Canada's round 7 opponents at the Olympiad: how New Zealand and Hungary play

Short answer

In round 7 of the 46th Chess Olympiad, on 23 September 2026, Canada face New Zealand in the open section, on table 24, and Hungary in the women's section, on table 11. Drawing on 733 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 733 classical over-the-board games from the last three years broadcast on Lichess, linked to each player by FIDE ID — 326 from the New Zealand squad and 407 from the Hungarian. The error profile measures each player's 12 most recent games with Stockfish 19 at depth 16. 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, Canada face New Zealand in the open section (table 24) and Hungary in the women's section (table 11). Canada are level on match points in both — 8 in the open, 9 in the women's — and ahead on game points in both. The line-up only appears in the official PGN when the round starts, so this dossier covers all ten players in both squads.
• Daqi Mao has the best numbers in either squad, 81.2% with White in 56 games and 71.1% with Black in 57, but his opponents averaged 1966, more than 300 points below his 2277. Most of his sample is Auckland weekenders and club championships.
• Felix Xie, New Zealand's top board at 2423, opens 1.c4 or 1.Nf3 in 35 of his 50 White games. With Black he plays the Sicilian in all 26 games against 1.e4, and the version with 2.Nf3 e6 gives him 43.3% in 15, flagged with high confidence.
• Szidonia Lazarne Vajda has lost three games in three years, and two of them came at this Olympiad: from a level position against Song Yuxin in round 4 and against Varvara Poliakova in round 6, played today. In round 3 she let a +10.04 position slip into a draw.
• Kata Karacsonyi opened 1.e4 in all 56 White games in the sample. With Black she scores 44.4% in 71 games, and against 1.Nf3 she drops to 25.0% in 10.

The round 7 picture

Canada reach the seventh round after beating Cyprus 3.5–0.5 in the open section and Mexico 2.5–1.5 in the women's. Both matches are level on match points, and in both Canada lead on game points. In the open, Canada and New Zealand have 8 match points, with 16 game points to New Zealand's 15; in the women's, Canada and Hungary have 9, with Canada on 18 game points to Hungary's 15.

On rating, the two matches point in opposite directions. In the open, Canada are seed 41 and face seed 71, a New Zealand team that lost only to Norway and to Uzbekistan's second team. In the women's section Canada are seed 38 and face seed 18: a Hungary that lost 0.5–3.5 to China, beat Brazil 3–1 in round 5 and drew 2–2 with Belarus today.

The sample column in the table below says how much of each dossier holds up. Kata Karacsonyi has 127 broadcast classical games in three years, the largest in either match; Tom Middelburg has 30. None of the ten falls below 10, the point at which the dossier flags low confidence. That is 733 games in total, 326 from New Zealand and 407 from Hungary.

Only Davaademberel Nomin-Erdene has played all six rounds. Middelburg has not played since round 4, Karacsonyi sat out rounds 4 and 6, and Lazarne Vajda missed rounds 2 and 5. Olympiad teams rotate their squads, which is why the dossier covers all five on each side instead of trying to guess four.

New Zealand: a flank player on top board and a number that needs context

New Zealand's squad does not share a repertoire; each player brings his own. Tom Middelburg opened 1.e4 in all 13 White games, Daqi Mao 1.d4 in 51 of 56, Felix Xie 1.c4 or 1.Nf3 in 35 of 50, and Sravan Renjith and Daniel Gong split between 1.e4 and a second move. The Black replies are just as fixed: Middelburg the Caro-Kann, Mao the French and Xie the Sicilian, each in every game against 1.e4 in the sample.

Felix Xie is the top board, at 2423, the New Zealander with the largest sample against strong opposition — 101 games, against opponents averaging 2309 — and the one with the most useful weak line in the match. Against 1.e4 he played the Sicilian in all 26 games and scores 51.9%, below his 56.9% Black average; within that, after 2.Nf3 e6 he scores 43.3% in 15 games, a line the dossier flags with high confidence. Against 1.d4, the 1...Nf6 2.c4 e6 set-up gives him 45.0% in 10. With White he makes 75.0% with 1.c4 in 18 games and 64.7% with 1.Nf3 in 17; when he opens 1.d4, the number falls to 50.0% in 13.

