlation and causation?

Every once in a while, the data can turn up ways that football hasn’t changed that much after all.

In 1970, Brazil demolished Italy so beautifully in the final that one newspaper compared their achievement to the Moon landing.

They did it with short passing (less than 10 per cent of Brazil’s passes were long balls) and circulation (fewer than five forward passing yards for every one backwards). Each backward pass wasn’t a sign of fear but the start of another, better move — except for the one where Clodoaldo inexplicably tried a no-look backheel in his own half and gifted Italy their goal. That was a worse, but very Brazilian, move.

Brazil still took 19 of 27 shots from outside the box, and their disconnected centre-backs are a relic of the long-ball era.

Still, the team’s clearly defined shape, with its narrow wingers, well-connected midfield and attacking right back, looks a lot like a contemporary 4-2-3-1. To watch footage of the game is to discover a strange thing both ancient and modern, like a first-gen iPod stuffed in the back of a drawer. Maybe it can still play that Nelly album?

🇧🇷Carlos Alberto & @cbf_Futebol , Mexico City, 1970 😍

🔬 One of the greatest #WorldCup goals of all time goes under the microscope in the first of our new series ‘Anatomy of a Goal’, brought to you by @Hyundai_Global pic.twitter.com/g3nEXjhTVy

— FIFA World Cup (@FIFAWorldCup) February 25, 2018

---

The midfield era: 1974-82

Look, these ‘eras’ are made up. Nobody drew a bright line between one World Cup and the next and decided that everyone would play football a different way now. But some time between Bobby Moore lobbing artillery shells towards Geoff Hurst and Gerson conducting the Brazilian midfield like prime Xavi one day would, the tournament’s whole vibe started to shift.

It sounds dumb, but one way to track how the game is played is to measure where it’s played.

In 1966, the middle of a football pitch was a yawning abyss, a desolate landscape for the ball to gaze down on from its aisle seat at 30,000 feet (maybe the ball is listening to a first-gen iPod). Stare into a ’60s midfield and the midfield stares back at you.

By the late 1970s, the heatmap had flipped completely. Instead of lobbing the ball from end to end, teams played straight up the gut, barely even bothering to use the wings.

What changed? Maybe attacking teams noticed they were being given an awful lot of acreage in some prime real estate and it would be a shame not to do something with it.

Between 1966 and 2018, the average time between receiving a pass and playing one fell by almost 20 per cent, from 3.1 seconds to 2.6. More time on the ball at early tournaments meant more space, and ball carriers in the 1960s and ’70s took advantage by gobbling up free yardage before a defender closed them down. It’s the first thing you’ll notice if you flip on an old game — ‘Why is everyone standing around off the ball like they are Paris Saint-Germain’s front three?’.

Teams gradually realized that the most effective way to play through midfield wasn’t just to dribble the ball but to set up a structured passing game there, the way Brazil had. In 1974, Rinus Michels’ famous Netherlands side gave the Selecao a taste of their own medicine.

The wingers were wider, in the Dutch style, but the midfield pair linking with a creative forward dropping in — Johan Cruyff playing something like Pele’s ‘ponta da lanca’ (spearhead) role — gave the side a strong central triangle to build around.

See how the centre-backs are getting involved in the more vertically-compact passing game, not just with long balls to the wingers but with patient two-way passing around the back? If Brazil’s brilliance hadn’t finally convinced the world to leave the long ball behind, Total Football soon did.

---

The defensive era: 1986-90

Michels’ teams left another legacy besides pretty passing triangles: the offside trap. As defences scrambled to shore up the midfield against opponents who actually wanted to play there now, they learned to compress the space the way the Dutch did, by bringing their defence closer together and using the linesmen to their advantage.

Offside calls rose steeply enough in the 1970s and ’80s that by 1990 the laws of the game were changed to allow an attacker who was level with the last defender to be judged onside. That only seemed to encourage more gambling on both sides of the line in the short term, as offside calls spiked to a record 8.5 per game at Italia 90, but things eventually settled down.

