USA's round 6 opponents at the Olympiad: how Norway and India play

Short answer

In round 6 of the 46th Chess Olympiad, on 21 September 2026, the United States face Norway in the open section on board 5 and India in the women's section on board 4. Drawing on 1,097 classical games broadcast by Lichess over the last three years, this article sets out each player's repertoire by colour, the lines where each of them scores below their own average, where the engine finds the mistakes, and what the online repertoire adds — always with the number of games beside every percentage, because that is what says how much of the dossier holds.

Event: 46th FIDE Chess Olympiad

How this was compiled: Compiled on 20 September 2026, with round 5 finished and round 6 not yet played: it starts on 21 September at 3pm Samarkand time (10:00 UTC). The pairings were checked on chess-results; the official PGN that names each team's line-up is only published when the round starts, so this dossier covers both full squads rather than four boards each. The repertoire uses the 1,097 over-the-board classical games from the last three years broadcast by Lichess and linked to each player by FIDE ID — 691 for the Norwegian squad and 406 for the Indian one. The mistake profile measures each player's 12 most recent games with Stockfish 19 at depth 16; every move quoted in this text was re-measured at depth 24, with a dedicated search and a cleared hash. The blitz and rapid repertoire comes from verified Chess.com accounts, over a 90-day window.

The essentials in a few lines

• In round 6, on 21 September 2026, the United States face Norway in the open section (board 5, both teams on 8 match points) and India in the women's section (board 4), where the Americans lead on match points, 9 to 8, against the top seed. Line-ups are only published when the round starts, so this dossier covers all ten players in the two opposing squads.
• Tor Fredrik Kaasen offers the most solid recommendation of this whole series: against 1.e4 he scores 50.0% across 38 games, high confidence, and the Spanish with 2...Nc6 takes him down to 35.0% in 10.
• India's three grandmasters post the lowest percentages in the squad and the cleanest endgames of the round — 10, 11 and 13 centipawns per move for Koneru, Vaishali and Deshmukh. Their sample is the Candidates, Norway Chess and the Cairns Cup; Vantika Agrawal and Savitha Shri B, whose percentages are higher, played the open circuit.
• The one large mistake the engine confirms in Humpy Koneru's sample cost her a game against Alice Lee — the American board two she may meet on Monday. At move 30 of their Cairns Cup game she was +3.3 and played 30.Qg4; the evaluation fell to −4.4 and she lost.
• Elham Amar's online account carries the only high-confidence weak line in the online section of this series: with Black, after 1.e4 e5 2.Nf3 Nc6, he scores 38.5% across 13 games in 90 days, against a 52.6% average with that colour.

The round 6 picture

Both American teams arrive from a round that did not go to plan. The open team lost 1.5–2.5 to Germany, the match this series previewed a day earlier, and now sits on 8 match points with 13.5 game points — level with Norway on match points and a point and a half behind on the first tiebreak. The women drew 2–2 with Uzbekistan and still lead their section on match points, 9 to India's 8, which is why the top seed is waiting for them on board 4 rather than board 1.

India lost 1.5–2.5 to Poland in round 5, the first defeat for the tournament's top seed, and that is the whole reason this pairing exists. Norway's route was steadier and had one stumble: a 1.5–2.5 loss to Singapore in round 3, between four comfortable wins. On seeding, the United States are the favourites in the open — top seed against the 21st — and the underdogs on paper in the women's, fifth seed against the first, while leading the section.

The sample column in the table below says how much of each dossier holds, and this is the strongest set of numbers the series has had. Every one of the ten players has at least 69 classical games broadcast in three years, none is flagged as low confidence, and Aksel Bu Kvaloy's 162 games are the largest sample of any opponent covered so far. That is 1,097 games in total, 691 Norwegian and 406 Indian.

One warning belongs here rather than in a footnote, because it governs how the whole women's section of this article reads. Percentages compare a player with the field they played, and these five Indians did not play the same field. Humpy Koneru's 72 games are Norway Chess, the Cairns Cup and the Women's Grand Prix — closed elite events, all of them. Divya Deshmukh's 69 are the Candidates, Norway Chess, the Cairns Cup and the open Grand Swiss. Vantika Agrawal and Savitha Shri B played Asian championships and the European open circuit. Comparing the five by percentage compares calendars. The comparison that works is each player against herself.

Norway: the fifth board is the one the numbers can reach

This is the most uniform sample of the round, and that is what makes it usable. All five Norwegians play the same kind of calendar — Nordic opens, the Norwegian elite championship, the Stockholm and Fagernes GM tournaments, team leagues — so comparing them with each other is not comparing calendars. When one of them scores ten points below another, the difference means something.

Tor Fredrik Kaasen is the opponent the dossier describes best and the one it can attack. Across 157 games he makes 62.7% with White and 55.1% with Black, the lowest figure of the squad, and the gap has a location. With Black against 1.e4 he scores 50.0% over 38 games — the largest sample behind any single recommendation this series has published — and inside that, the Spanish with 2...Nc6 brings him to 35.0% in 10 games, twenty points under his own average. Two separate measurements pointing the same way, both with real samples.