Tom Middelburg has the smallest sample, 30 games, and the clearest asymmetry: 57.7% with White and 41.2% with Black. Against 1.e4 he played the Caro-Kann in all 11 games and scored 36.4%, with five losses; the Advance with 3...Bf5 4.h4 gave him 16.7% in 3. It is the most readable Black repertoire in the squad, and one of the weakest.

Daqi Mao is where the table needs a correction. He scores 81.2% with White in 56 games, with 41 wins, and 71.1% with Black in 57 — the best numbers in either squad. His opponents, however, averaged 1966, more than 300 points below his 2277, and most of his games come from Auckland weekenders, club championships and the North Island championship. After 1.d4 d5 he won all 19 of his games, a figure that describes the field rather than the opening. Against opposition at his own level the dossier has almost nothing to say about him, and the error section adds a second reason for caution.

Sravan Renjith is the most balanced: 62.5% with White and 64.3% with Black, with only six losses in 41 games. Against the Caro-Kann with White he scores 40.0% in 5. Daniel Gong scores 56.8% with White and 44.7% with Black; against 1.d4 as Black he drops to 40.0% in 5, and against 1.c4 he lost all three games in the sample.

Hungary: a squad of 1.e4 and a player who is not herself this Olympiad

Zsoka Gaal, top board at 2357, has the most balanced and least exploitable profile in the Hungarian squad: 59.8% with White in 46 games and 51.0% with Black in 50, against opponents averaging 2310. With White she opens 1.d4 in 29 games, 1.Nf3 in 11 and 1.c4 in 6. With Black she answers 1.e4 with the French in 12 of 26; the Winawer with 4.e5 gave her 40.0% in 5. Nothing in her dossier reaches the medium-confidence line.

Davaademberel Nomin-Erdene, who has played all six rounds, is the Hungarian with the clearest weak spots. With White she opens 1.e4 in 35 of 47 games and scores 68.1%, but against 1...e5 2.Nf3 Nc6 and against the Sicilian with 2...Nc6 she falls to 50.0% in 8 and 50.0% in 7. With Black she scores 53.7% in 54, and the Queen's Gambit Declined with 3.Nc3 gives her 35.7% in 7, flagged with medium confidence. Her Caro-Kann is worse still, 16.7% in 6, but the sample is small.

Kata Karacsonyi has the largest sample in the match and the most predictable White: 1.e4 in all 56 games, with 57.1%. Against the French she scores 33.3% in 6, and against 1...e5 2.Nf3 Nc6, 46.2% in 13, a line flagged with high confidence. With Black she is the weakest in the squad: 44.4% in 71 games, with 28 losses, against opponents averaging 2173, about her own level. Against 1.Nf3 she scored 25.0% in 10, the best starting move against her in the dossier.

Petra Papp is the most drawish: 10 draws in 17 White games, 15 in 25 with Black. Against 1.d4 with Black she has not won a game in the sample — 38.9% in 9, with seven draws — and her King's Indian, 1.d4 Nf6 2.c4 g6, gives her 30.0% in 5. Against 1.e4 she plays the Caro-Kann in 8 of 10 and scores 62.5%.

Szidonia Lazarne Vajda has the fewest losses in the squad, three in 41 games, and the best Black score, 62.5% in 16. The engine section explains why this Olympiad is an exception: two of those three losses came in Samarkand, in the last three games she played. With White she opens 1.d4 in 16 of 25 and scores 71.9% there.

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 between 5 and 8 for nine of the ten players; Daniel Gong is the exception, at 11. In the middlegame the range is wide, from 17 for Renjith and 18 for Gaal to 42 for Papp and 57 for Mao. Mao's 80 in the endgame is the highest number in the table and the least informative: his five worst moments were all played in positions with mate or a decisive advantage already on the board, against opponents who were often 300 points weaker. Felix Xie's 44 in the endgame has the same cause — his two worst moments came in positions with mate on the board, one for each side.

Szidonia Lazarne Vajda's case is the opposite, and it was all played in Samarkand. In round 3, against Janelle Mae Frayna of the Philippines, she was +10.04 when she played 21.Rd8 instead of 21.Rxe8+; the evaluation dropped to +0.14, and two moves later 23.Rd1 let a +6.44 position slip in the same way. She drew. In round 4, against Song Yuxin of China, she reached a knight and pawn against queen ending that the Lichess tablebase calls a draw, with three saving moves — 66.Nf5, 66.Ng8 and 66.Kg6. She played 66.Ng4, which the same tablebase gives as a loss, and lost. Today, in round 6, she was level against Varvara Poliakova of Belarus and played 57.Re8, taking the evaluation to −6.26; she lost again, in a match that ended 2–2.

