Brazil's round 7 opponents at the Olympiad: how Tajikistan and Vietnam play

Short answer

In round 7 of the 46th Chess Olympiad, on 23 September 2026, Brazil face Tajikistan in the open section, on table 21, and Vietnam in the women's section, on table 15. Drawing on 528 classical games broadcast on Lichess over the last three years, this article shows each player's repertoire in both squads, the lines where each scores below their own average, where the engine finds the most errors and what the online repertoire adds — always with the number of games next to each percentage, because that number says how much of the dossier holds up.

Event: 46th FIDE Chess Olympiad

How this was compiled: Researched on 21 September 2026, with round 6 finished and round 7 still to be played: it starts on 23 September at 3 pm Samarkand time (10:00 UTC), after the rest day. The pairings were checked on chess-results; the official PGN with each team's line-up only comes out when the round starts, so this dossier covers both opponents' full squads rather than the four boards. The repertoire uses the 528 classical over-the-board games from the last three years broadcast on Lichess, linked to each player by FIDE ID — 363 from the Tajik squad and 165 from the Vietnamese. The error profile measures each player's 12 most recent games with Stockfish 19 at depth 16; for Mustafokhuja Khusenkhojaev it was 11, because one game could not be read and was logged as a failure. Every move quoted in the text was re-measured at depth 24, with a dedicated search and a clean hash, and endgames of up to seven pieces were checked in the Lichess tablebase. The blitz and rapid repertoire comes from verified Chess.com accounts, over a 90-day window.

The essentials in a few lines

• In round 7, on 23 September 2026, Brazil face Tajikistan in the open section (table 21) and Vietnam in the women's section (table 15). All four teams have 8 match points. The line-up only appears in the official PGN when the round starts, so this dossier covers all ten players in both squads.
• Nine of the ten opponents open 1.d4 in most of their White games, and in the 270 White games in the sample 1.e4 appears five times, never twice from the same player. The exception is Mekhriddin Sharifov, who opens 1.Nf3 or 1.c4 in 15 of 19.
• Three Tajiks score below their own average after 1.d4 Nf6 2.c4 e6: Muhammad Khusenkhojaev scores 38.9% in 9 games, Jaloliddin Ilkhomi 56.5% in 31 and Mustafokhuja Khusenkhojaev 65.2% in 23. With 3...c5 after 3.g3, Mustafokhuja drops to 50.0% in 7 and Ilkhomi to 44.4% in 9.
• Luong Phuong Hanh scores 30.4% with Black in 28 games against opponents who are, on average, at her own level. In the Sicilian with 2...e6, her reply to 1.e4, she scored 10.0% in 5 classical games, and the same hole shows up in her rapid games on Chess.com: 23.5% in 17.
• In round 6, played today, two Tajiks got away with draws from endgames the tablebase calls decided. Mekhriddin Sharifov had a won ending against David Navara and played 64...Kf4, which draws; Jaloliddin Ilkhomi walked into a lost ending with 74.Ka4 and drew anyway.

The round 7 picture

Brazil reach the seventh round after drawing 2–2 with Belgium in the open section and beating Paraguay 4–0 in the women's. All four teams in these matches have 8 match points. In the open, Tajikistan have 14.5 game points to Brazil's 13; in the women's, Brazil have 17 to Vietnam's 16.

On rating, the two matches point in opposite directions. In the open, Brazil are seed 31 and face seed 61. In the women's section it is the other way round: Brazil are seed 44 and face seed 21, a team that opened the Olympiad with three wins, lost 1–3 to Poland, beat Norway 4–0 and lost 0–4 to France in round 6. Tajikistan are also coming off a defeat: 1–3 against the Czech Republic, with two draws that the engine section covers in detail.

The sample column in the table below says how much of each dossier holds up. Jaloliddin Ilkhomi has 131 broadcast classical games in three years, the largest in either match; Vo Thi Kim Phung has 15 and Bach Ngoc Thuy Duong 16. None of the ten falls below 10, the point at which the dossier flags low confidence, but the two Vietnamese players only played Olympiads in this window: all 15 of Vo's games and all 16 of Bach's come from the Budapest and Samarkand Olympiads. That is 528 games in total, 363 Tajik and 165 Vietnamese.

Only Ilkhomi and Bach have played all six rounds. Mustafokhuja Khusenkhojaev, top board on the Tajik list, sat out round 6; Nguyen Thi Mai Hung sat out rounds 4 and 6. Olympiad teams rotate their squads, which is why the dossier covers all five on each side instead of trying to guess four.

