Courtois, Philippe Albert and the Data Gap in the 'Best Goalkeeper on the Planet' Debate

In the notebook on Real Madrid I have kept since the 2026-19 season, there is one page I have never closed. It tracks Thibaut Courtois's appearances, updated by hand after every La Liga round, next to an empty column labelled PSxG. That column stayed empty for months, because at the time I could not find a clean post-shot data source for the Spanish league.

This week, reading Goal.com's interview with Philippe Albert, I reopened that page. Albert — a former Newcastle United and Fulham centre-back, one of the names attached to Belgium's golden generation — said Courtois remains the best goalkeeper on the planet. He added a line I read three times: “Over the last seven, eight years, no one is better than him.”

I believe him. And I have nothing to prove it with.

That is the paradox I want to write about. The Goal.com piece places Courtois on the scales against David Raya, Alisson Becker, Gianluigi Donnarumma and several others. But when I stripped the article down paragraph by paragraph, looking for a single performance metric — any metric — to place beside that claim, I found none. No save percentage. No PSxG. No crosses claimed. No pass completion under pressure. Not even a penalty record.

All the piece has is a debut age, transfer fees, appearance counts and a trophy list. That is career data. Current level is a different question, and it requires a different kind of evidence.

So I sat down, reopened the spreadsheet, and did what I have done for ten years whenever a large claim appears without support: I checked what the claim is standing on.

A weighty claim resting on a foundation with no measurement in it

Philippe Albert is not an outsider. He played in the Premier League in an era when the goalkeeping position had not yet been dissected by data, he faced some of Europe's best forwards, and he came from a football culture that has produced a remarkable number of good goalkeepers. When a man like that speaks about a compatriot goalkeeper, his words carry real professional weight.

But professional weight and evidential weight are two different things. And I need to separate them clearly, because that is exactly where most football arguments go wrong.

Albert's statement belongs to a category I call the authority claim. It rests on the speaker having watched an enormous amount of football, having been inside a professional environment, and having a basis for comparison with his own eyes. Such claims have value. They are a form of compression: thousands of hours of observation packed into a sentence no one could ever turn into a table.

But they carry a structural weakness: they cannot be verified, cannot be refuted, and cannot be broken into components that can be argued separately. When Albert says “no one is better”, we do not know what better means here. Reflex shot-stopping? Box command? Passing under pressure? Reading crosses? Organising a back line? Or simply presence — the sense of security a team feels when it looks back at its own goal?

Each of those components can be measured by something different. And if we do not separate them, we will argue forever without anyone being wrong.

This is why I always start by listing what can be measured. For goalkeepers, that list is now fairly long: PSxG (post-shot expected goals), PSxG± (the gap between goals conceded and expected goals conceded), cross-claim rate inside the six-yard box, cross-claim rate inside the penalty area, pass completion under high pressure, successful sweeper actions outside the box, penalty save rate, and average time on the ball before distribution.

Not one of those appears in the Goal.com article. That is a notable gap, because Goal.com is not an outlet short of data resources. They could have placed Courtois's PSxG beside Raya's and Alisson's. They chose not to. Perhaps because the format is an interview, and interviews privilege words over tables. But the consequence stands: readers receive a ranking conclusion without receiving any basis on which to judge it.

An article published after the 2026 World Cup, and why that changes everything

There is one detail I have to reconstruct before going further, because it fixes the entire timeline.

In the piece, Ederson is described as a “former Manchester City star”. Unai Simón is referenced as having “become a World Cup winner with Spain”. And there is a ten-man shortlist for the 2026 Yashin Trophy.

Those three signals, placed together, are consistent with only one moment: the second half of 2026, after the 2026 World Cup had finished. Which means Courtois — born 11 May 2026 — was 34 when the article was published.

Why does that matter so much? Because the entire “best on the planet” argument changes in nature once the subject turns 34. At 28, the question is whether he is at his peak. At 34, the question is how far he is from his peak, and what that distance means for the way his team plays.

I once believed in absolute numbers, until a World Cup taught me that emotion is a variable too. That lesson came in 2026, when I was seventeen, a schoolboy in Madrid with a homemade spreadsheet, betting a friend that Spain would beat Russia 3-0 in the round of sixteen. Spain had 75 percent of the ball, completed hundreds of passes, and lost on penalties on the hosts' own ground. When I pulled the data, they had generated roughly 0.7 xG from twenty shots. A dismal figure. From that day I understood that data only means something in the right context, and that what is obvious in front of your eyes is often the least informative thing available.

At 34, Courtois is still an elite goalkeeper. But elite and best are different tiers, and the gap between them is exactly what an article with no data will slide straight past.

Ten names on a list, and two names left off it

This is where the article contradicts itself, and I want to be blunt about it because it is my most important finding.

The opening frames Courtois against David Raya, Alisson Becker and Gianluigi Donnarumma. The sentence structure implies a direct comparison between four leading goalkeepers. The reader is invited into a picture with four figures in it.

Then comes the ten-man shortlist for the 2026 Yashin Trophy. Alisson Becker is not on it. Donnarumma is not on it. Matvey Safonov is. Vozinha is.

I am not arguing that Safonov or Vozinha are undeserving. I am arguing about the article's logic. If the two central figures of the comparison frame are absent from the official shortlist the article uses as its reference point, then one of two things must be true: either the award's selection criteria diverge sharply from expert consensus, or the article built its own comparison frame without checking it against reality.

Both possibilities weaken the premise of being “ranked against”. You cannot rank a man against people who are not in the same measurement system.

To me this is a familiar professional error, one I have seen many times in ten years of following European sports journalism: the headline and framing are written by one person, the body is assembled by another, and nobody cross-checks. The result is a text shaped like a comparison with no spine of a comparison.

Data does not give answers; it surfaces the questions we are brave enough to ask. The question here is: if we genuinely want to know who the best goalkeeper is, what do we need to measure, and how do we measure it without fooling ourselves?

Why raw save percentage is the biggest trap in every goalkeeper comparison

Suppose the article had given us save percentage. Would that solve the problem? No. And this is the technical part I want to spend time on, because it explains why almost every public goalkeeper debate goes nowhere.

A goalkeeper behind a possession-dominant defence faces fewer shots than a counterpart at a mid-table club. But the shots he does face are usually of higher quality, because they arrive from situations where the defence has been fully stretched, from counters, from moments when the team has accepted high risk to push forward. In other words: fewer shots, each more dangerous.

In that case, a good goalkeeper's raw save percentage can be lower than a mid-table goalkeeper's, even though the good one prevents more goals in probability terms. This is why PSxG± became the standard: it compares actual goals conceded with the goals an average goalkeeper would concede from exactly those shots. It separates shot quality from goalkeeper quality.

I learned this lesson the hard way during the empty-stadium season of 2026. I was a remote intern at a small sports data company in Madrid, tasked with comparing Real Madrid's home performance before and after crowds returned. The result made me rewrite the report three times: with empty stands, Real averaged 1.9 goals per game; with crowds