Fritz 20 and the false FIDE president headline: chess's credibility gap, seen from a Da Nang coffee shop

5:42 in the morning. A coffee shop on Nguyen Chi Thanh Street, Da Nang. The ceiling fan turns so slowly I can hear its steel blades cutting through the humid air. Across the table, a thirteen-year-old boy named Huy opens an old laptop, its cooling fan howling like a broken hive. He pushes the pawn to e4. The cursor blinks for three seconds. The machine returns +0.31.

Then my phone buzzes against the wooden tabletop.

A headline appears: "FIDE President Timur Turlov meets Cypriot President Nikos Christodoulides in Nicosia." Directly beneath it, in the same article: "Fritz 20 — your personal chess trainer. Your toughest opponent. Your strongest ally."

I put the phone down and look across the table. Still +0.31. The machine knows nothing about Cyprus. It does not know who sits in the chair of any federation, does not know where Nicosia falls on the emotional map of chess. It only knows that after e4, the game is still open, and there is a great deal left to say.

Two screens glowing at once in an empty café. One whispers. One shouts.

I have sat between those two things for forty years, and I know which one is usually right.

In 2026 I left the newsroom after eighteen years because a twenty-seven-year-old editor cut my piece and told me I was "writing literature inside a sports report." I thought I had been exiled from football; football turned out to be the road deeper into myself. But football was only half of it. The other half was the board — the thing I learned at twelve in a small town where people played on front porches and nobody knew what Elo meant.

So when I read that headline, my first reflex was not anger. It was curiosity. Because an article that lies about the head of an international federation and still sells software tells you something about the buyer.

The headline and the gap behind it

The piece was a product introduction. It carried three slogans for Fritz 20: personal trainer, toughest opponent, strongest ally. It carried a meeting in Nicosia. It carried two big names. And it carried very little else.

Not a single game. Not a single variation. No engine-strength metric, no search depth, no neural-network architecture, no comparison with any competitor on the market.

On the personnel side: the sitting FIDE President is Arkady Dvorkovich, elected in October 2026 in Batumi and re-elected in 2026 in Chennai. That is public record, stable for years, checkable by anyone. Timur Turlov is a Kazakhstan-born billionaire, chief executive of Freedom Holding Corp, a man publicly associated with chess sponsorship in Kazakhstan. He may well have a role in chess. He does not occupy the chair the headline gave him.

Nicos Christodoulides is real, and he has indeed been President of Cyprus since early 2026. Nicosia is indeed the capital. True details were placed beside a false one, and the way they were blended is the thing worth discussing.

A typo cannot produce this structure. Producing it requires a writer who needs an authority hook and a reader who has no time to check.

I remember 2026, when I still organised tournaments and wrote news for the domestic chess scene. Everything passed through two gates: a federation secretary and a landline telephone. Get a president's name wrong and someone called to shout at you. Now an article travels from an unknown page to the screen of a thirteen-year-old in Da Nang in two seconds, and nobody stands in between.

How many lifetimes Fritz has lived through

To understand why a piece of chess software needs to borrow political credibility, go back to where it started.

Fritz appeared in 2026, written by Frans Morsch, a Dutch programmer, and published by ChessBase in Hamburg. ChessBase was founded in 2026 by Frederic Friedel and Matthias Wullenweber, starting from an idea that seemed almost mad at the time: storing games in a computer so they could be searched.

Fritz 3 won the World Computer Chess Championship in Hong Kong in 2026. Then came the period I consider the most beautiful and most naive in modern chess history.

In October 2026, in Bahrain, Deep Fritz drew 4–4 with Vladimir Kramnik over eight games. Kramnik was then world champion, the man who had beaten Garry Kasparov in 2026. Chess breathed out: the humans still hold.

In November 2026, in New York, X3D Fritz drew 2–2 with Kasparov over four games, played in three-dimensional space through glass. I remember watching those four games three times over, just to hear commentators talk about a machine the way one talks about a person.

Then in November and December 2026, in Bonn, Deep Fritz beat Kramnik 4–2. The second game of that match entered history through a mate in one. Kramnik, world champion, made a mistake a club player would spot once shown.

After Bonn, the man-versus-machine story died. Not because the machine won. Because the machine winning became boring.

Stockfish emerged from an open-source branch and became the community standard. Leela Chess Zero arrived in 2026, learning the way AlphaZero did after the December 2026 paper. In 2026, Stockfish switched to NNUE — an efficiently updatable neural network — and the gap between free software and commercial software became a chasm no amount of strength could bridge.