Tajikistan: four 1.d4 players and one exception

Preparing for Tajikistan starts with a rare certainty. Mustafokhuja Khusenkhojaev opened 1.d4 in 43 of the 44 White games broadcast over three years, Muhammad Khusenkhojaev in 39 of 43, Jaloliddin Ilkhomi in 60 of 70 and Amin Kabilov in 8 of 10. It is not only the first move that repeats. Mustafokhuja's and Ilkhomi's most-played openings are the same three — the Catalan, the Fianchetto Benoni and the Bogo-Indian with the bishop retreat — all born from 1.d4 Nf6 2.c4 e6 3.g3. The team has a home, and it is on g2.

That home is exactly where the most useful edge in the match shows up. After 1.d4 Nf6 2.c4 e6, Muhammad Khusenkhojaev scores 38.9% in 9 games, against his own 62.8% White average; Ilkhomi scores 56.5% in 31, and Mustafokhuja 65.2% in 23, below his 69.3%. When Black meets 3.g3 with 3...c5, the number falls further: 50.0% in 7 games for Mustafokhuja and 44.4% in 9 for Ilkhomi, with six draws. These are samples of seven and nine games, a medium-confidence hint rather than a finished plan, but it is the same hint in two different players, which is worth more than a large number for just one.

Mustafokhuja Khusenkhojaev has the best White score in the squad, 69.3% in 44 games, and is its most predictable player. With Black he drops to 53.7% in 41 and has no clear weak spot: against 1.e4 he scores 52.3% in 22 games, close to his own average, answering 1...e5 in 21 of them. The Italian with 3...Bc5 4.d3 gave him 33.3% in 3, a sample that supports nothing.

Muhammad Khusenkhojaev is the opposite case, and the one the table most misleads about. He scores 69.8% with Black in 43 games, the best Black score in either squad, in a sample where his opponents average 2259 — about 150 points below him. It is the number of someone who played a lot of opens and beat the weaker players, not of someone who neutralises a master's White. Against 1.e4 he drops to 63.6% in 22, and within that 1.e4 e5 2.Nf3 Nc6 gives him 57.9% in 19, the team's most reliable weak line by sample size.

Jaloliddin Ilkhomi has the largest sample in the match and the most exploitable asymmetry. With White he scores 60.7% in 70 games; with Black, 40.2% in 61. Against 1.e4 the number drops to 32.0% in 25 games, with two wins, and the French he chooses in 15 of them gives him 33.3% — 28.6% in the 14 that continue 2.d4 d5. The Caro-Kann is no better: 31.2% in 8.

Mekhriddin Sharifov is the squad's top-rated player, at 2419, and the only one who does not play 1.d4: he opens 1.Nf3 or 1.c4 in 15 of his 19 White games and scores 68.4%. With Black he drops to 40.9% in 22, and against 1.d4 to 28.6% in 7. His rating has risen quickly — the median in the sample is 2259 — so the older games describe a weaker player than today's. Amin Kabilov, the fifth, has 20 games and the longest in the team: a median of 101 half-moves, with 75% of games going past move 40.

Vietnam: five 1.d4 players and a Sicilian that does not work

Vietnam repeat Tajikistan's pattern with even less variety. Nguyen Thi Mai Hung opened 1.d4 in all 21 White games in the sample, Bach Ngoc Thuy Duong in all 9, Pham Le Thao Nguyen in 19 of 20, Vo Thi Kim Phung in 6 of 7 and Luong Phuong Hanh in 21 of 27, with 1.Nf3 in 5 of the other 6. Brazil go into the women's match knowing the first move of the opposing White player on four boards, and the work shifts to a different question: what to do with their own White.

That is where the best news of the match lies. Luong Phuong Hanh scores 30.4% with Black in 28 games, and the number is not an effect of strong opposition: her opponents in the sample average 2196, against her own 2210. Against 1.d4 it is 12.5% in 8, with seven losses; against 1.e4, 36.7% in 15. Against 1.e4 she answers with the Sicilian, and the 2.Nf3 e6 version gave her 10.0% in 5 games. Five games are few, and the dossier flags low confidence — but the online accounts section shows the same hole in a different sample, and two independent samples pointing to the same place are worth more than one.

