Samarkand 15/16 and the Najdorf Update: Two Stories Glued Into One Headline

I sat in front of a screen in Shenzhen at two in the morning, scanning the live scoreboard from Samarkand. The United States women's team had just closed Round 4 with 15 out of 16 board points. I reopened the source page to find the player list, the board numbers, the opponents, the colours. Nothing. The body of the piece was devoted entirely to a Najdorf opening video update by Daniel King, complete with a list of named move orders.

One page, two souls. The upper half carries a genuine competitive signal. The lower half carries product information. Both are worth reading, but they belong to different stories, and the fact that they share a single headline is the first lesson for anyone working with sports data: unverified data is just noise arranged neatly.

Context: board points, the Swiss system, and a sixty-year-old opening system

Team chess scores by board, not by match. A win is worth 1 board point, a draw 0.5, a loss 0. With four boards per match, the ceiling for a single match is 4 points, and the ceiling after four matches is 16. The United States women's team reached 15/16, meaning that across four matches they barely dropped anything.

Two layers of data need separating here. The first layer is match points, which decide the standings. The second layer is board points, which reflect the margin of victory. A team can win a match 3-1 and collect 3 board points; a team losing 1-3 still keeps 1 board point. The official standings run on the first layer. The 15/16 figure cited in the headline belongs to the second layer, and the second layer is always the more flattering one.

Samarkand, Uzbekistan, has become a familiar venue for continental and international federation team events in recent seasons. The women's event there brings together national teams and operates on the Swiss system: after each round, teams on similar scores are paired. The Swiss system has a property viewers routinely overlook — opponent quality rises with each round. Early rounds bring weaker teams, late rounds bring stronger ones. The difficulty curve slopes upward; it does not lie flat.

The rest of the source page belongs to an entirely different world. Daniel King is a veteran opening-content editor at ChessBase, known for the Power Play series. This update supplements the Power Play 18 DVD, a project that builds a Najdorf system for Black. The update names specific branches: 6 h3 e5 7 Nde2; 6 Bc4 with English Attack ideas; 6 Be3 against ...a5; 6 Be2; and early f3 setups. The author states that the DVD can be understood independently, though a foundation from the original set helps.

The Najdorf is one of the largest systems in the Sicilian Defence, named after Miguel Najdorf, reached after 1 e4 c5 2 Nf3 d6 3 d4 cxd4 4 Nxd4 Nf6 5 Nc3 a6. The move ...a6 is the signature: Black controls the b5 square, prepares ...e5 or ...e6, and keeps every queenside expansion option open. The English Attack is White's setup with Be3, f3, Qd2 and often 0-0-0 — an attacking structure that entered top-level practice in the late 1980s and has never gone out of fashion.

A repertoire — a pre-planned set of openings — is the central concept of this content. Strong players do not study isolated games; they own a decision tree validated against databases and analysis engines.

Core analysis: where 93.75 percent sits on the distribution

15/16 equals 93.75 percent of board points. In elite women's team events, a strong team facing a mixed field typically hovers around 55 to 65 percent of board points. A rate of 93.75 percent sits a distance above that band which historical data rarely sustains across a full tournament. An extreme rate in a small sample is always an invitation to regression toward the mean, not proof of absolute dominance.

The sample here is very small: 16 games. With 16 games, turning a single draw into a loss drops the rate to 87.5 percent; two games drop it to 81.25 percent. The noise band of a small sample is wide enough to swallow the entire gap between a champion team and a fourth-place team.

Opponent composition matters more than the number itself. Under the Swiss system, Rounds 1 through 4 are the softest stretch of the event. A top-seeded team usually meets lower-ranked teams in the first half, and in those matches the Elo gap between boards can reach 300 to 400 points. At that distance, the expected score of each individual board exceeds 80 percent, and the expected team board score exceeds 3.2 out of 4.

Missing data also deserves to be named. The source page provides no player list, no individual Elo, no board-by-board results, no opponent names, no time control (classical, rapid or blitz), no ACPL — average centipawn loss — and no engine match rate. Without those four data groups, no one can calculate a performance rating, and no one can distinguish a team playing well from a team enjoying an easy schedule.

I once spent three months working on opening data for a club. I spent three months learning that a beautiful chart is worth less than a correct process. My group built a frequency table of Najdorf branches appearing in more than ten thousand top-level games. The table showed that a branch favoured in one event can vanish after the next, simply because one leading player found a way around it. The shelf life of opening theory is measured in months, not years.

