FRITZ 20 and the Data Gap: How Free Chess Tools Rewrote the Training Market

One December evening, in a small chess club tucked deep in an alley off a Hanoi street, I sat behind a fourteen-year-old boy. He had no laptop — only an old phone. On the screen, Stockfish was running: free, open-source, strong enough to beat almost any player at the board. The boy rewound one of his own games, stopped at move twenty-three, pointed at an empty square and told me: "I went wrong here. The knight should have come to this square." He did not know what ACPL means. He did not need a software box costing a few hundred thousand to a few million dong. He only needed a tool strong enough to answer his question — and that tool had been free for years. Weeks later, I read the announcement for ChessBase's FRITZ 20. It called itself "a training revolution," a "personal chess trainer" for beginners and professionals alike. I read it twice, then opened my notes. ChessBase is no stranger. Founded in Germany in 2026, the company was once the benchmark of chess software: game databases, analysis interfaces, and the Fritz engine — the name that beat Garry Kasparov in the famous Man vs. Machine match in Bonn in 2026. Fritz was, at the time, the symbol of a machine surpassing a human at an intellectual sport. But the market has changed. Stockfish, the open-source engine released in 2026, now leads virtually every engine-strength ranking. Leela Chess Zero uses neural networks and reinforcement learning. Chess.com integrates AI analysis directly into its online playing platform. Lichess offers free, ad-free analysis. Chessable sells structured opening courses. Each major player has chosen a niche and taken it. Against that backdrop, FRITZ 20 arrives with a broad claim: for everyone. But reading the release carefully, I found not a single number. No ACPL figure, no engine match rate, no technical specifications, no comparison table. Only adjectives. This is where I want to pause a little longer than others. In chess, every claim about strength can be measured. ACPL — the metric for the average deviation of each move from the best move — is the currency of analysis. Engine match rate, search depth, response time per position: all are verifiable numbers. A product that wants to convince chess players should lead with data, not adjectives. I am not denying that FRITZ 20 may be useful. I am denying how it is being sold. Look at the structure of the pitch. It says the product is for "ambitious players and professionals," and also for "players taking their first steps." In chess, those two groups have completely different needs. Beginners need principles: control the centre, develop pieces, keep the king safe. Professionals need deep analysis tools, opening databases, customisation. A product that embraces both usually means it is optimised for neither. This is not only my judgment. It is the logic of the software market in general. When a tool dares not compare itself with rivals, it is usually because it knows it would lose a direct comparison. When a product calls itself a "revolution" without offering an operational definition of the word, "revolution" is just a word. The board has its geometry. A move does not exist in a vacuum — it opens one square, closes another, changes how many squares a piece controls, shifts the centre of gravity of the whole position. A good training tool is one that teaches a player to see that geometry. Not a tool that says "this move is wrong," but one that shows why it is wrong, by drawing the controlled squares before and after the move. That is the standard I use to measure any chess software. And by that standard, the FRITZ 20 release answers none of the questions. I tried applying that standard to how free platforms present their analysis. Lichess draws the best move as an arrow on the board, shades controlled squares, lets players step through each move and read the explanation. Chess.com displays an evaluation graph move by move, marking where the game turned. Both are free at the basic level, and both answer the question "why." A paid product that wants to compete must clear that bar — it cannot stop at declaring itself a revolution. What does my data say? Based on my experience following games, I have spent many evenings replaying moves with the sound off, watching only the pieces and the squares they control. I learned that a tool's strength lies not in how many moves ahead it calculates, but in how many of them it can explain to a human. An engine that beats you without teaching you anything is a useless engine for a learner. The real value of training lies in the gap between "the best move" and "the move you understand." People say girls know nothing about tactics. So I write for them to read. And the first thing I want them to read is: do not trust a pitch that lacks data. Here I want to argue against myself. There is an easily overlooked angle: FRITZ 20's failure to present data may not be a product weakness but a strategic shift. Free Stockfish has fully captured the engine-strength market. Nobody buys an engine just to own the strongest one anymore. So ChessBase must sell something else: experience, interface, integration. The problem is that it has not yet found the language to sell that experience with evidence. But there is a larger problem, and it belongs to the chess community itself. We — the players, the writers about chess — have stopped demanding evidence. We have grown used to reading a marketing release and copying it as if it were news. Each time we do, we rob readers of the chance to judge the product properly. Worse, we teach software companies that they need prove nothing at all. For Vietnamese chess players, the issue is more concrete. Most young players in Hanoi, Da Nang or Ho Chi Minh City come to chess through free tools, through small clubs, through coaches without international titles. They do not have the budget to buy expensive software and then test whether it was worth it. For them, a pitch without data is not merely an academic matter. It is money that may be poured into something that brings no progress. The next game I watch will still be with the sound off. And I will still ask the old question: where is the evidence? Tactics never lie; only the reader gets it wrong. For chess software, that means: do not read the promise, read the number. If one day FRITZ 20 publishes its ACPL, its engine match rate, or a comparison with Stockfish and Leela, I will be the first to open it and verify. Until then, a revolution without data is only an unproven revolution. Sometimes, to understand a product, you must stand in silence longer than others can bear.