FRITZ 20: When the Chess Machine Stops Being an Opponent and Becomes Your Personal Coach

On the night of November 25, 2026, I sat in front of a screen watching game six between Vladimir Kramnik and Deep Fritz. Kramnik, then world champion, collapsed after a mistake he himself called unforgivable. The final score, 4-2, went to the machine. Fifteen years later, I still tell that game whenever someone asks why I left the board to sit in a broadcast booth. But the story I want to tell today is not about a human losing to a machine. Today's story is about a machine changing roles.

FRITZ 20 has just been announced with three lines of description: your personal chess trainer, your toughest opponent, your strongest ally. Those lines are not marketing slogans. They describe a turning point the chess world has been half-walking for years without naming it properly. The analysis engine is abandoning its role as referee and stepping straight into the learning process.

Three decades of shifting roles

Fritz's history is welded to the history of computer chess. In 2026, Frans Morsch and Mathias Feist released Fritz 1 under the ChessBase roof. Over the three decades that followed, the name Fritz travelled alongside milestones any chess player knows by heart. In 2026, X3D Fritz drew Garry Kasparov 2-2 in a four-game match in New York. In 2026, Deep Fritz beat Kramnik 4-2 in Bonn.

That was the era when a machine was defined by strength. Every new version was a declaration that computers had moved one more step past humans. But when every top engine has passed the 3000 Elo mark and the gaps between them shrink to a few dozen points, the strength race becomes meaningless for ordinary players. A 1800-rated player does not need a stronger machine. He needs a machine that understands him.

That is exactly the gap FRITZ 20 is aiming at. Not stronger, but closer. Based on my experience watching hundreds of games and commentating live for four years, what young players lack is not calculating power. They lack a feedback system that knows how to ask the right question.

Three adjectives and three promises

The publisher uses three adjectives: more efficient, more intelligent, more individual. I read those adjectives differently from a press release.

Efficiency lives in the speed of the learning loop. An amateur player can spend three hours analysing a game and extract two lessons. A good training tool must extract twenty lessons in thirty minutes. That is not about calculating faster. It is about filtering signal from noise. In chess, the signal is the moment when the evaluation swings abruptly. The noise is the hundreds of neutral moves in between. A machine that points at the three decisive moments of a game teaches more than a machine that points at the whole game.

Intelligence is the ability to recognise repeated error patterns. This is the hardest part, and the part pure engines have never done well. An engine analysing your game at depth thirty will find the best move. But it will not tell you that you have made the same type of mistake fourteen times in your last twenty games. Humans do that. If FRITZ 20 truly does it, it has crossed the line between tool and teacher.

Personalisation is the most dangerous promise, because it is easily reduced to decoration. Real personalisation is not an interface that changes colour to taste. It is a machine building a roadmap from your own weaknesses. If you are weak in rook endgames, it must push you into rook endgames. If you tend to lose composure at move twenty-five, it must build games that carry you to exactly move twenty-five.

In the opening phase, the real value of a training tool is not handing you an opening book of millions of moves. Everyone has huge databases. The value is showing that you have chosen a system that does not fit your style. An attacking player will struggle with dry positional openings. A patient player will be frustrated by early force trades. A machine that reads that mismatch will save you years of wandering.

In the endgame, the story is even clearer. The endgame is where pure technique decides results, and also where amateurs skip the most work. A tool that generates hundreds of endgame positions based on exactly the endgame type you are weak in turns a blind spot into a strength within months.

I have seen a generation ready to press, and I believe they will never take a step back.

A companion that knows how to lose

There is a technical detail I consider more important than engine strength, yet it is rarely mentioned at product launches. A good training machine must know how to lose at the right moment.

When you train against an opponent that always plays at the highest level, you learn very little, because every game ends the same way. When you train against an opponent that can adjust to create positions suited to that day's lesson, you improve. This sounds small, but it changes the entire nature of training.

In four years of chess commentary, I have watched young players beat an engine on easy mode and believe they had become strong. I have also watched people crushed on hard mode lose all confidence. Both are failures of design, not of ability. A machine that adjusts so you win just enough, then lose just enough, teaches more than a machine that knows everything.

Strength is dead, pedagogy is the battlefield

Chess has a stubborn habit: measuring everything by Elo. When an engine launches, the first question is always how strong it is. That question has expired. Stockfish, Leela, Komodo and Fritz are all far beyond any human player. Raw strength is no longer a competitive advantage. It is the minimum requirement.

The real question lies elsewhere: can a machine teach without creating dependency? In the past fifteen years, chess has seen a generation of young players raised alongside engines, astonishingly precise in calculation, yet gradually losing the ability to read a position by intuition. They argue well and judge poorly. They are strong with a machine beside them, weak when they must decide alone.

A good training tool must solve that paradox, not deepen it. The only way is to design so the learner moves first and the machine moves after. If FRITZ 20 encourages players to commit to their own judgement before opening the analysis pane, it is doing exactly what the whole field needs. If it turns every session into watching a machine explain, it is only a prettier screen.

In the world of chess, the only boundary is the boundary of skill.

The gender gap on the board

There is an aspect product launches rarely mention. Chess is one of the few intellectual sports that does not split by gender on the open stage. Yet in the list of top players in the world, the share of women stays low in a way that cannot be explained by ability alone. The causes lie in access, in training environments, and in invisible biases that push girls away from the board very early.

I once spent a summer searching for women's faces that had no name on the sports media map. Those forgotten queens did not lack talent. They lacked an audience. An affordable personal training tool can be a lever. When the quality of instruction no longer depends on whether you can pay for a private coach, the gap starts to narrow.

Imagine an afternoon when your little sister sits down at the board without needing anyone's permission.

A machine cannot save anyone

FRITZ 20 will not turn a beginner into a grandmaster. No tool can, and anyone promising otherwise is selling an illusion. But it can change how a serious player spends his time. Instead of three hours wrestling alone with lines, the learner can spend those three hours understanding himself.

I once thought machines were humanity's opponent. In 2026, I believed it. But the best teacher I ever had was not the person who beat me. It was the person who showed me where I had beaten myself. If FRITZ 20 can do that, it deserves the name companion. On the chessboard, a true companion is worth more than a strong opponent.