The Vanishing Touchline Winger: Reading Football's Homogenisation Through PPDA, xG and Transfer Fees

Minute 63, the ball goes out to the left flank. The winger collects it tight against the touchline, some 35 metres from goal. In front of him is a two-metre corridor along the line, completely empty. The opposing full-back has dropped inside, the centre-back has shifted across to cover. A winger from the 2000s would have attacked the line, pushed the ball toward the byline, and crossed. This winger takes the ball with the inside of his left foot, turns half a circle, and goes the other way.

I noted the moment in my notebook. It was the seventh time in the match he had refused the byline. The eighth came on 78 minutes, when his team were a goal down and needed a cross more than anything else. He cut inside again, laid the ball to a central midfielder, and the attack died at the edge of the box.

That night I sat down with the data. Across the match, the team played 19 passes into the penalty area and attempted only two open-play crosses. Fifteen years ago, in the same game state, that number would have been 11 or 12. I did not need more evidence to believe I was watching a disappearance rather than a poor performance.

Do not rush to trust a number before it has told the story from the beginning.

How I count, and why the method matters more than the number

I began following football in 2026, when the first data pages were still typed by hand by part-time students. Twenty-eight years later, my job sits between two worlds: the numbers broadcasters put on screen, and the raw data that clubs buy for six figures a season.

The gap between those two worlds is wider than people assume. Broadcast data is designed to tell a story in 90 minutes. Internal data is designed to answer a question across 38 rounds. When I say the touchline winger is vanishing, I am not speaking from feeling. I am speaking from three families of traceable indicators.

The first family concerns space. I record the average position where a winger receives the ball, on both the horizontal and vertical axes. A traditional winger receives around 88 to 92 percent of pitch width, hugging the line. An inverted winger receives around 70 to 78 percent. That twenty-percent band, multiplied by thousands of touches a season, is a continental-scale migration.

The second family concerns pressure. PPDA, the passes allowed per defensive action, tells you whether a team presses hard or lazily. It also tells you where a team is willing to let the opponent play. High-pressing sides force opponents wide, where space is tight and sideways passes are blocked. Low-pressing sides invite them inside. Both choices push the winger infield.

The third family concerns value. I compare transfer fees against actual output, normalised into expected goals and assists per 90 minutes. This is where I found what I consider the most important part of the whole story.

But before the numbers, one methodological point. Every dataset I use has a source. Every conclusion I draw carries a note on the data's limitations. No exceptions, even when I am writing for a mass-market outlet whose readers do not care about the footnote.

The arithmetic of the inside cut

An inverted winger creates three measurable advantages, and all three are geometric rather than technical.

The first is shooting angle. Hugging the touchline, a player's shooting line to goal forms an acute angle, usually under 20 degrees. Cut inside to the edge of the box and that angle opens to 40 or 55 degrees. In a model I built for a club in the Chinese top flight in 2026, I calculated that for the same player, same shot power, same goalkeeper, moving the shot location from 90 percent of pitch width to 72 percent raised the scoring probability from 0.031 to 0.087. Nearly three times. No player and no coach can ignore that ratio.

The second is passing options. A winger on the line has two worthwhile directions: down the line to cross, or back to the full-back. A winger in the inside channel has five: shoot, slip a through ball, switch play, combine with a central midfielder, or lay it back. More options mean a lower turnover probability, because defenders cannot anticipate. This is why the pass completion of elite wingers has risen steadily for fifteen years while crossing accuracy has fallen.

The third advantage, rarely discussed, is vertical stretching. An inverted winger forces the opposing full-back to choose: follow inside and vacate the corridor, or hold the corridor and leave him free. Most modern full-backs choose the second and rely on a holding midfielder to cover. The defensive line is stretched into two tiers, and space opens in the middle.

Add the three together and the conclusion is unavoidable: in a system that prizes efficiency over aesthetics, the inverted winger is the optimal choice. And when every team picks the same optimum, football starts to look the same everywhere.

When the full-back becomes the winger

There is a consequence most analysis skips. When every winger cuts inside, the flank does not disappear. It simply changes owner.

The new owner is the full-back. Over the past twelve years, the full-back role has changed more than any other on the pitch. They run further, cross more, and above all must decide faster. A full-back of the 1990s mainly defended. A full-back of the 2020s must do both, and in many systems must attack better than he defends.

I tested this with running data. In a sample of 240 Premier League matches I collected between 2026 and 2026, high-speed distance covered by full-backs rose by an average of 21 percent. Their touches in the final third rose 34 percent. Over the same period, touches in the final third by wingers fell 12 percent.

In other words, the work did not vanish. It was transferred from one player to another, and the new owner has to run further to do the same job.

This has a recruitment consequence that few clubs are prepared for. If you sign a full-back because he defends well, you are signing a player for a role that no longer exists. If you sign a full-back because he runs fast and passes well, you have an old-fashioned winger with defensive duties. Both arguments are coherent, but only one fits the present.