It wasn’t just the linesman’s flags that made that era dull. Now that they weren’t being coached to knock the ball forward at all costs, teams discovered they really liked passing backwards to the guy who was allowed to pick the ball up and boot it wherever he wanted.

Backpasses to the goalkeeper were safe. They were a good way to waste time when you needed to, and one team or another always needed to. They were awful.

When officials saw that the 1990 World Cup averaged just 2.2 goals per game, down from 2.8 eight years earlier, they banned the backpass — a change that would be transformative. By the next World Cup, passes to the goalkeeper had fallen by almost two-thirds.

The decade before those two rule changes isn’t exactly famous for its sparkling football. The great team of the 1980s, Diego Maradona’s Argentina, abandoned the style and structure of the ’70s and became, well, pretty much just Diego Maradona’s Argentina. When opponents can squeeze your passing game with an offside trap and play keep-away with their goalkeeper, maybe dribbling through midfield isn’t such a bad strategy.

---

The transition era: 1994-2006

Remember that touches heatmap? Let’s take a second to gawk at the white-hot flurry of midfield activity in the 1994 World Cup.

Without the safety of backpasses to bail them out, teams had to learn how to play forward again, and they had to do it against the latest in anti-midfield technology: Arrigo Sacchi’s zonal 4-4-2 press.

Part of the reason attackers enjoyed so much time and space in earlier decades was that even when defences did want to press, which wasn’t that often, the orthodox way to do it was man-to-man. A polite way to characterise that kind of defending is “high risk, high reward”. Another way is how former Netherlands and Austria manager Ernst Happel put it: “If you mark man-to-man, you’re sending out 11 donkeys.”

Without careful coordination, man-marking schemes can get stretched and ripped open in all kinds of unfortunate ways. Sacchi’s Italy defended differently, in organised banks of four that would look a lot more familiar to modern fans. As other teams caught on, the effectiveness of World Cup pressing surged and pass completion rates plunged from 80 per cent in 1986 to 75 per cent in 2002 — the lowest since 1966.

That dip owed something to more aggressive attacking play (vertical yards per pass rose steadily from 1982 to 1998 and 2002) and lots and lots of crosses (32 per cent of the goals at the 2002 tournament came within three actions of a cross, by far the highest rate in the data). You remember the old 4-4-2 with natural wingers, right? It was probably the last time football felt distinctly different from how it is now.

This was the period when football got hooked on transitions: teams pressed harder and attacked faster than before, trying to catch their opponents off-guard. The quick, choppy action produced a lot of set-piece goals, especially from free kicks (lately, corners and penalties have been more important).

It was also a transition era in another sense, as the game started to morph into something more modern.

By the middle of this century’s first decade, that creative second forward who had been dropping to link with midfield since Pele and Cruyff’s time finally became a full-time attacking midfielder to outnumber the flat 4-4-2’s central midfield pair. Soon, wingers were creeping inside, too, increasingly playing on their inverted side so they could cut in and curl shots from distance with their stronger foot.

Portugal were one of many teams at the 2006 World Cup that ditched the old 4-4-2 for a 4-2-3-1 or 4-3-3. They had three men in midfield, not two, and Cristiano Ronaldo sometimes played as an inverted goal-scoring left winger opposite Luis Figo’s traditional crossing winger on the right.

Things were starting to look familiar.

---

The circulation era: 2010-18

It’s not hard to peg the team who kicked off our current era. From 2008 to 2012, Spain were basically unbeatable, and they did it by doubling down on all the long-term trends that had brought us to this point.

Instead of playing with a third central midfielder, they played with four or five of them, stashing extra midfielders on the wings, at striker and possibly in the overhead baggage compartment on the way to South Africa (these were pretty small guys).

Spain completed more passes than almost anyone ever had (as of 2010, their 525 per game was the second-highest on record) and attempted fewer long balls than just about anyone (7.7 per cent, at the time the second lowest). They played fast (2.5 seconds between passes) and short (16.7 yards per pass). They were Guardiola’s — OK, Vicente del Bosque’s, but let’s be real about who most of these players spent their workdays with between international get-togethers — perfect circulating machine.