That is the real break. Not 2026. 2026.

From the moment the strongest engine in the world became a free download for anyone with a connection, strength stopped being a commodity. And when strength becomes free, the only thing left to sell is what surrounds it: interface, curriculum, the feeling of being guided, and a story.

What is actually being sold

I am not disparaging Fritz 20. I am saying it is being sold on precisely the thing it is no longer central to.

A modern top-tier engine has moved so far beyond every human player that estimated Elo numbers lose practical meaning. When you are three hundred points stronger than the best person alive, being twenty points stronger than your colleague has nothing to do with the buyer.

So what are they paying for?

They pay for designed friction. A levelled mode, so a beginner is not crushed in twelve moves. An explanation mode, so a move has a reason in human language. An annotated opening library. An opponent that errs the way people err, in the places people actually err.

The value of a training machine lies in the quality of the friction it creates, not in its Elo.

And here is the misalignment. All three slogans in that advertisement orbit strength: best teacher, toughest opponent, strongest ally. None of them mentions being allowed to be wrong.

But the right to be wrong is exactly what a learner needs.

I have sat beside Huy for three months now. Every Saturday morning he brings a game he played online; we switch the machine on and walk through it move by move. Once the engine pointed to a better move at move nineteen and he nodded along. Once he argued. He said: "But if I play that, I don't understand why."

That sentence was more accurate than any evaluation the engine produced.

The paradox of the perfect teacher

A machine that is never wrong cannot teach anyone how to correct. It can only teach lookup.

I thought the machine had taught me chess; it had taught me how to sit still.

Before machines, finding out whether a move was good or bad meant waiting for a tournament, asking someone older, losing once and remembering. That process was slow, painful, and extremely effective — because it forced me to build a hypothesis before I got an answer.

Now the answer arrives before the hypothesis. The boy pushes a move, has not finished thinking, and the evaluation bar jumps. He learns that there is an authority that always knows first. And he never learns the feeling of seeing it himself.

That is why I think a product selling "the toughest opponent" is selling against the need. What a thirteen-year-old needs is an opponent that errs at the right moments, errs in the places players at that level genuinely err, so the boy can discover the error himself.

Fortunately modern software all ships with strength limiting, and Fritz has a long tradition here — it was once the program coaches used to build exercises for students. But the distance between what the software can do and what the advertising says about it is enormous, and that distance is always filled with ornate language.

"A revolution in training." There is not a single measurement behind that sentence.

I have watched chess for forty years and learned one rule: whenever a product calls itself a revolution, look for the comparison table. No comparison table, no revolution.

The beautiful metric and what it fails to measure

The same problem at a different layer.

People now use ACPL — average centipawn loss per move — as a measure of form. The logic is simple: compare each move to the engine's best, convert to points, sum, divide.

It sounds scientific. And it measures exactly one thing: the distance between you and the machine.

ACPL measures distance from the engine's move, not the quality of a human decision.

A player makes a slightly inferior move objectively but places the opponent in a position with no time to think — that move is scored as a loss. A player plays the engine's exact choice at the wrong moment, with eleven seconds on the clock — that move is scored as perfect.

I have watched games where the winner had a worse ACPL than the loser. I have watched games where an apparently mistaken move was the decisive one, because it traded a sliver of objective advantage for the opponent's entire composure.

This is where I hold my position, unchanged across years of writing about football and about chess: expected-value metrics are very useful for describing the past and very easily abused to explain the present. They give you a number instead of a story, and numbers print more easily than stories.

Good training software must teach this. It must show that some moves the engine calls poor should not be chosen by humans, because they demand a precision nobody sustains for four hours.

If Fritz 20 does that, it is worth the money. If it only offers an evaluation bar and a green arrow, it is an expensive mirror.

Adult tempo and a child's tempo

There is something I have tracked for years and still find painful.

Based on my experience following domestic matches and youth tournaments, I see a repeating pattern: a twelve-year-old memorises twenty-five opening moves of a complex system, plays well for fifteen moves, then collapses the moment he leaves the book.

That child is not weak. He has been pushed into adult tempo before his body and his thinking were ready.

The machine is part of that pressure. When answers are available, memorising becomes easier than understanding, and memorising always wins in the short run. Parents see a junior title. Coaches see the rating tick. Nobody sees the price, because the price arrives later, at nineteen, when a young player suddenly has nothing to lean on.

