Wesley So's Fourth US Chess Championship: A Tiebreak Win Over Caruana, and the Empty Ledger Beneath the Headline

The file landed on my desk around eleven at night. I opened it in Pune with cold tea on the coaster and a headline on the screen: Wesley So beat Fabiano Caruana in tiebreaks, wins fourth US Chess Championship. I started counting paragraphs. The first was about So. Paragraphs two through nine — eight of them — never mentioned him. They contained a Bulgarian IM, a six-video course, and Black's plans against 1.d4 sidelines. The arithmetic is simple. One of nine information points addressed the event. Eleven percent of the payload was carrying one hundred percent of the headline claim. I am not writing this because advertising offends me — chess media is financed by courses and everyone in the trade knows it. I am writing it because the story the headline threw away and the technical material the body buried both have ledgers, and my oldest professional habit is to reconcile a ledger the same day it goes wrong. The first number I trust is the one with a timestamp and a correction. This file had no timestamp. No publisher. No date. No score. The match being reported had no round, no game, no tiebreak time control. Everything below therefore splits in two: what was claimed, and what should have been there to support it and was not. The structure of the US Championship decides what the title means, so the structure first. It is an eleven-round round-robin at the St. Louis Chess Club — one hall, one stage, the same spectators in the same seats three days running. Classical control: ninety minutes for forty moves, thirty minutes to finish, thirty-second increment from move one. That control is how chess tells you who is durable. Whoever doesn't crack across eleven rounds is the champion. But a round-robin that ends level has a second half, and that half is the tiebreak. The American sequence has been roughly stable for years: two rapid games at twenty-five minutes with a ten-second increment, then two blitz games at five plus three, then Armageddon if needed — a colour bid, five minutes against four, draw odds to Black. A classical round-robin, built to measure slow endurance, settles in a room where the clock moves fast. So against Caruana is a matchup built for that structure, and it is not only about ratings. Both are elite classical technicians. So's strength is positional solidity, a low-theory, low-variance framework that has kept him near the top ten for years. Caruana's is calculation-heavy, preparation-driven chess, the same profile that made him the 2026 world title challenger. In the games I have logged between them, the draw count has been high. That is not a weakness. It is what happens when two players know each other's soft spots and see no reason to open them early. Which is where the tiebreak's real function begins. When two classical technicians cannot separate across eleven rounds, the tiebreak does not extract the difference between them. It asks them a different question. Not who plays better chess, but who handles a faster clock better. Those are two measurements and calling both of them 'champion' is the oldest habit in the record book. Caruana is not unfamiliar with that question. In November 2026 in London, all twelve classical games of the World Championship were drawn, and Magnus Carlsen won the rapid tiebreak three to nothing. What happened that day was not a verdict on chess strength; it was a verdict on format. Now, seven years later, in a national title, the same structure, the same player, the same outcome. Two events do not make a trend — I will not take that easy conclusion. But two events do establish that the format is stable, and that a stable format produces an outcome distribution different from the classical one. That is arithmetic, not complaint. On the upper floor of the St. Louis club on tiebreak day something happens that no scoresheet records. The hall has wooden tables, a digital clock on each with a red light visible at the board's edge, and in the corridor stand the people with no game to play — parents, coaches, club volunteers, an elderly couple who come every year and stand in the same corner every year. No scoresheet records any of it. I will not claim it matters. I will only say that part of the room where a title is decided is invisible to every analytical tool I own, and it reminds me annually that my spreadsheet is incomplete. Now the player who won it. So's rapid record is not a discovery of mine — it is filed. In December 2026 in Moscow he won the World Rapid Championship. When a player holds a world title in the exact time control that decides a national title, calling the tiebreak luck is an insult to the data. That is the field where he is measurably strong. There is a further layer, and it runs back to 2026. That year the US Championship was played online because of the pandemic, and So won that title too. I logged that period's online events in my notebook, including the FIDE Online Olympiad, where India and Russia were declared joint winners after server failures. That experience