Why every wonderkid list undercounts how hard this is

This page answers two different questions and never lets them touch. The first is about real football, and it is answered with structure rather than statistics: there is not one percentage about the real world anywhere on this page, ours or anybody else's. The second is about a video game we make, and its numbers describe that game and nothing else. The section where the second one begins says so in its own heading.

And nobody is named. Not a player, from any list, in any year, and not the outlets that publish them. The subject here is a format. Most of the people that format ranks were children on the day it ranked them, and a page arguing that hype is unreliable does not get to make its case out of their names.

What a wonderkid list actually is

The shape is stable enough to describe from memory, because it barely changes between outlets or between years. A fixed number of players — fifty is the common size, though it varies — ranked from one downwards, every one of them under a birth-date cutoff, published in roughly the same month each year, with a mostly new cast each time. Some editions carry a note about who topped the previous ones.

Compiling one is real work. It means watching youth football that almost nobody watches, talking to people who will not be quoted, and committing in public to an opinion that is checkable later.

The word itself arrived in football sideways. "Wonderkid" is a term of art from football management games, where it means a young player the engine has secretly rated very highly, and it leaked out of the games into football writing, where it now means roughly "a teenager we expect to become famous." That history matters more than it looks: the label was born in a context where the future rating was knowable, because a programmer had typed it in. Out here nobody has typed it in.

And here is the structural fact the rest of this page hangs on. A list is a selection. It names people. It does not describe a population, and it was never built to.

A ranked list has no denominator

To answer "how often does someone like this make it", you need two numbers: how many players looked like this, and how many of them got there. The first is the denominator, and it is the hard one. A list hands you a hand-picked few candidates for the numerator and no denominator at all.

That is not a failure of the people compiling it. There is no slot in the format for a base rate, and adding one would not help, because a ranked list's unit is the individual while a base rate is a property of a population. You cannot get the second out of the first however good the scouting is. It is a category difference, not a diligence problem.

The obvious move at this point is to pull up a list from some years ago and count how many of them are where it implied they would be. We are not going to, for two reasons. The first is that it would mean writing about named people who were mostly children when somebody else put them in rank order, and that is not our material.

The second is the one worth arguing about: it would not answer the question anyway. A follow-up on a previous edition gives you a success rate for players who had already been selected. That is a fact about the selection — it tells you how good that list was, which is genuinely interesting — and it is not a fact about the population the selection came out of. "How accurate was this list?" and "how hard is this?" are two questions, and only the first one can be answered by re-reading a list.

Three things the format leaves out, none of them its fault

Take the criticism off the compilers entirely and it still undercounts, because of what the frame excludes before anyone starts typing.

• Everything upstream has already happened. By the time a player is eligible for a list like this, he has been selected over and over: into a club, into an academy, past each of the ages at which a club decides whether to keep him. The list starts at the top of a pyramid whose base is never printed, and the base has meetings in it — what actually happens to players an academy releases is a separate subject with real reporting behind it, and it is where most of this pyramid's population is.
• "Making it" moves while you argue about it. A list implies one outcome — the very top of the game, the reason you would rank teenagers at all — and gets defended two years later with a different one: he is a professional footballer, he is playing in a good league, that is hardly failure. Both statements can be true at once. Swapping between them mid-conversation is how a hype list is never wrong about anybody.
• The verification horizon is shorter than the claim. A list is published in one season and litigated in the next, about a claim whose subject is a career that runs for decades. Judged that early it is judged on the wrong thing — and because nobody re-reads a ten-year-old list, it is never judged late either. The window in which the format could be checked is not the window anybody looks in.

Why this page will not give you a real-world number

You will have noticed there is not one number above about real football. That is deliberate, and it is the most important paragraph on the page.

We have never measured the real world. We have a tool that plays a simulation of a football career several thousand times and counts what happened in it, and that tool has no contact of any kind with real academies, real clubs, real lists or real children. Taking a figure out of it and putting it in a sentence about real teenagers would produce something that looks exactly like a statistic and is in fact a fact about a video game wearing a lab coat.

And it would not stay contained. A number like that gets quoted once, by a blog that did not read the method, and then quoted from the blog, and by the time anyone traces it back the project's job is to publicly disown its own citation. The strongest thing we own is that our numbers come with their method attached; the fastest way to lose that is to let one of them describe something it never measured.

We are not going to borrow anybody else's number for it either. A real-world rate quoted second-hand on a page full of our own figures reads, after one screenshot, as one of our figures. If you want the real-world side of this properly, the page for it is what happens to players released by an academy, which links the people who report on it and which is deliberately the one page on this site that also quotes no percentage.

