Semenyo's Brace and the 0.14 xG: What the Data Table Missed at Manchester City

Minute 43. The ball broke out of the Sunderland penalty area, spinning backwards, bouncing once before it touched the grass again. On the television feed, Antoine Semenyo did not rush in. He stood still for about half a beat — long enough for the Sunderland defender to turn his head, short enough for nobody to close the angle. Then he planted his standing leg, pulled back his right foot and placed the ball beside the far post. The scoreboard flicked from 2-2 to 3-2.

I watched that sequence seven times, frame by frame. There is nothing spectacular in it. No bicycle kick, no thirty-yard thunderbolt. Just a man choosing to wait for half a beat that most Premier League footballers no longer have the patience to wait for. That half beat is this entire article.

Fourteen minutes later he scored his second. This time from a loose touch by Sunderland right on the edge of the box, a hard strike that stretched the lead to 4-2. Sitting between those two goals was an assist for Rayan Cherki. Three moments, one Sunday afternoon on 20 September 2026, in the Premier League.

To read this performance properly, you have to go back to the state of the match before it happened. Sunderland travelled with a deep defensive block, their two lines no more than ten metres apart, willing to concede the ball and wait for the counter. Manchester City held possession, shifting it from flank to flank, and for most of the first half those rotations opened nothing of consequence. By minute 43 the score was still 2-2.

What stands out is this: the goal that unlocked the game did not come from a rehearsed combination. It came from a loose ball, from a situation no coach can draw on a tactics board. Sunderland defended exactly as planned for forty-two minutes, then collapsed in a moment the plan did not cover.

Semenyo is not the kind of attacker who grew up on loose balls. He is normally described in terms of pace, dribbling and surges down the wing. But what I observed in him here was something else entirely: the patience of a predator who knows the chance will come, and knows precisely where he must stand when it does.

Statistically, his performance is recorded as two goals and one assist. His Expected Goals for the match was 0.14. His Expected Assists was 0.17. His Expected Goal Involvement total was 0.31.

To make sense of those numbers, they have to be placed beside what actually happened. An Expected Goals figure of 0.14 means that, based on position, angle, ball speed and the pressure applied by defenders, the model valued the entire set of Semenyo's shots in this match at roughly one seventh of a goal. He scored two.

That is a multiple overperformance. For those who read football through a spreadsheet, it is a signal of luck, of an afternoon when everything fell into place. For those who read football with their eyes, it is a signal of a skill the spreadsheet has not yet learned to measure.

Semenyo's real value lies in second balls — the kind every modern Expected Goals model systematically misprices.

Start with the shot in minute 43. The ball was bouncing on grass. That is the hardest kind of ball to strike cleanly, because the contact point shifts from instant to instant. The usual solution is to hit it up, accepting the risk of skying it. Semenyo chose the harder route: keeping it low, finding the far corner, forcing the Sunderland goalkeeper to move across his line before he could dive. That technique appears in no model.

The second goal in minute 57 was more about reading the game. When the Sunderland defender miscontrolled the ball on the edge of the box, Semenyo had already moved before the mistake happened. He did not react to the error; he anticipated it. The distance between those two things is the entire difference between a player who chases the game and a player who stands one step ahead of it.

The assist for Cherki follows the same logic. An Expected Assists figure of 0.17 means that pass, according to the model, was unlikely to become a goal. It became one not because the model is wrong, but because the model prices the pass and cannot price the tempo and the drop point the passer chose. A pass only becomes dangerous when the receiver is in the right place at the right moment, and that does not live in the passer's data.

This is where the numbers reveal their limits. Modern football has become very good at measuring the volume and quality of chances. It remains very poor at measuring preparation. No model captures a player standing still for half a beat to watch the flight of the ball before he strikes it.

The concept of the second ball, put simply, describes situations in which order has already dissolved. A shot comes off the post, a ball ricochets off a defender's shin, a ball drops after a duel. In those moments a defensive system stops being a system and becomes a collection of individuals trying to rebuild order. Whoever reacts fastest wins.

Based on my experience covering matches in England across several seasons, I keep noticing the same pattern: players who live on second balls are rarely rated highly by models, yet they are the ones who decide the most deadlocked games. They do not create chances. They pick up the chances everyone else discards.

Something also needs to be said about Semenyo's position, because it speaks to a wider trend in European football. Over the past fifteen years the traditional winger has been almost wiped out. Academies now produce wingers who know one direction only: receive in the half-space, turn inside, shoot with the stronger foot. That uniformity makes defences easier to read than ever, and it also makes players who can go both ways far more valuable than the market prices them.

Semenyo does not sit entirely inside that template. He still goes outside when required, still uses his left foot when required, still accepts touches that look less elegant. That unpredictability is what put him in positions the Sunderland defence did not anticipate — outside the box in minute 43, on the edge of the box in minute 57.

Tactics are prose, moments are poetry — and a match is where the two swallow each other whole.

After minute 57, Sunderland could no longer touch the shape of the game. That tells part of the story too: the visitors had prepared meticulously for an orderly match, and they only lost it when that order was broken by an individual who worked better than anyone else inside the chaos.

Semenyo was voted Player of the Match in the official Premier League poll, with thirty-three per cent of the votes. It is a deserved award. But thirty-three per cent also means sixty-seven per cent of voters picked someone else. A performance with two goals and one assist still left two thirds of the room sceptical.

The pitch never lies — only the storyteller knows how to hide his loneliness behind every goal. And the loneliness of a player who lives on second balls is a very particular kind: he exists only in the moments when somebody else makes a mistake, and is otherwise almost invisible.

Read closely, that scepticism is not unreasonable. A player who generates value from loose balls depends on two things outside his control: the opponent's error and a teammate's touch. When both arrive, he is the hero. When they do not, he vanishes into the crowd without leaving a meaningful statistical trace.

That is why an Expected Goals figure of 0.14 has the right to exist. It is not lying. It is warning that afternoons like this one are hard to reproduce mechanically. But football is not decided in the middle of a distribution. Matches are decided in the tail — in the minutes when order collapses and somebody has to be the one who clears up.

There are talents that never appear in a ranking — they hide in the eyes of someone who believes in what has not happened yet.

Intuition is a dead star whose light is still travelling — and I choose to stand beneath that sky and receive it. When I watched Semenyo waiting outside the box in minute 43, I did not see a player gambling. I saw a man who had finished his calculations before the ball reached him.

What I want to track over the coming matchdays is not Semenyo's goal tally. It is the situations in which he chooses to wait, the second balls he claims, the movements that happen before the opponent's mistake takes shape. If the model keeps undervaluing those moments, then the problem lies with the model, not with the player.

Half a beat of waiting before the shot. That is what every data table will skip, and the only thing I want to keep from the afternoon of 20 September.