taught me something I still repeat: online Elo and over-the-board Elo do not belong in the same column. They are different games that were given one name. And the same applies to counting titles. The headline says fourth. Four is an integer, and the integer is the most comfortable lie in any record. His list shows 2026 in the classical format in St. Louis. It shows 2026, played online. The remaining editions had their own conditions. The 'four' is therefore a sum of four things that were not won under one rulebook. I am not saying the fourth is small — it is complete and it carries its own weight. I am saying that if a career is to be compressed into one digit, the digit needs a footnote, and no headline prints footnotes. I left the copy desk, but the copy desk never left the model. Seventeen years ago on a Pune sports desk I set results tables, and there was a rule that survives in no style guide: a number that has not been checked does not go on the page. A count of titles is exactly that kind of number. Writing four takes a second and a half; explaining which four, under which conditions, takes a paragraph. Now the money. Who paid for this title is a more interesting question than the podium. The American structure is private — the St. Louis Chess Club, and behind it Rex Sinquefield's foundation, which turned a city into a chess capital inside a decade. Recent prize funds have sat near two hundred thousand dollars with the winner's share announced separately. No state association, no election cycle, no suspensions, no public-sector job ladder. A trust writes, a club runs, a hall hosts. For an Indian reader the comparison matters, because our title-producing machine works differently — state government cash awards, the ONGC and Railways job ladder, the Velammal school pipeline, AICF election cycles. There the money passes through many hands and every hand leaves a receipt, which makes audit hard. Here it passes through one hand, which makes audit easy, and still nobody does it. Both systems make champions; only one lets you trace the money with a finger. Back to the file that started this. List its nine information points and the picture resolves. Point one: So, Caruana, tiebreak, fourth title. Point two: White players avoid Slav, Semi-Slav, QGA and Orthodox Queen's Gambit theory and therefore decline 2.c4. Point three: these systems grew from that avoidance. Point four: four anti-theory systems are 'extremely dangerous'. Point five: Black must prepare for each. Point six: six instructive videos. Point seven: one plan per opening. Point eight: the author is IM Valeri Lilov, a Bulgarian trainer. Point nine: Catalan, Blackmar-Diemer Gambit and Richter-Veresov are also covered. That list is not news. It is a product that used news as a doorway. The form is news-hook-led native advertising, and in chess media it is common enough that nobody flinches: news traffic arrives, the reader moves to a course page, the course sells, next month's coverage gets a budget. No conspiracy, just a business model — and in a small ecosystem close to a necessity, because long-form chess journalism does not run on subscriptions. It runs on course sales. But the reader pays in a currency that is not money. He cannot tell which claim has been checked and which was written to sell. When a title claim and a product claim sit in the same paragraph in the same tone, the reader's filter shuts off. The trade name for this is an unattributed source, and the cure for an unattributed source has never been censorship. The cure is disclosure. Now the technical audit, because the course is the only solid material in the payload and therefore deserves the scrutiny. IM Valeri Lilov is a Bulgarian player and trainer, known online as 'Tiger Lilov', with a coaching-oriented rather than competition-oriented profile. That matters: coaching material teaches plans rather than forced lines, and the article's own phrasing — one relevant plan for Black for each opening — matches that profile. The first problem is arithmetic. The article promises six videos and then lists seven openings: Colle, Trompowsky, Torre, London, Catalan, Blackmar-Diemer Gambit and Richter-Veresov. Possibly one video covers two systems. If so, the article should say so, because six against seven is a difference in coverage and runtime for the buyer. I will not spend more than this paragraph on it — it changes no result, no ranking and no payout. I file it because a product with a counting error raises the question of who checked the lines. The second problem is serious and it is categorical, not numerical. The Catalan is placed in this list as an anti-theory sideline. The Catalan is not an anti-theory sideline. It runs 1.d4 Nf6 2.c4 e6 3.g3 d5 4.Bg2 and it is a mainline elite opening contested on 2700 boards, with its own vast theory and its own strategic character. Grouping it with London, Colle and Torre misdescribes both its theory load and its nature — the equivalent of putting a fast bowler and a spinner in one table and calling both of them bowlers. The third problem is structural and the most expensive for a buyer. The article's entire framing sits on 1.d4 d5: White does not play 2.c4, so what should Black do. But three of the listed openings arrive through 1.d4 Nf6. Trompowsky is 1.d4 Nf6 2.Bg5. Richter-Veresov is 1.d4 Nf6 2.Nc3 d5 3.Bg5. Catalan starts 1.d4 Nf6. Only Colle, London and the Blackmar-Diemer Gambit genuinely fit the 1.d4 d5 frame. A buyer expecting a coherent anti-1.d4 d5 repertoire gets heterogeneous move orders and must decide himself what to play when. The fourth problem is an absence, and absences are informative. No sample line. No engine-verification claim. No ACPL. No database statistics. No runtime. No sample video. 