Nguyen Ngoc Truong Son became a grandmaster before his fifteenth birthday, in 2026. Le Quang Liem won the World Blitz Championship in 2026 in Khanty-Mansiysk and spent a stretch near 2740 Elo. Both came out of an era when engines existed but had not yet reached every household.

My point is not that the old days were better. Memory is useful only when it shines down on the present; when it merely praises the past, it is a bad habit.

My point is that both of them grew up in an environment where every move had to be defended in words. Now a child can produce the correct move without defending anything, and that is a trade, not a pure advance.

Good training software should take the side of slowing down. It should force the learner to explain before revealing the answer.

Which road the money takes

There is one more layer, and it explains why that headline exists.

World chess runs on sponsorship. FIDE has a budget, events, organising costs, and its money comes from wealthy corporations and individuals. A Kazakh billionaire with a US-listed company and chess sponsorship activity at home is an entirely plausible figure in chess governance.

And here I keep the stance I carried over from football: sponsorship money bypasses the oversight corridor the way a free-agent contract bypasses financial fair play.

When a large sum enters a sports organisation as sponsorship, it does not pass through the gate built to inspect it. There is no transfer contract, no fee, no balance sheet to reconcile. There is a ceremony, a handshake, a photograph.

And when the organisation depends on that money, the price does not appear in the books. It appears in the questions that do not get asked.

A headline that assigns the presidency to a potential sponsor is not harmless. It is advertising first, verification later. It softens the ground for a story that may be told differently a few years from now.

I have no evidence of a conspiracy. I have a pattern, and that pattern has held too many times in my career for me to ignore it.

The blind spot of collective memory

Chess remembers the wrong breaking point.

Collective memory places the moment the machine passed the human at Bonn 2026, when Kramnik fell to Deep Fritz. It is a beautiful story: proud humanity, cold machine, one game as an emblem.

The real break was 2026, when top-level engine strength became a free software package anyone with a computer could download. After that moment, every commercial product in the field had to sell something else. And the easiest thing to sell is always borrowed credibility.

The second blind spot: we treat a wrong headline as a technical error. It is a market signal. Nobody invents a political hook unless the hook sells. If the Vietnamese chess audience numbered three thousand readers, nobody would write that headline. Someone wrote it, which means there are enough readers for an unknown page to earn advertising money from it.

That is good news wrapped in bad. Chess has grown large enough to be farmed. It has not yet grown large enough to protect itself.

The third blind spot, and the one I weigh heaviest: we import software without importing method. A club in Da Nang buys a licence, installs it, and uses it as a scoring machine. Nobody writes the students a single A4 page on how to ask the engine a question before opening the engine.

I thought I had been exiled from the board; the board turned out to be the road deeper into myself.

I first wrote that line on the Co May blog, in a piece about the final I lost to injury at seventeen. It still holds. But it holds differently now: I was not pushed away from the board by age or by profession. I was pushed away by a screen that always knows the answer before I do, and I have to find my own way back.

What to watch

Four things I will be watching over the next six months.

First, whether FIDE responds. How an international organisation reacts when its leader's name is falsely attached to a viral sales piece says a great deal about its communications professionalism.

Second, whether the publisher of Fritz releases any independent measurement for version 20. A comparison against Stockfish, a figure on neural architecture, a dataset on latency across skill levels. If it exists, good. If not, the story remains a story.

Third, Timur Turlov's actual role in chess. If there is a formal role, it should be announced in a press release, not implied by a wrong headline.

Fourth, and to me most important: whether any club in Vietnam writes down a coaching process that uses engines without turning a child into a memorising machine. One sheet of paper. Thirty lines. How to ask the question before switching the machine on.

I asked Huy that morning, after finishing the headline: "If you were not allowed to see the evaluation bar for the first twenty moves, would you play differently?"

He thought for a long time. Then he said: "I would play slower. But I think I would understand more."

I switched off my phone, folded it, and set it face down on the table. Across from me the machine still returned +0.31, patient, not arrogant, needing nobody to believe in it.

However strong a machine becomes, it cannot write its own headline. That work remains human, and it remains where humans fail most.

The thirteen-year-old in the coffee shop on Nguyen Chi Thanh understood what an article with two big names did not: slowing down does not weaken your chess. It only makes you responsible for what you think.

If Vietnam's next generation of players grows up with a machine in their pocket and a question in their head, we will get a very different generation. If they grow up with a machine in their pocket and a false headline on their phone, we will get more strong players and fewer people who know why they play at all.

I choose to believe in the child who already knew to tell himself to slow down.