The first three branches on the list deserve separate reading. 6 h3 e5 7 Nde2 is a structure where White sends the knight to e2 instead of f3, keeping the f3 square free for pawn advances and sidestepping familiar tactical blows in the centre. 6 Bc4 avoids the English Attack template altogether, steering toward a position where White castles short and presses on f7. 6 Be3 against ...a5 is the classical English Attack with a twist: Black starts queenside action before White can castle long. Each of those branches already has thousands of games in the database, which is precisely why the value of an update lies in pointing out where the engine assessment has shifted.

On the opening-content side, the data situation is similar but differently shaped. The update names specific branches without attaching any statistics: no win rate, no draw rate, no game count from a database, no engine evaluation. The claim that the material is "developing all the time" is an author's opinion, not a statement backed by figures. Engine match rate and ACPL are the two standard measures for assessing the depth of opening material, and both are absent.

What deserves credit is the level of specificity. 6 h3 e5 7 Nde2, 6 Bc4, 6 Be3 ...a5, 6 Be2, 5 f3 — these are branches sitting at a high theoretical layer, where any single move can be refuted within months. This is an update aimed at a narrow audience: players who handle Black in the Najdorf and already own the original set. The content does not claim to be a new system; it is a continuation of an old one. That acknowledgment matters, because it shows the product is positioned more honestly than its headline.

Contrarian view: one headline, two disconnected facts

The structure of the source page is itself a form of data, and it is flashing a warning. The headline concerns the United States women's team in Samarkand. The body concerns a video product. The two subjects share no causal link, no data source, no author. Correlation in page position is not causation in content, and that is the trap fast readers fall into most often.

The most likely operational scenario is that a short report on the Round 4 result was published, and the product promotion was attached by an automated publishing workflow. A reader skimming past will fuse the two and believe the American women are advancing on some theoretical breakthrough. There is no basis for that belief.

Another counterpoint lies in the label "new ideas". In the opening-content industry, every update is called new. Real value only surfaces when the material is checked against current databases and recent top-level games. If over the next six months leading players begin playing exactly the branches the update recommends, that is verification. If not, the product may still be useful to amateur players, but it is not a theoretical turning point.

The same counterpoint applies to 15/16. A Round 4 score predicts very little about the final outcome. Team-event history contains teams that led early and collapsed over the final three rounds as the schedule steepened. It also contains teams that opened around 60 percent, accelerated, and won. Board points from Rounds 1 to 4 and board points from Rounds 7 to 9 are produced by two different competitive environments.

One more blind spot deserves stating plainly: even the phrasing "15/16" has not been confirmed as a cumulative figure after four rounds. It could be cumulative board points, it could be a different event indicator, and the source page does not define it. In data work, a number without a clear definition is worth only as a starting point for questions, nothing more.

This is where I remind myself of the limits of the tool. A Chinese club taught me that data is not the destination, but a walking stick. The stick helps a traveller stay upright on a steep stretch, but it does not know where the destination lies. Judgments about opponents, about form, about timing must still come from the person in the analyst's chair.

At the same time, I want to be fair to the number. If the United States women's team genuinely holds a 70 percent board-point band through Rounds 6 and 7, then the original 93.75 percent is no longer noise but the crest of a stable curve. A small sample can be the beginning of a large story. The problem lies in telling those two possibilities apart — and that distinction only arrives in the next three rounds.

Signals to track, and one open question

The first signal to follow is the pairings for Rounds 5, 6 and 7 in Samarkand. If the United States women's team meets the top seeds and still holds above 75 percent of board points, the regression-to-the-mean model begins to weaken. If they drop a 2-2 match or a loss, the set-the-pace narrative closes very quickly.

The second signal lies in the game database. Track whether the branches 6 h3 e5 7 Nde2, 6 Bc4, 6 Be3 ...a5, 6 Be2 and 5 f3 appear in top-level games over the next six months. This is the most objective test of the theoretical value of an opening product, and it does not depend on the manufacturer's own description.

The third signal is community response to the Power Play 18 update. Store ratings, discussion in chess forums, and independent reviewer feedback will indicate whether the audience-driven opening-content model remains commercially viable. It is a business model worth watching, because it turns a very specific demand from a narrow group of players into a product with a short life cycle that can be renewed.

Drawing on years of watching team events, I set three early-warning thresholds for myself. The first is the United States women's board-point rate falling below 70 percent in two consecutive rounds. The second is the appearance of a loss on board one or board two, because that is where opponent quality shows most clearly. The third is the emergence of a second team chasing with an equivalent board-point differential. If two of the three thresholds activate together, every conclusion about superior strength must be rewritten from scratch.

Data does not speak on its own. It only answers when we ask the right question, at the right moment, with enough patience to wait for the endgame. When the data does not lie, we are the ones lying to ourselves. The open question remains: what will change our assessment of the United States women's team in Samarkand — a defeat, or a run of draws on the middle boards?