What makes that number interesting is that Kaasen is not a player who blunders. His mistake profile is among the cleanest of the ten: 15 centipawns per move in the middlegame over 303 moves and 13 in the endgame. He is not losing those games, he is drawing them — 41 draws in 79 games with White, 36 in 78 with Black. A team that needs a win on board five has a target; a team that needs a draw does not.

Johan-Sebastian Christiansen, at 2656 the strongest of the squad, posts 86.7% with White across 49 games: 38 wins, 9 draws, 2 losses. With Black he makes 72.2% in 54. There is no line in the dossier where he drops meaningfully below himself with a sample worth quoting — the two weak lines registered have five games each. Against him the honest reading is that the repertoire route is closed, and what remains is the clock and the position.

Aksel Bu Kvaloy inverts the usual shape: 68.6% with Black in 78 games against 63.7% with White in 84, with 43 draws in those White games. He is the most drawing player of the round and the one with the most weak lines registered — five with each colour — which is a function of the sample size rather than of frailty: with 162 games, the dossier has enough repetitions to find four-game clusters. The one with a usable sample is his 57.1% in 7 against the Spanish with 3...a6, and against him as White the dossier points at 1.Nf3, where he scores 55.0% in 10.

Frode Urkedal makes 75.4% with White and 67.5% with Black, and has the clearest structural weakness of the four grandmasters: against the Catalan-style 1.d4 Nf6 2.c4 e6 3.g3 he scores 42.9% in 7 games, twenty-five points below his own average with Black, and the Bogo-Indian with the retreat variation gave him 30.0% in 5. Two pointers in the same family — quiet queen's-pawn set-ups with g3 — and one of them with medium confidence.

Elham Amar, with 137 games, sits between the two blocks: 73.6% and 62.7%, no weak line with a sample above five, and the widest tournament calendar of the squad. His interesting numbers are online, and they are in their own section below.

India: five players, five calendars, one comparison that works

Start with the trap, because every number in this section sits on top of it. Humpy Koneru's 72 games in this sample come from seven events, and six of them are closed elite tournaments: Norway Chess in 2025 and 2026, the Cairns Cup in 2025 and 2026, and two legs of the Women's Grand Prix. There is no open in her sample at all. Divya Deshmukh's 69 games are the 2026 Candidates, Norway Chess, the Cairns Cup, the open Grand Swiss and the World Cup. Vaishali's 77 mix the Candidates and Norway Chess with a handful of opens. Vantika Agrawal's 81 and Savitha Shri B's 107 are Asian championships, European opens and norm events.

So the fact that Savitha Shri B posts the highest White percentage of the squad, 68.5%, and Divya Deshmukh the lowest, 37.1%, says nothing about which of the two is harder to play. It says they spent the last three years in different rooms. This is the same failure the series avoided with Vincent Keymer in round 5, and it is worth naming again: a percentage is a statement about a player and a field, and the field is not printed in the number.

Compared with herself, Humpy Koneru has one clear line. With Black against the Spanish — 1.e4 e5 2.Nf3 Nc6 3.Bb5 — she scores 16.7% in 6 games, twenty-six points below her own average with that colour, medium confidence. Against her as White the dossier finds no first move that opens an edge; against her as Black she meets 1.d4 with 1...Nf6 in 18 of 26 games and scores 66.7% doing it. She is, by the numbers here, a much more dangerous opponent with White than with Black.

Rameshbabu Vaishali shows the same asymmetry — 52.5% with White, 43.2% with Black — and two pointers with medium confidence. With Black against the Italian with 3.Bc4 she scores 21.4% in 7 games; with White, after 1.e4 e5 2.Nf3 Nc6, she scores 38.9% in 9. Both numbers sit twenty points or so below her own averages, and both come from the open games she has to face at the top level.

Divya Deshmukh's 37.1% with White deserves one more sentence, because it is the most extreme figure in the table: 4 wins, 18 draws and 13 losses in 35 games. Her online account, over the same period, shows 62.5% with White across 32 blitz and rapid games. The repertoire did not change between the two; the opposition did. With Black over the board she makes 60.3%, and the one line the dossier flags with a sample is 1.e4 e5 2.Nf3 Nc6 as White, at 21.4% in 7.

Vantika Agrawal is where the dossier is most useful to an opponent, because her calendar and her sample agree. As White she scores 25.0% in 6 games against the Slav with 2...c6 and 33.3% in 6 against 1.d4 Nf6 2.c4 e6 3.Nc3, both medium confidence; as Black, 35.0% in 10 against the Indian with 2...e6. And the game plan is unusually direct: against her, 1.d4 has produced 38.5% for her across 13 games, high confidence, against a 51.4% average with Black.