And here I return to the Vietnamese clubs I have followed most closely over the past three seasons. Most V.League sides still recruit full-backs on the old criteria: physique, duels, straight-line speed. Meanwhile, clubs in Thailand and Indonesia have shifted to the new criteria: long passing, running into the inside channel, one-twos. That gap does not show up in the table immediately. It shows up in qualifying rounds, when international football forces teams to play at a tempo twenty percent higher.

The cross is dying, and the byline with it

Let me be precise: the cross is not dead. It has lost its status as the default option.

In data I compiled from five major European leagues between 2026 and 2026, open-play crosses per match fell from around 22 to around 15. Their conversion rate into goals fell far less, from 2.4 percent to 2.1 percent. The cross remains nearly as effective as before; it is simply chosen less often.

This paradox took me a while to understand. If an option is still effective, why abandon it?

The answer lies in risk, not in effectiveness. A cross is a high-variance action. It can produce a goal, but more often it produces a counter-attack for the opponent, because the ball travels away from goal while many players have pushed up. A modern side, organised to minimise transition risk, chooses the lower-variance option.

When probability collapses, what remains is the nature of the match.

And the nature is this: modern football does not optimise for expected goals created. It optimises the difference between expected goals created and expected goals conceded. In such a calculation, an action worth 0.09 expected goals but also worth 0.07 to the opponent is less attractive than one worth 0.05 with nothing conceded.

The cross is the first type. The pass into the inside channel is the second. And so, without anyone announcing it, teams have collectively turned their backs on the byline.

This carries a coaching consequence I consider serious. If nobody crosses, nobody teaches crossing. If nobody teaches crossing, we will produce a generation of wingers who cannot use their weaker foot to deliver into the box. And one day, when a team needs exactly that skill to break a deep block, there will be nobody left to call.

The price of a profile

Here is an old story. In 2026 I analysed Hulk's transfer from Zenit Saint Petersburg to Shanghai SIPG for a reported 55 million euros, at the time one of the largest fees in Asian football history.

I built a simple model. I took his entire shooting record across four seasons in Russia, normalised it for the quality of chances he was served, and produced an expected-goal output. The result showed his real finishing output running some 40 percent below what Chinese media were projecting.

The article drew fierce backlash. But three scouts from three different clubs contacted me within two weeks asking for the full report. I learned something then that I still repeat to younger colleagues: accurate numbers find the people who need them.

I do not look at the price tag, I look at the signature of the money.

That signature, in this case, sat somewhere else entirely. The club did not pay 55 million euros because they believed he would score 30 goals a season. They paid because they needed a name that could sell tickets, sell shirts, and attract international media attention during a period when their league was trying to prove its status. The fee was a marketing cost, booked into the transfer account.

This is why I say the transfer race among big clubs is largely a brand arms race. Real sporting value sits elsewhere, in deals where money spent and output produced match each other.

Those deals tend to happen at small clubs. A small club cannot afford to pay for a brand, so it is forced to pay for ability. It cannot buy a player because he is famous. It must buy one because he solves a specific problem on the pitch.

In transfer data I have tracked since 2026, output per euro spent among clubs with revenue under 100 million euros runs about 2.3 times higher than among clubs with revenue above 500 million. That does not mean small clubs work better. It means big clubs are paying for things that are not on the pitch, and those things have a price.

The empty-stadium experiment

In 2026, when leagues paused and then returned in empty stadiums, I realised I had a natural experiment nobody could have designed.

I collected Premier League data from 2026 to 2026, 1,140 matches in total, and compared it with the post-lockdown run. I controlled for the usual variables: team quality, schedule, injuries. Home win rate fell from 46.2 percent to 38.4 percent. Average goals per match rose by 0.6. Average yellow cards fell slightly.

The stadium was empty, but data never lacked an audience.

The implications go far beyond home advantage. It showed that a substantial share of what we call club identity, of what we call home spirit, is a measurable psychological effect that can disappear. With crowds gone, refereeing decisions became more neutral, home players lost part of their drive, and away teams played with more confidence.

I sent a 40-page report to a club fighting relegation. They did not hire me for attacking analysis, which is my speciality. They hired me for set pieces, the only part of the match that does not depend on a crowd. A corner is taken the same way whether 60,000 people are watching or nobody is.

That was when I left my media pundit role to work directly with coaching staffs. It was also when my writing style changed: shorter, drier, fewer adjectives, because the reader was no longer a spectator but a coach and a scout.

A match lasts only 90 minutes, but its story outlives a season.

What the data does not see

Now I have to argue against myself. If the story of the vanishing winger were this simple, every team would look identical and there would be nothing worth watching. Reality is not like that, and the reason lies in the blind spots of data.

The first blind spot is pass quality. A ball into the box from wide is recorded identically whether it threaded three defenders or not. Models cannot see that some players deliver through gaps others cannot perceive. That skill has no metric. And when a skill has no metric, it is undervalued in the transfer market. That is a market inefficiency, and inefficiency always creates opportunity for those who can see it.

The second blind spot is match context. A cross on 85 minutes when your team is a goal down is worth something entirely different from a cross on 15 minutes. Most models I have worked with do not distinguish the two. As a result, the traditional winger is undervalued, because he appears most often exactly when the match needs him most.

