Petroff Defence Gets Patched, Positional Play Sold Twice: ChessBase August 2026

On 13 September 2026, ChessBase published its roundup of August releases. Four content groups: two positional-play courses, a set of calculation and strategic decision-making courses, an updated database family for the Petroff Defence, and an opening-repertoire section described as active. Three names behind them: Bahne Fuhrmann, Andrew Martin, Kostya Kavutskiy.

I read the piece three times. Once for work. Once out of curiosity. The third time because of a question that would not leave — across a whole month's catalogue, exactly one opening area was touched, and it was the Petroff.

The Petroff Defence, 1.e4 e5 2.Nf3 Nf6, is not for players who like adventure. It is for players who do not want to lose. Symmetrical, solid, low theoretical risk, and in the right hands it drains White's initiative without a single decorative move. A database update for it is not breaking news. But it says more than two positional-play courses combined.

A release schedule, not a product

ChessBase does not sell a product. It sells a rhythm. One new batch each month, each batch covering several topics, each topic attached to one or more named authors. A subscription platform survives only if paying users keep seeing something new. Churn is the real enemy, not the competitor.

Chessable goes the other way: courses tied to big names, sold one product at a time, each course its own media moment. ChessBase chooses breadth. Looking at the August batch, the two-halves structure is obvious. The first half is the theory race for strong players. The second half is the skills market for the crowd.

The Petroff update belongs to the first half. The two positional-play courses belong to the second. The fact that content is segmented by playing strength tells you the target is a mass market, not a narrow elite tool.

In India, where I work, the second half is most of the business. After Gukesh's world title and the wave of Praggnanandhaa, R. Vaishali and Arjun Erigaisi, the number of new chess learners has grown fast. Most of them do not need a new line at move 15 of the Petroff. They need to know how to decide with twelve minutes on the clock.

The trap sits inside the word update

The updated Petroff Defence database is the only opening item in the August batch. That does not mean the Petroff is rising. It means that at the elite level, theory maintenance has become a revenue-generating side profession.

The Petroff is the kind of opening the machine froze long ago in many branches, yet small move-order gaps remain. For a 2600 player, a reordered move can change the whole plan. For a commercial database, that is enough reason to sell an update. ChessBase gives no engine evaluation, no specific line, no move-quality metric. Only the word update.

That is the most notable detail in the entire release note.

I have spent years replaying Petroff games from open tournaments. What caught my eye was never the beautiful moves, but the number of moves the two sides used to cancel each other out. Mistakes are not frightening; crouching down at 2 a.m. to pick the tape back up is what makes a forecaster.

Three authors, three doors

Bahne Fuhrmann, Andrew Martin, Kostya Kavutskiy. Martin is a veteran ChessBase author, an IM with a distinctive commentary voice and a long-standing place in the catalogue. Kavutskiy is a US-based IM, comfortable as a commentator and online teacher. Fuhrmann is less familiar to a mass audience.

Placing three names side by side in one batch is a distribution decision, not an academic one. Each one pulls in an audience already used to their voice. One catalogue, three doors, one invoice.

That structure is also a career risk for the publisher. Content value depends on whether the authors keep their standing. If a name leaves, it takes part of the catalogue with it.

Turning a game into a curriculum

Look closer and the August batch is not a set of unrelated products. It is a syllabus: openings covered, calculation covered, strategic decision-making covered, positional play covered. That is a full training cycle, not a single themed release.

The way the content is described also follows the template that has become the market standard: theoretical explanations, annotated games, practical exercises. That triad is now the default. It is no longer a differentiator; it is a minimum requirement.

I once mispronounced a player's name, then read the game correctly when the slow-motion replay appeared in front of me. At this age I am not afraid to say it plainly: a template catalogue is not automatically a good catalogue.

Two positional-play courses, one question

Two positional-play courses in a single cycle is the most arguable point. On one hand it shows depth: multiple levels, multiple approaches. On the other it signals catalogue overlap — two products from one publisher, same subject, released too close together.

In a crowded market, internal overlap can be a market-coverage strategy, or a sign of exhausted new ground. From outside there is no way to tell without seeing the content itself.

When free becomes the default

Lichess opens its entire database, a strong engine costs nothing, and no one pays a cent. That is the ground every content publisher stands on. You cannot sell what a website gives away. You can only sell what comes with it: learning structure, learning order, a guide.

That is the real reason the August batch leans on skills. Not because the publisher loves skills. Because skills are the only part with margin left.

The release calendar is chess's transfer window

In football, the transfer window is when noise overwhelms signal. In chess, the content release calendar plays the same role. One batch a month, one announcement each, a little noise each time. The reader is put in a position of having to filter.

My filter has three questions. Is there any data. Is there any specific content. And is the party reporting also the party selling. The August batch fails all three. No engine evaluation, no specific line, and the reporter is the seller.

Where I might be wrong

If I look at two positional-play courses and call that overlap, I might be misreading. Perhaps one is aimed at club-level players and the other at those with a foundation. Two courses does not automatically mean two identical courses. I have no internal content to verify.

The second place I might be wrong is the self-praise problem. ChessBase both publishes and writes the introduction. The images come from ChessBase. The tone is supportive. Every quality claim should be read as advertising, not independent assessment.

But if I call that a problem, I have to be honest: it is an old problem, and not unique to ChessBase. Buyers are used to sellers writing about their own goods. The more worrying part is on the reader's side.

What is genuinely worrying

We consume roundups like this as news. A product catalogue is read as an academic signal. A release schedule is read as a trend in chess. That shift is where the trouble lies.

Positional play is not something you download like a theory file. Calculation is not something you copy from machine to human. That is exactly why they became sellable goods. Once engines and open databases gave away opening theory for free, what remains sellable is what the machine cannot hand over directly: the skill of conversion.

That empty stadium years ago, I found a tactical diagram resting quietly under a seat instead of the roar of the crowd. In chess, that empty stadium is the analysis room. No spectators there, only moves. And it is precisely there that people pay for what cannot be copied.

When I ask who certifies the quality of a training product, the answer is usually silence. The question I ask is not the kind a woman asks, but the answer they dodge is not the kind a man gives.

Two halves of one model

Taken together, ChessBase is selling two kinds of goods to two kinds of people. To strong players it sells reassurance about theory: you will not be surprised at move 15. To the crowd it sells skill: you will decide better at move 30.

The first kind has a short life. Theory ages with every tournament. The second has a long life. Skill does not expire with the season.

If I had to bet on the direction of this catalogue over three years, I would bet on the second kind. Not because it is better, but because it sells for longer. In a subscription model, selling for longer is everything.

Ten years standing in the middle of arguments, and the fire lit by a question in 2026 is still burning. That year the question was whether someone understood pressing. This year the question is whether a product catalogue counts as a signal.

The right name for a table of contents

Silence applause can kill emotion, but it cannot kill tactics. A product roundup is the same: it cannot kill the market, but it creates no tactics either.

What I mean is not that ChessBase does poor work. What I mean is that we are calling a table of contents by the name of an analysis.

What I will be counting

I will say it in advance so there is something to check against. Over the next twelve months, count the ratio between new opening content and skills content in ChessBase's monthly batches. If the ratio drifts toward skills, the market has clearly said what it needs.

If another Petroff update appears without a single specific line, call it by its proper name. A table of contents. Not a discovery.