Pham Le Thao Nguyen, the squad's top player at 2348, has the most balanced profile: 57.5% with White and 52.5% with Black, 20 games each. Against 1.e4 she drops to 40.0% in 10, and the Caro-Kann she plays in 6 of them gives 41.7%, with three losses. With White, after 1.d4 Nf6 she scores 50.0% in 13, below the 60.5% she makes when she opens 1.d4; 1.d4 d5 gives her 83.3% in 6. Whoever has Black against her meets more resistance after 1...d5 than after 1...Nf6.

Nguyen Thi Mai Hung plays more outside the Olympiad than her teammates — Biel, the Doeberl Cup and the Melbourne Open, in 36 of her 39 games — and has the most middling numbers: 54.8% with White, 44.4% with Black. Against 1.e4 she scores 38.9% in 9, almost always with the French, which gives her 42.9% in 7. The Slav gave her 20.0% in 5, a small sample.

Vo Thi Kim Phung and Bach Ngoc Thuy Duong have the prettiest numbers and the narrowest samples. Bach scores 77.8% in 9 White games and 64.3% in 7 with Black; Vo, 64.3% and 56.2%. All of it comes from two Olympiads, where part of the rounds are against much weaker teams. The only slice the dossier can register is Bach's in the Nimzo-Indian, 66.7% in 6, below her White average — and still a high number.

Where the engine finds the errors

The table shows the average evaluation loss per move in each phase, with the number of moves measured alongside. The opening column is the most uniform in both squads, from 5 to 9, and confirms what the repertoire already said: these are prepared players who reach the middlegame without having lost anything. The middlegame is another matter. Sharifov loses 41 centipawns per move over 368 moves, the largest middlegame base in the match, and Muhammad Khusenkhojaev 42 over 221. Jaloliddin Ilkhomi and Pham Le Thao Nguyen sit at 21, the cleanest in that phase.

The most expensive error in the sample was played today. In round 6, against David Navara, Mekhriddin Sharifov reached a rook-versus-three-pawns ending with his king on e4 and a pawn on h5. The Lichess tablebase, which gives the exact result with up to seven pieces, says the position was winning and that four moves won — 64...h4, 64...Rf1, 64...Ra5 and 64...Ke5. Sharifov played 64...Kf4, which the same tablebase calls a draw, and the game was drawn. In July, at the Oskemen Open, he was +5.25 against Assel Serikbay and played 51...Bc3: the evaluation went to −9.70 and he lost.

In the same match, Jaloliddin Ilkhomi went the opposite way. Against Nguyen Thai Dai Van, in a rook versus rook and two pawns ending that the tablebase calls a draw, he played 74.Ka4 and walked into a lost position — mate in 57 by the tablebase. His opponent did not find the way and the game was drawn. In round 3, against Libya, the story was similar: 46...Re2+ took the evaluation from −0.31 to −3.53 when re-measured at depth 24, and Ilkhomi drew. At Oskemen in July he let two five-pawn advantages slip — 41...Rxa2 against Aziz Degenbaev, which ended in a draw, and 60.Kb3 against Erasyl Zhakan, which he won anyway.

On the Vietnamese side, the pattern is an advantage that gets away. Against Niina Koskela of Norway, in round 5 of this Olympiad, Luong Phuong Hanh was +6.77 when she played 25...Bh2+, and the evaluation went to zero; she won all the same. At the 2026 Asian Women's Championship, against Zol-Uchral Erdenesaikhan, she was +5.20 and played 42...Qe3: the evaluation went to −6.14 and she lost. Nguyen Thi Mai Hung drew with GM Diptayan Ghosh in Biel after two moves the engine calls losing, 52...h4 and 60...a3 — the second took the evaluation from −0.15 to −10.73. Pham Le Thao Nguyen lost to Srija Seshadri at the 2025 Asian Championship after 70.Ra8, which took an ending from −0.64 to −5.34.

The high numbers in the table do not always describe errors. Amin Kabilov's endgames (43 centipawns over 156 moves) and Luong Phuong Hanh's (46 over 111) are the most expensive in the match, and the re-measurement shows why: Kabilov's three worst moments and Luong's two worst were played in positions that were already lost, several of them with mate on the board. The same goes for Vo Thi Kim Phung, whose five worst moments all come in decided positions. It is the pattern that repeats every round in this series: at depth 16, a settled position produces losses of thousands of centipawns for moves that change nothing, and the errors that changed a result showed up with modest losses.

The blunder count stays out of the table for the same reason: the classification uses the uncapped loss, so a lost player who picks the opponent's fastest mate adds a blunder without having changed anything. The average loss per move caps each evaluation at a thousand centipawns and does not suffer from the same problem.