On the New Zealand side, the most costly Olympiad moment was Renjith's. In round 2, against Elham Amar, he was +2.97 when he played 48.Kd3; the evaluation fell to +0.05, and the game was drawn — the only half-point New Zealand took from Norway in a 0.5–3.5 defeat. In round 1, Daniel Gong let two positions of about +5 go against Jerome Mboula Mboumba of Gabon, with 39...Rg3 and 41...Rxg5, and drew. Tom Middelburg's worst game was at the 2026 Bob Wade Masters, against Li Tian Yeoh: he was +5.87, played 67...Nc7 and dropped to −0.41, and over the next three moves the position twice reached a forced mate against him that his opponent did not see. The game was drawn.

The Hungarian with the longest list outside the Olympiad is Nomin-Erdene. In the Turkish league in August, against Hatice Asli Mustu, she let advantages of +7.03, +9.93 and +6.44 fall below one pawn in the same game and still won it, in a game that went past move 140. In the same league, against Mai Narva, she was level and played 35...h6: the evaluation went to −6.48 and she lost. Zsoka Gaal, by contrast, has only one move in her profile that took a level position to a losing one: 31...Bc3 against Michail Karamichos at Paleochora, from +0.07 to −3.67, in a game she still drew.

The blunder count stays out of the table: the classification uses the uncapped loss, so a player who is already winning and picks a slower mate adds a blunder without having changed anything. Mao's and Xie's profiles are full of exactly that. 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.

Daniel Gong is the one online sample worth reading. On the itheonlyme account, with an FM title verified by the platform, he played 300 games in 90 days and scored 48.7% with White and 50.7% with Black. With White he plays the London: after 1.d4 Nf6 2.Bf4 g6 he scored 34.8% in 23 games, and after 2...c5, 30.8% in 13 — both flagged with high confidence. With Black, 1.d4 Nf6 2.c4 g6 gave him 30.0% in 20, also high confidence. Over the board he opens 1.d4 in only 9 of 22 games, so the London is a hint of what he knows rather than of what he will play.

Sravan Renjith has 54 games on the slipperyspeedster account, with 38.9% as White and 53.7% as Black. The account lists Australia, against his New Zealand FIDE registration; it was included because the Chess.com player page, which links the athlete to the account, points to it. Szidonia Lazarne Vajda has 11 games and Petra Papp 4 — too few for any line.

Two Hungarians have a verified account and no games in the last 90 days: Zsoka Gaal and Davaademberel Nomin-Erdene. The account exists and is idle, which is different from the account not existing. Four players have no account that could be verified — Felix Xie, Tom Middelburg, Daqi Mao and Kata Karacsonyi — and stay out of this section. A matching username is not proof, and publishing a stranger's repertoire under a player's name is worse than publishing nothing.

What to watch in round 7

In the open, Canada are the higher-rated team, and the dossier offers two clear targets with Black players who never vary against 1.e4. Middelburg plays the Caro-Kann every time and scores 36.4% with it; Xie plays the Sicilian every time and drops to 43.3% in the 2...e6 lines. Against Mao, the table is the wrong guide: his 81.2% was earned against a field 300 points weaker than an Olympiad board, and his error profile is that of a player who wins won positions slowly, not one who never errs.

In the women's section Canada are the lower-rated team but lead on game points. Karacsonyi's 25.0% in 10 games against 1.Nf3 and Nomin-Erdene's 35.7% in 7 in the Queen's Gambit Declined are the two lines with a sample behind them. Papp without a win against 1.d4 in 9 Black games is a third hint. Lazarne Vajda is the unknown: her three-year record says she almost never loses, and her last three games in Samarkand say the opposite.

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. Two matches level on match points, and in both of them the most impressive number in the table is the one that deserves the least trust.

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 733 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
• Szidonia Lazarne Vajda v Song Yuxin, round 4 — broadcast game — Lichess
• Daniel Gong (itheonlyme) — 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