'Extremely dangerous' is asserted without one illustration. In my experience, repertoire products that have been engine-checked advertise the fact loudly — it is a selling point. Where the claim is silent, I treat the silence as evidence, and I say plainly that this is inference, not proof. Against all that, and this is a correction I owe: the market is real and large. Below 2200, White players rationally avoid Slav, Semi-Slav, QGA and Orthodox Queen's Gambit theory, because they have four hours a week, finite memory, and an opponent who is equally prepared. The London System did not grow because players got lazy. It grew because the system delivered on its promise — less theory, clear plans, a repeatable structure. On public opening databases the frequency of London-type systems below 2026 tells exactly that story, and I have had that figure written in my notebook for years. Here my old inversion applies, and I will state it flatly. Video courses have done more for opening literacy at club level in the last decade than any new edition of an opening encyclopedia. That is not opinion, it is a fact about distribution: knowledge moved to a format where it can be watched, paused, rewound and watched again. The generation that learned theory alone at a board from a book undervalues how fast a plan is absorbed on a screen. I will not stand on the nostalgia side of this argument, because I know the pre-2026 reality: no audience, no circuit, one specialist desk at a single newspaper. Calling that era golden is writing history without dates. But the inversion has a second edge, and it stops me from praising the course. Video increased reach, and reach replaced verification. When the medium is video, the buyer sees only what is shown — he cannot pause a line and put it into an engine, cannot match a variation against a database. Verification therefore sits on the seller more heavily than ever, and it is the thinnest thing in this article. Now the part where I have to argue with my own reflexes. My first instinct says So won this title in a tiebreak, so it is half a title. I will write the strongest version of that before knocking it down: Wesley So lost to nobody across eleven rounds, he did what a round-robin demands, and the tiebreak is part of a structure everyone knew before the first move. Changing the rule after the result is the definition of an excuse. That argument is strong and it is correct. So I challenge the label, not the title. The problem is not that So won. The problem is that we call the result a classical championship when it was settled in rapid, blitz and possibly Armageddon. It is the same error we make with bowling figures when we put ODI and T20 economy rates in one table — both numbers are right and neither is comparable. A reader never told which time control decided the title will draw a wrong conclusion about the player who won it. And here I have to fight my second reflex, the contrarian one. It is easy to say the media sold the story. The truth is less dramatic: in a small ecosystem there is no other buyer for long-form journalism. Course sales are the money that keeps the coverage alive. So 'sold out' is a false charge. The accurate charge is that the money path was hidden — and hiding it was unnecessary. One sentence would have done: this material promotes a commercial course. The absence of that sentence is the whole problem, not the advertisement. So next time I read a tournament report I will do what I always do. I will count. I will look at the clock. I will write down the time control the game finished under, because that is what tells you what was actually measured that day. So's fourth title stands, and his career ledger accepts it. But if next year's round-robin ends level again and the title is decided on a twenty-five-minute clock, we will have to ask what we expect 'national champion' to measure. If the answer is endurance, either the tiebreak changes or what we call things changes. For the other half — where courses sell and news does not — I have no reform to propose, only an expectation I am not putting forward as evidence. Next time a trainer tells me a system is extremely dangerous, let three moves be shown, let one line be engine-checked, and let the page say the material is commercial. Then it is no longer information I have to audit. Before I call it an edge, I make the spreadsheet confess. This time the spreadsheet said one out of nine. Next time we will see what it says.