So: everything below this line is about the game.

Separately, and about the game only: what our simulation measures

Everything in this section describes a video game. None of it is a measurement of real teenagers, real lists or real careers, and it should not be quoted as though it were.

Wonderkid is a football life simulator: one tap is one year, you are the player rather than the manager, and the career can simply not happen. The simulator underneath it has no interface attached, so a harness plays whole lives by the thousand and tallies the endings. The figures below come from that harness at 1,000 complete lives per play profile, 2,000 in total, run from a seeded random number generator against the shipped 609-card pack on 30 July 2026. The full method, the seed and both complete tables are published on the odds page; this page borrows three rows and no more.

Those two play profiles have to be named alongside every figure, because a percentage without its profile means nothing:

• Random — every decision taken at random and training ignored. The floor: what the game does to someone tapping through it without a plan.
• Engaged — trains every year and takes the ambitious option wherever one exists. A realistic ceiling for a player paying attention, not a perfect-play optimum.

Cut the run the way a hype list implicitly cuts a career — not "did he fail" but "how far up did he get" — and it looks like this:

One reading note that matters more than the numbers do: all three rows are shares of every career in the run, not of the row above. We have deliberately not divided one into another to produce a "share of professionals who reach a five-star club" — that is a different measurement, and the honest way to get it is to run the harness for it rather than to do arithmetic on a published table. Not doing that in public is the same discipline as not quoting a real-world figure we never took.

With that said, the gap is the whole point. Inside the game, turning professional is the ordinary outcome and being the kind of player a list would have been written about is not. 87.3% of random careers never play for a five-star club, and 93.4% never retire Immortal. Playing properly moves both a long way — the five-star rate doubles, 12.7% to 28.0% — and still leaves the top of the ladder rare: the Engaged profile, which trains every single year and takes the ambitious option every time one exists, finishes in the top legacy tier 15.4% of the time.

The fourth row of the same run — the careers that never turn professional at all, 27.1% on Random and 10.8% on Engaged — plus the second table on what the game gets players to do, and the raw output, are on the odds page with the rest of the measurement.

What that has and has not got to do with a list

It is not evidence about the lists. It cannot be: it is a measurement of a model somebody built, and a model agrees with its author far more readily than the world does. Anyone telling you a simulation has settled a question about real children is selling something, and it would be us.

What it is, is the same shape from the other side. A list is a numerator with the denominator cropped out of frame: real people, real judgement, no population. Our figures are the exact opposite — a denominator we control completely, describing a world we made up. Neither is the thing anybody actually wants, which is a measured base rate for real football, and the bodies best placed to produce that are the ones holding the registration records rather than either of us.

It does answer one thing within its own walls. We built a football career from scratch, tuned it until it felt fair, then counted what it produced — and even in a world we designed ourselves the top of the game stays rare. The drop-off between "he turned professional" and "he became what the headline said" was not something we had to add to make the game hard. It fell out of building the ladder at all.

Why a football game models the drop-off at all

Because the alternative is a game in which every career works, and a game in which every career works has nothing left with which to make the good ones feel earned. A five-star transfer is only a moment if it was not going to happen anyway.

That is the design reason. There is a second one that matters more to how the game handles it. A career that peaks two divisions below the one you wanted is not a loss screen in Wonderkid: the life carries on, and eventually a child carries the name and starts again with none of your ability. The game is rated for everyone and its audience starts at ten, so the one thing it will not do is treat a simulated kid whose football did not work out as a punchline.

Two things it also will not do, given what this page is about. No real footballer appears in the game — every player, rival, manager and agent is invented, generated inside your save, ageing and retiring alongside you, and none of them is modelled on a living person. And no real person's career is ever simulated, here or in anything we publish, which is the same rule that keeps every name off this page. If that sounds like the game you were looking for, the early-access list is where the invitations come from.

What the lists are genuinely good for

All of that is an argument about a format, not a complaint about it. A ranked list of the best teenagers in the world is an excellent answer to "who should I go and watch this season". It is a bad answer to "who is going to make it" — and, read closely, it usually does not claim to be the second. The reading does that. The format simply has nothing in it that pushes back, no line at the bottom saying and here is how many players who looked like this last year are anywhere near this now, because that line would need a number nobody has.

The people who compile them are usually right about ability, which is a genuinely hard thing to be right about. Ability is simply not the whole of an outcome — which is precisely why the outcome is worth measuring separately from the talent, even inside a model. So read the list, argue about the order, enjoy it for what it is. Just do not read it as a forecast. And do not read our numbers as a fact about anybody on it.