The third blind spot, and the one I consider most serious, is defensive value. An inverted winger produces more goals, but he also leaves the flank empty. When the team loses the ball in midfield, that flank becomes a motorway for the opponent. Attacking models do not price this in, because the cost is paid in another part of the pitch and at another moment of the match.

Exceptions, and why I refuse to write the same piece twice

There is a group of teams that still needs the traditional winger, and needs him rationally.

These are sides that sit deep and regularly face opponents with a well-organised block and a crowded midfield. Against such a block, the inside channel pass does not exist, because the inside channel is sealed. The only way to create space is to stretch horizontally, and the only way to stretch horizontally is to have a player on the touchline forcing the full-back to follow him.

In other words, the traditional winger has not disappeared. He has become a specialist tool, used in specific situations, like a defensive stopper or a free-throw specialist in basketball.

And precisely because he has become a specialist tool, his market value has fallen. That creates a paradox I consider the best opportunity in the transfer market right now: a good traditional winger, one who can change three or four matches a season, can be bought for the price of a substitute.

Three or four matches a season. In a 38-round campaign, where the gap between fourth and seventh is often four points, those three or four matches can be the entire season.

Correlation is not causation, and I nearly forgot it

I have to tell a mistake of mine, otherwise this piece becomes propaganda for data.

In 2026 I built a prediction model on PPDA and win rate. The correlation was strong: teams with PPDA under 8 won 61 percent of matches, teams above 12 won 34 percent. I presented this to a coaching staff and recommended they press more.

They did not follow it. They asked me a question I could not answer immediately: do low-PPDA teams press well, or do they press a lot because they have better players?

The answer is mostly the second. Strong teams tend to press more because they control the ball more, and controlling the ball more creates more pressing opportunities. PPDA is not a cause of winning. It is an accompanying marker.

Data never gets tired; only the people reading it do.

Since then I have set myself a rule: before recommending any change based on an indicator, I must answer the question about the direction of causality. If I cannot, I label the indicator a marker rather than a cause, and I recommend no action.

This makes my reports longer and less exciting. It also makes them more accurate, and in this job accuracy matters more than excitement.

Limitations of the data I am using

An analysis without this section is not credible, so I will be explicit.

The touch and average-position data I use comes from two different providers, and the two define pitch zones differently. The discrepancy between systems in the same match can reach 3 percent for average position. For conclusions based on large distances, this is irrelevant. For conclusions based on small differences, it can reverse the result.

The PPDA data I use for Asian leagues is significantly lower quality than for European leagues. In some leagues, defensive actions are still logged manually, and the logger must make a subjective call about whether an action counts as a defensive action. This is what I call human-entry error, and it is more dangerous than technical error because it tends to be systematic rather than random.

The transfer data I use for value analysis is reported fees, not actual fees. Many deals include add-ons, performance payments, and undisclosed swap components. Every conclusion about output per euro therefore carries an error band I cannot quantify precisely. I always state this when presenting, even though it makes the number less seductive.

Signals for the next cycle

Over the past three seasons I have been tracking four signals I believe will shape the next phase.

The first is the return of the attacking full-back as a genuine winger rather than a defender pushed forward. Youth setups in Europe are now training full-backs the way wingers were trained twenty years ago: crossing, one-on-one duels, positioning inside the box. This reverses the logic of development, and it will take five to seven years to produce results.

The second is the specialisation of set pieces. As open play becomes harder to convert because blocks are better organised, the value of set pieces rises. And set pieces are the only part of the match that can be coached independently of talent. A team without the best players can still own the best corner routine.

The third is the shift of transfer money south and east. Clubs in Southeast Asia and the Middle East buy differently from European clubs: they buy for specific needs rather than for brands, often because they cannot afford brands. Over the next decade I expect some of the world's most efficient deals to come from this region.

The fourth is the return of spatial metrics. As positional data gets cheaper, clubs will start measuring what nobody measured a decade ago: distances between passes, the speed at which a block shifts, the time a defensive shape takes to recover after losing the ball. These metrics will reprice a great many players, in both directions.

If you see a monk in me, read the numbers like scripture.

But I do not want to end with a declaration about data. What I have learned in twenty-eight years, and especially in five years working directly with coaching staffs, is this: every model is wrong at some point, and the analyst's job is not to find the right model but to know when his own model is failing.

The touchline winger is vanishing. He will return, in a different shape, once defensive blocks become dense enough to seal the inside channel. And when he returns, there will be a generation of players nobody taught to cross, and a generation of scouts nobody taught to price that skill. That gap is where the opportunity sits.

History never repeats itself exactly, but it very often stumbles over old data.

What I am waiting for in the next round is not a beautiful goal. I am waiting for a full-back, on 80 minutes of a match his team needs to win, to run to the byline, lift his head, and cross. If that happens in a major league, and if it produces a goal, I will reopen my notebook and start counting again.

Note: This article is based on the author's data-analysis methods and professional observations. The figures cited serve to illustrate method, not to provide predictions or advice for any form of betting.