FritzTrainer's New Author Kostya Kavutskiy: Two Courses, One Reviewer, Zero Sample

On my Moscow desk, the first thing the ChessBase announcement offered was a ratio, not a move. It contains two course titles, one reviewer's name, two comparison authors, and exactly zero games. Three names against zero moves — the whole story of a product review sits in that ratio. Kostya Kavutskiy, an International Master and FIDE Trainer, enters the ChessBase FritzTrainer line with “Mastering Calculation” and “The Fundamentals of Positional Play,” and a club-level reviewer places him beside Robert Ris and Ivan Sokolov. I went looking for the number, and the number is not there. That is the largest fact here: the announcement delivered a verdict and withheld the sample behind it. When a chess announcement carries no rating, no opponent, no move count, reading it as sports analysis means asking first: where is the base?

Context first. ChessBase is a major publisher of chess software, databases and training material; FritzTrainer is its video-course format, typically combining lectures, game analysis and interactive exercises. Kavutskiy's identity works on two levels: as a player he is an International Master, as a teacher he holds FIDE-recognized trainer credentials. Those two identities carry two different claims, and the review risks merging them. The comparison set — Robert Ris and Ivan Sokolov — are established names in the training market; Kavutskiy is the latest addition. This is not a tournament, not a championship; it is a product release, and its evaluation yardstick is entirely different. In a tournament you measure field strength, draw rate, performance rating; in a course release you measure price, length, curriculum density, and — hardest of all — the student's actual improvement. That last measure is rarely taken, and that is where the training market's biggest blind spot hides. Around this market, subscription-based spaced repetition, free YouTube lessons and online platform curricula are all crowding in; when a legacy publisher like ChessBase adds a new author, it lengthens the queue but leaves the quality question untouched.

A provenance note matters here. The original item is a ChessBase-related publication, and the reviewer is Jochen Schwarz, himself a club-level player and a familiar reader of the FritzTrainer series. The precise metadata — the date and the original address of first publication — is not in my hands. So I hold three items as verifiable, not proven: the course titles, the author's credentials, and the release timeline. Anyone quoting a figure from my sentences should know what it rests on.

My first lesson in deep analysis was accounting, not statistics. Before any claim, its base. When someone says “this course is good,” I immediately ask three questions: how many said it, on what measure, and compared with whom. The announcement answers the first clearly — one reviewer. It answers the second partly — a club player's eye. It answers the third, but incompletely: the comparison to Ris and Sokolov is made, yet nobody states the size or direction of the gap. A review is a claim; its base is a sample of one, and a sample of one is not a trend.

I learned that rule the hard way, coding matches year after year. In May 2026, when the Bundesliga returned behind closed doors, I coded the first sixty matches. The home-win rate fell from 43.2% to 32.7%. A clean number — but it stood on sixty matches, directly against equivalent fixtures a season earlier. Empty stadiums gave us a control group we would never have dared to create, and when the crowd disappeared, the data stopped lying about home advantage. That base is exactly what the course review lacks. If the review's crowd is one person, the data stays silent — and I do not treat a silent dataset as a basis for decisions.

The second level is credentials. FIDE Trainer is a formal certificate that should be verifiable in FIDE's trainer registry; International Master is a playing title, usually implying a rating above 2400. A strong instructor and a strong player are not the same claim, and the review did not separate them. Playing strength does not guarantee course quality; neither does a trainer certificate. I admit a suspicion of my own here. Former stars opening academies reads to me mostly as branding, while genuine grassroots coach education is chronically underfunded. Entering the ChessBase author roster is precisely that branding-level event — respectable, useful, but not a system for producing a coaching generation. I accept the FIDE credential as a quality signal; I simply do not place the signal where the outcome belongs.

The third level is money. ChessBase's FritzTrainer line is monetized educational product; adding a new author expands supply and gives club players more options. Here my follow-the-money principle applies. The question is not “is the course good”; it is “who is paying for it, and what do they get back.” If a club player spends thirty to forty euros for two to six hours of video, the test is usable knowledge per euro — just as I measure expected goals per shot in football. Price, length, number of interactive exercises: three figures absent from the announcement, yet sitting at the center of the buyer's decision.

There is a revenue-model question too. FritzTrainer courses are usually bound to the ChessBase database software ecosystem, meaning a course can also push software adoption. That bundle effect is not mentioned in the announcement, but commercially it is central. When educational content becomes a driver of software sales, the quality of the course and the platform's interest do not always point the same way. My suspicion here is long-standing: when data and content are sold together under one roof, an independent voice becomes essential to measuring the buyer's interest.

