Manchester United 1-1 Fulham: Cunha's 89th-Minute Equaliser and What the Scoreboard Doesn't Say

In the 89th minute, Matheus Cunha collected the ball more than twenty metres from Bernd Leno's goal. His strike was not clean; the ball struck a Fulham defender's leg, changed direction and settled in the net. Craven Cottage fell silent for a beat before the away section erupted. Michael Carrick rose from the bench. A point rescued at the 89th second — and if you read only the scoreboard, you would tell a story about a narrow escape. I watched the whole match with my tracking notebook open, and what I recorded does not look like that. This was matchday five of the 2026-26 Premier League season. Fulham, managed by Álvaro Arbeloa, hosted Michael Carrick's Manchester United at Craven Cottage. The hosts arrived with a single point from four matches, without a win, sitting near the bottom of the table. The visitors had four points from four games and were under pressure from their own stands. Seven years ago I began building a standard tracking sheet for every match: expected goals, shot counts, possessions entering the final 25 metres, passes into the box, and PPDA — the number of opponent passes allowed before a defensive action. PPDA is not a measure of spirit; it is a measure of honesty in pressing. A team that claims to press aggressively must prove it with numbers, not with post-match quotes. Fulham set up in a 4-2-3-1 with two deep central midfielders, conceding the middle third inside a medium block. Manchester United held 58% of the ball in the first half yet created no clear chance from open play in the opening 25 minutes. The ball circulated laterally between the centre-backs and full-backs and was stopped at the edge of the final 25 metres. I call that sterile control, and I have spent years separating it from dangerous control. The turning point came in the 63rd minute. Senne Lammens parried a Fulham shot but could not control the rebound. The ball spilled in front of goal, inside the zone where every centre-back is coached to be decisive. Lisandro Martínez lunged to intervene, and his touch sent the ball into his own net. Fulham led 1-0 in a manner no coach wants. I replayed that phase seven times. The own goal was not a random individual error. It was the consequence of a measurable chain: United's midfield had been leaving gaps in the inside channel for ten minutes; the defensive line had dropped an average of four metres deeper; and once the block was compressed, every rebound became a grenade inside the box. Lisandro Martínez was simply the last man to touch it. More telling was the reaction. After conceding, United raised their tempo for about ten minutes, pushing their PPDA from 9.4 to 12.1. Then it fell away. Between the 75th and 85th minutes their PPDA hit 14.8 — meaning Fulham were allowed close to fifteen passes before being engaged. For a side chasing a game, that is the number of fatigue or of disorganisation, and usually both. Bernd Leno kept the old script from collapsing. The Fulham goalkeeper produced at least five significant saves, three of them from attempts my model valued above 0.15 expected goals each. He did exactly the job of a keeper whose team holds only 42% of the ball: stop what can be stopped, and accept that the rest belongs to probability. At the other end, Senne Lammens left a question mark over his handling of rebounds. It is a point I have tracked since the start of the season: the reflexes are good, but the rate at which he pushes the ball back into dangerous areas remains high. The phase leading to the own goal was the clearest example. A good goalkeeper does not only save; he redirects the point of landing. Saving is reflex. Redirecting is thinking. In the 89th minute, Matheus Cunha changed everything. His shot from outside the box carried an initial expected-goals value of roughly 0.06 — the kind of chance nobody builds a plan around. The ball struck a Fulham defender, changed trajectory, and turned a near-harmless effort into an unstoppable goal. The score finished 1-1. When the stadium falls silent, we hear the voice of probability most clearly. Here I must place the two goals side by side. The opener was an own goal. The equaliser was a deflected long-range shot. Both belong to the category of outlier events, the kind probability models assign a low frequency. The conclusion is not that the match was good or bad, but that it carries little information about the true quality of either side. A 1-1 scoreline is too small a data sample to conclude anything. According to my live tracking model, Manchester United generated about 1.62 expected goals across the match, Fulham about 1.08. Split the game into twenty-minute blocks and the picture changes. Between the 45th and 75th minutes — the phase that decided the match — United produced roughly 0.14 expected goals. Most of their cumulative output came in the final twenty minutes, when Fulham had dropped deep to protect the lead and conceded control. United recorded 31 possessions entering the final 25 metres; Fulham recorded 24. But passes into the box were reversed: Fulham 11, United 9. The team with more of the ball delivered less of it into danger. My dangerous-control index — possessions entering the final 25 metres per 100 possessions — gave United 11.4, below the range of 15 to 18 I usually record for sides competing for Champions League places. This is where I have to step away from the scoreboard. Five points from five matches is exactly one point per game. In the Premier League that rate corresponds to the bottom half, not to the European places. Numbers never lie; only the people reading them deceive themselves. And the most common self-deception here is reading an 89th-minute equaliser as evidence of character. The character to equalise in the 89th minute is real; I do not deny it. But character is not an index. It does not explain why United allowed Fulham nearly fifteen passes per possession in the closing ten minutes. It does not explain why a team chasing a game let the opponent dictate tempo for most of the second half. Bias is a match played without data. I choose to bet on the number. Fulham, conversely, should not be too hard on themselves. Two points from five matches, second from bottom, still without a win. But my data shows they created enough to have earned more than two points. Their problem is not organisation; it is the quality of the final ball and the ability to finish when the opponent is stretched. This match was their season in miniature: control during the decisive phase, then a single outlier moment taking two points away. I still track this kind of team the way I tracked Italy at a major tournament a few years ago: by measuring how far each possession travels toward the opponent's goal, not by the possession percentage on the big screen. Fulham were structurally far better than their second-from-bottom position suggests. Structure usually arrives before results. For Manchester United, the problem is not the defence, even though the goal conceded was an own goal. It is the ability to create quality chances against a settled block. A side that wants Champions League football must generate 15 to 18 on the dangerous-control index, not 11.4. That number cannot be compensated by a deflected long shot in the 89th minute, nor by two weeks of training. Both teams shared the same blind spot: they judge matches by results. Fulham see two points dropped. United see one point gained. The truth sits in between, inside the twenty-minute data blocks nobody replays on television. For United, the signal to watch next matchday is their PPDA over the opening thirty minutes. If it stays around 13 to 15, that points to a structural problem rather than a bad day. For Fulham, the signal is passes into the box. If they hold above 10 while maintaining their defensive structure, points will come, because structure always pays dividends before results do. I do not predict football. I only describe probability before it happens. And the probability at Craven Cottage said that two teams shared a point neither truly controlled. The question left behind is not who deserved what, but this: when a goal in the 89th minute is a team's only goal of the second half, is that character — or merely the echo of a deflection? Assumptions and latency: the indices in this article come from my live tracking and modelling, with an estimated error of about 0.15 expected goals per match. PPDA data should be updated after matchday six, since pressing structures usually stabilise only after roughly six rounds. No conclusion here should be read as a fixed forecast before the sample is large enough.