FRITZ 20 and the Chess Engine Pivot: When Selling Strength Stops Working

Hook

In November 2026, in Bonn, Vladimir Kramnik sat down at the board against Deep Fritz and missed a mate. Not a twelve-move combination, not a sacrifice requiring calculation to the endgame. One move. A child three months into chess would have seen it. The machine did not win by calculating deeper than the world champion; it won by sitting still long enough for the world champion to drop himself.

I was nineteen that year, sitting in front of a CRT monitor in Binh Duong, rewinding that game a few times in the night. What haunted me for nearly twenty years was not Fritz. It was where the human went wrong, and the fact that the machine was only a mirror patient enough not to judge.

Nearly twenty years later, I read the FRITZ 20 announcement published by ChessBase itself. No date, no specifications, no price, no performance comparison. Only very beautiful phrases: “your personal chess trainer,” “a training revolution,” “efficient, intelligent, individual.” I realised I was reading exactly what anyone writing about chess must read most carefully: a piece of advertising dressed as news.

Context

Fritz arrived in 2026, developed by Frans Morsch and Mathias Feist and brought to market by ChessBase, the chess software company founded in 2026 in Hamburg. Through the 1990s and 2000s the name Fritz made grandmasters wary. In 2026, at Brains in Bahrain, Deep Fritz drew 4-4 with Kramnik. Four years later, in Bonn, Deep Fritz beat Kramnik 4-2. That was the era when engine strength was still sellable goods: every engine Elo point was a line of advertising, every win over a world champion a page of catalogue.

Then the market turned. Stockfish appeared, open source, free, and quickly took the top of the engine rating lists. Leela Chess Zero launched in 2026 with AlphaZero-style neural network thinking. Lichess gave analysis away. Chess.com folded play, study and analysis into one subscription. In 2026 Chess.com bought Chessable, which meant buying most of the structured chess learning market.

Against that backdrop, ChessBase presenting FRITZ 20 in the language of a training platform is a rational move. The question is what it says, and what it does not.

What the machine can no longer sell

The commercial value of engine strength has gone to zero, and what remains sellable is a method that helps people see their own errors. That conclusion comes from how the announcement itself is written. A company that once boasted of an engine that beat a world champion now offers not a single number on Elo, move-match rate or search depth. That silence is calculated. Putting FRITZ on the same strength scale as Stockfish means accepting defeat in the first line.

But what is offered in its place is vaguer than any number. “Individual,” “intelligent,” “revolution” are words that cannot be checked. I have spent years commentating on chess and organising tournaments, long enough to know one thing: in chess, teaching cannot be shared across all levels. A method for a beginner cannot be a method for a grandmaster, and claiming to serve both at once is the sign of a product that does not know who it is for.

More concretely: a 1200 player needs pattern recognition, needs basic tactical motifs drilled into reflex. A 2500 player needs an opening repertoire maintained, needs endgame precision, needs things where one percent wrong costs half a point. Those two problems do not share one answer. The announcement says the product is for “ambitious and professional players,” and immediately after says it is for “players taking their first steps.” Those two sentences cannot both be true of one bounded product.

When I was still competing, Fritz was a training partner. Its value lay in the clock, in repetition, in never getting tired and never going easy on me. But there is one thing it could not do, and still cannot: it told me my move thirty was wrong; it did not tell me why, at that moment, I thought the way I did. The gap between an analysis engine and a teacher sits exactly there.

Based on my experience following and commentating on domestic and international chess events, most amateur players buy the strongest software and use it as a marking machine. They finish a game, look at the graph, see the dip at move twenty, nod, and make the same mistake next game. A machine does not fix habits. Only a method fixes habits, and a method has to be designed for a specific group of people.

On many nights during the empty-stadium stretch of 2026, I reopened old games and remembered writing: “Sixty days of listening to recordings, I imagined the roar of empty seats.” Watching an old game again feels the same: the applause is gone, only the question of why that move was chosen remains. When there is no crowd, football returns to the sound of its own breathing. So does the board.

The counter-view

The familiar explanation is that free engines killed commercial chess software. I do not entirely believe it.

Stockfish took away the price of strength, but it did not take away the teaching market. It made that market bigger: online chess exploded after 2026 and brought tens of millions of newcomers to the board, and newcomers need to learn, not a stronger engine. The death of engine strength is the life of the coaching market.

FRITZ 20's real risk lies elsewhere. It lies in an announcement that will not admit what it is. A product serving everyone from beginner to professional, with no price, no specifications, no release date, living on the reputation of a name from 2026. The buyer is asked to trust the memory of Fritz more than the spec sheet of FRITZ 20. Age does not extinguish love for the pitch, it only changes its colour — but for a product, changing colour is not enough.

To be fair, not publishing head-to-head figures is the sober choice. In a pure strength race, no one beats an open-source project with thousands of unpaid contributors. Leaving that race is the only remaining road. The problem is that once you leave, you need something else to sell, and the announcement has not shown what that is.

What I want to know

I do not need FRITZ 20 to be stronger than Stockfish. I need it to answer one question: after I lose, can it tell me why I played that stupid move at move thirty, when I was tired, when the clock had four minutes, when I had calculated eighteen moves ahead — or does it just show a red question mark on the screen? The game ends, but the applause still rings in my memory. With chess software, that applause only rings when it teaches someone something beyond the board.