The content itself is not in the wrong place. Calculation and positional play — these two middlegame fundamentals give club-level players the highest return. Not opening novelties, not memorized engine lines, but calculation accuracy and positional decision-making: that is where an amateur's rating moves fastest. Based on years of watching and coding games, I can say the majority of a club player's losses come from two sources — a calculation error and a positional misread. In that sense the two courses are not aimed at the wrong market. But if the target audience is club players, the success measure should also be the student's rating change before and after the purchase. A course advertisement is a story; a student's rating gain is a number. I wait for the number. The announcement lacks that number, because at release it cannot yet exist — but as a culture of verification, where and when it will be measured should be declared alongside the course.

If I were to write a real evaluation protocol for these two courses, it would look like this: record each student's current rating and their accuracy on a fixed tactical puzzle set; then one arm uses the course while a control arm uses a different course or none; after at least eight to twelve weeks, measure the rating delta and puzzle-accuracy difference between the arms; six months later, measure retention once more. Confounders — study time, prior rating, motivation — must be separated out. Most important, no publisher would fund this test from their own pocket, because a negative result is rarely published. In a product market where negative results go unpublished, I hold every positive review as a potentially biased sample. That is my null-result discipline: if the vanishing trend is not published, every remaining number becomes suspect.

Look at the comparison structure. Robert Ris and Ivan Sokolov both have long publishing histories in the training market. Placing Kavutskiy beside them tells the buyer: “this is a serious author, comparable to known names.” But the announcement never states Sokolov's title, gives no rating gap, shows no methodological difference. When a comparison carries no numbers, it becomes a marketing device rather than analysis. I am not saying the comparison is false; I am saying it is incomplete. Sokolov's over-the-board profile is publicly much higher — but that is my inference, not the announcement's fact, so I keep it in the data-pending drawer.

This is why the competition in the training market is not really a rating contest; it is a contest for attention and trust. Authors compete for the buyer's belief, and in that market the word “established” is itself a product. A new author entering is a mild signal of creative generational renewal — but it is never comparable to a young player breaking into the 2700 club. The scales differ, and I do not add them together.

This is where my standing spreadsheet earns its keep. I refresh it weekly: the list showing Bangladesh has produced five grandmasters — Niaz Murshed in 2026, Ziaur Rahman in 2026, Reefat Bin-Sattar in 2026, Abdullah Al Rakib in 2026, Enamul Hossain Razib in 2026 — and not one more in the following decade. The federation's annual budget sits near eleven lakh taka, and from 2026 to 2026 home norm tournaments were nearly invisible. I write these figures from my own record, with sources. A course can be bought; a pipeline cannot be bought, only built through years of investment.

From Russia I see the mirror image — a system where the state and clubs run chess schools as infrastructure, treating coach education as recurring cost. The gap between the two cultures is one of accounting, not talent. Yet I am not willing to treat the old null result as permanently true. What has shifted in Bangladesh late this decade — a fourteen-year-old national champion, a fifty-nine-year-old winner — belongs in the same frame as my old baseline. A model is a receipt, not an oracle; when the record changes, the model changes. ChessBase's new author is no receipt for that shift; it is a market row, not pipeline proof.

Now the counter-warning. A positive review is not evidence of quality — it is only a claim, and behind the claim sits one person's taste. A club player's praise is a demand-side signal, not supply-side proof of quality. The easy trap here is mixing correlation with causation: if the course sells well, someone will say “see, the course is good,” when the sales could come from marketing, from the author's existing audience, or from ChessBase's bundle strategy. None of those three causes is directly tied to the course's actual teaching quality.

There is another trap. The “established versus new” frame can itself be a marketing device. By seating a new name beside old ones, the publisher captures attention on both sides. For the buyer it is a convenience, because the comparison frame is ready-made; and a disadvantage, because the frame becomes a substitute for proof. I take Schwarz's review as a single voice, not a single truth. His identity as an experienced reader of the FritzTrainer series raises the credibility of the comparison, but it does not fill the missing numbers.

I append a paragraph to every piece: what this data cannot see. Here it is — no sales figures, no refunds, no completion rates, no student rating deltas, no independent second reviewer. On the basis of a single review I will reach no conclusion. The first public xG model was a receipt, not a prophecy; this announcement is a receipt too — only its sample has not yet been built. I built the Russian Premier League model because the silence was measurable; here the silence cannot be measured, because the publisher released no measurement.

Three signals to watch ahead. Customer ratings and sales signals on the course pages — if they surface, we will know whether Kavutskiy is entering a long-term relationship with ChessBase. Independent reviews — blogs, YouTube, forums — if multiple independent voices support or refute Schwarz's assessment, the quality narrative will hold. And a third course — if it appears within a reasonable period, the publisher's confidence will look durable. The question is not whether Kavutskiy can teach; the question is who is measuring whether the teaching landed, and who is keeping the receipt.