What the online accounts show, and what they do not

This section answers a different question from the previous one, which is why it stands apart. A blitz and rapid repertoire is usually wider than a classical one, and what someone plays with three minutes is not what they prepare for a five-hour game. It tells you what they know and where they feel uncomfortable, not what they will play on Wednesday.

The most useful case is Luong Phuong Hanh's. On the amylazy account, with a WGM title verified by the platform, she played 287 games in 90 days and scored 35.4% with White and 37.0% with Black. In the Open Sicilian she plays with 2...e6, 1.e4 c5 2.Nf3 e6 3.d4 cxd4 4.Nxd4, she scored 23.5% in 17 games, which the dossier flags as high confidence. It is the same structure in which she scored 10.0% in 5 classical games. Two samples, two time controls and the same result do not prove she will lose again, but they do say she is not at home there.

Pham Le Thao Nguyen, on the fh2411 account, repeats the Caro-Kann she plays over the board: it is her most frequent Black opening in 298 online games. Against the Exchange Variation, with 3.exd5, she scored 0.0% in 6. In classical games, the Caro-Kann gives her 41.7% in 6. With White she opens 1.d4 in 136 of her 147 online games, and the dossier records 27.8% in 18 after 1.d4 Nf6 2.Nf3 d5.

Nguyen Thi Mai Hung plays online at the same level as over the board, 54.3% with White and 46.6% with Black in 300 games, and opens 1.d4 in 149 of her 151 White games — the loyalty to the first move holds in blitz too. Jaloliddin Ilkhomi has 32 games, with 47.1% as White and 46.7% as Black: too small a sample for lines, enough to say he opens 1.d4 online as well, in 15 of 17. Bach Ngoc Thuy Duong has 2 games in the window, and two games are not a sample.

Five opponents have a verified account and no games in the last 90 days — Mustafokhuja and Muhammad Khusenkhojaev, Mekhriddin Sharifov, Amin Kabilov and Vo Thi Kim Phung, whose account has been inactive since 2022. The account exists and is idle, which is different from the account not existing. Sharifov's lists Palestine as its country, against his Tajik FIDE registration; it was included because the Chess.com player page, which links the athlete to the account, points to it.

What to watch in round 7

The practical question in the open section is the defence against 1.d4, and the dossier answers with a direction, not a variation: Black players who go 2...e6 and accept the Catalan game find the Tajiks below their average, and 3...c5 after 3.g3 is where Mustafokhuja and Ilkhomi score least. With White, the target is Ilkhomi, who scores 32.0% in 25 games against 1.e4. Muhammad Khusenkhojaev is the warning: his 69.8% with Black looks frightening and comes, in large part, from opponents 150 points weaker.

In the women's section Brazil are the lower-rated team, and the dossier does not change that. What it offers is two concrete points. Against Luong Phuong Hanh, 1.e4 leads to the 2...e6 Sicilian that fails in both of her samples; against Pham Le Thao Nguyen, the Caro-Kann earns her little in classical games and nothing against the online Exchange. Both are hints built on five to seventeen games — enough to choose an opening the night before, not to promise a result.

When the round starts and the official PGN brings the line-up, the dossier file attached to this article will show the four boards and each table's game, with no need for the text to change. Until then, what is here is the full squad. Ten opponents, 270 White games and five 1.e4s: rarely does the eve of a round come with such a well-defined question.

Frequently asked questions

Update log

• 2026-09-21 — First published, with the round 7 pairings checked on chess-results and the round still to be played: repertoire from 528 classical games broadcast on Lichess, error profile with Stockfish 19 at depth 16 and the blitz and rapid repertoire of the verified accounts. The dossiers cover both opponents' full squads; when the line-up comes out, the same file will show the four boards and the round's game.
• 2026-09-29 — The Olympiad ended on 27 September: Uzbekistan won the open and China the women's event. This coverage is no longer updated; the final result is in the standings story.

Sources

• 46th Chess Olympiad Samarkand 2026, Open — round 7 pairings — Chess-Results
• 46th Chess Olympiad Samarkand 2026, Women — round 7 pairings — Chess-Results
• David Navara v Mekhriddin Sharifov, round 6 — broadcast game — Lichess
• Luong Phuong Hanh (amylazy) — titled profile that verifies the account used for the online repertoire — Chess.com

Related guides

• How to build an opening repertoire: 1.c4 as White and the Sicilian as Black
• Corporate chess leagues: what changes when the point belongs to the team
• How to bounce back after a loss in a chess tournament: a three-step mental reset