The point of all that passing wasn’t just to rack up high scores in Opta’s new event data, although people did talk a lot about possession percentages back then.

Getting really good at passing sideways and backwards in order to go forward made Spain difficult for opponents to press, which made it easier to bring all 10 outfield players together into the opponent’s half. Overlapping full-backs could provide width in the attack while wingers overloaded the top of the box and centre-backs pushed up to establish a high line.

And once they got set up like that, it was practically impossible to take the ball away from them and hang onto it for very long.

The sharp increase in successful counter-pressing at the World Cup, defined here as winning the ball back within eight seconds in the highest 60 per cent of the pitch, probably gets at the real advantage of the circulation era. But it also reflects a calculated response to so many teams trying to play like Spain: other, weaker teams did the opposite.

Instead of wasting energy pressing top possession sides and leaving space between their defensive lines, opponents decided to sit deep and stay as compact as possible, hoping to hit the space in-behind on the counter. Even some very talented teams surveyed the stylistic landscape and figured they’d have a better chance without the ball.

And so we come full circle to a very weird 2018 World Cup final, where France played direct (19 per cent of their attempts were long balls) and relied on goals from set pieces (a free kick and a penalty) to get them going and generally looked like they were performing some kind of existentialist theatre about the futility of midfield on, like, a spiritual level.

Croatia-1-1.png” alt=”” width=”1500″ height=”2550″ />

Their passing network was a cry for help. Their non-penalty expected goals number that day against Croatia was a pitiful 0.3. And none of that mattered because, guess what, they won. Deal with it, nerds.

---

The leaderboards

No World Cup data review would be complete without an excuse to sit around and name old sportsmen and just have the best time.

The leaderboard below is mostly based on possession value, which is far from perfect for reasons we’ve already talked about (it’s based on simple averages, it’s not era-adjusted and it can’t see what’s happening off the ball) but also generally better than just counting things. It helps us value ball progression, which is how teams create (and avoid conceding) goals.

The three stats at the bottom — successful take-ons, tackles and interceptions — are included because they don’t really progress the ball but they do change goal probabilities in a way that’s more difficult to value. We’ll settle for counting those.

If that’s too many stats to digest all at once, how about a simple scatterplot? The “on-ball value” vertical axis below adds up the kinds of ball progression a guy can do on the ball (passing, crossing, carrying) and plots them against how he can help his team off the ball (receiving passes).

The player who’s done the most of all of it combined is, by a mile, Cruyff.

  

Is it fair to compare crude ball progression stats across eras with wildly different standards of defending? No, probably not, but this exercise is not about justice. This is about remembering some guys.

Remember Mesut Ozil and Alexis Sanchez, how they got their names dragged through the mud when they came to play their club football in England? Those dudes were golden gods at the World Cup.

Remember Pierre Littbarski? Ariel Ortega? If not, you’re probably clicking over to YouTube right now to find out if they really were better than Pele and Maradona, then coming back scratching your head to read the explanation of how possession value works again (hint: we’re not penalising unsuccessful actions here).

Finally, there are the teams.

This may not look flattering to teams such as Canada and New Zealand who sit at the bottom of the cumulative goal-difference colour scale, but just remember that means they did better than all those blank white countries by making it to the tournament in the first place.

As for the green end of the scale, it won’t blow anyone’s mind to learn that western Europe and South America have historically been the most successful regions at the World Cup, or that the Brazilians and the Germans, the two countries who have been to the most finals, have the best goal difference (the poor trophyless Dutch are just behind).

You want an overall champ? OK, fine.

Since 1966, Brazil’s goal difference is an astonishing plus 1.01 per game. But you know what they say, it’s a simple game.

Twenty-two men, one ball, 90 minutes, and at the end the Germans (+1.03 GD) win on 50 years of cumulative goals.

(Data visualisations: John Muller/Photos: Getty Images/Top image design: Sam Richardson)