Savitha Shri B has the largest Indian sample, 107 games, and the widest spread of results. Her weakest documented line with White is the French, 41.7% in 6, and with Black the Sicilian with 2...e6 followed by 3.d4, at 33.3% in 6. Against her, 1.e4 has held her to 46.4% across 28 games — high confidence, and the second-largest sample behind a recommendation in this article.

Where the engine finds the mistakes

The table gives average centipawn loss per move by phase, with the number of moves measured beside each figure. The second number matters as much as the first: Savitha Shri B's 48 rests on 87 endgame moves, while Humpy Koneru's 10 rests on 164, and they are not equally solid statements. The opening column runs from 4 to 9 across all ten players and describes preparation rather than play.

The women's column contains the round's most useful correction to the previous section. The three Indian grandmasters whose percentages look modest have the three cleanest endgames of the ten: Koneru at 10, Vaishali at 11, Deshmukh at 13, all over samples above 160 moves. Their percentages are low because of who they played; their technique, measured against the engine rather than against a field, is the best on the board. An American team reading only the percentage column would prepare for the wrong match.

The one exception on the Indian side is Savitha Shri B's endgame, 48 over 87 moves, and one of the moves behind it is worth reporting because of what re-measuring did to it. At depth 16 the pass flagged a loss of 8,027 centipawns at move 60 of a game in the Kavala open. At depth 24, the move she actually played is the engine's own first choice, and it leads to mate in 14 in her favour. It is not a mistake at all; it is the arithmetic of an evaluation crossing into mate.

Humpy Koneru's profile contains one real mistake, and it is the most expensive move confirmed anywhere in this series. In round 6 of the 2026 Cairns Cup, on 12 August, she had a winning position against Alice Lee — plus 3.3 at depth 24, with 30.Nxg7 available. She played 30.Qg4, the evaluation fell to minus 4.4, and she lost the game. Alice Lee is the United States' board two in Samarkand. Whether the two meet again on Monday depends on two line-ups that do not exist yet, but the file is there.

On the Norwegian side, Aksel Bu Kvaloy has the cleanest profile of the ten — 4, 15 and 17 — and Tor Fredrik Kaasen is close behind, with the lowest endgame figure of the squad at 13. Kaasen's worst moment is also a real mistake, and a smaller one on paper: in round 5 of the 2026 Landsturneringen he was plus 7.1 against Kjetil A. Lie and played 37.Rb3 instead of 37.Bb6+. The evaluation went to dead level and the game was drawn. Johan-Sebastian Christiansen has the highest Norwegian endgame figure, 30 over 211 moves.

The pattern is worth stating because it repeats every round of this series and runs against intuition: the worst moments carrying enormous losses at depth 16 are nearly always artefacts of already-decided positions, while the mistakes that actually changed a result show up with modest numbers — 582 and 669 centipawns, in the two cases above. That is why no worst-moment figure reaches this text before being re-measured at depth 24 with a dedicated search and a cleared hash.

Blunder counts are deliberately absent from the table. The classification uses uncapped loss, so a winning player who picks a slower mate collects a blunder for a move that changes nothing — exactly what happened to Savitha Shri B above. Average centipawn loss caps each evaluation at 1000 and does not suffer from that.

What the online accounts show, and what they do not

This section answers a different question from the one above, which is why it is kept separate. A blitz repertoire is usually wider than a classical one, and what somebody plays with three minutes is not what they prepare for a five-hour game. It tells you what they know, not what they will play tomorrow.

Elham Amar's account, twitch_elhamblitz05, has the 300-game cap of the window and the most pointed result in it. With Black, after 1.e4 e5 2.Nf3 Nc6, he scores 38.5% across 13 games against a 52.6% average with that colour — the only online weak line the dossier has ever marked high confidence in this series. With White he is at 60.5% in 148 games, and the open games with 2.Nc3 and 3.d4 hold him to 50.0% in 11.

Aksel Bu Kvaloy's 121 games add a second concrete pointer: against the Petroff with 3.Nxe5 he scores 25.0% in 10 games, well below his 52.5% online average with Black. Tor Fredrik Kaasen's 67 games show the same shape as his classical dossier, 58.8% with White and 48.5% with Black. Johan-Sebastian Christiansen and Frode Urkedal have verified accounts with no games in the window — the account exists and is idle, which is different from the account not existing.

On the Indian side, Vaishali and Savitha Shri B both hit the 300-game cap, at 47.3% and 49.7% with White; Humpy Koneru has 44 games and posts 42.9% with Black online, almost exactly the 42.6% she posts over the board. Vantika Agrawal has 11 games, which is barely a sample and is reported as such.

Divya Deshmukh's online figures are the ones that matter most, because they close the argument of the second section. Over 64 games in 90 days she scores 62.5% with White and 64.1% with Black. Over the board, in classical chess, the same colour gives her 37.1%. No repertoire changed between those two numbers. The field did.

Every account used here was verified one by one against the Chess.com API before publication: verified title, real name and country matching the FIDE record. None of the ten needed a weaker criterion, and no account was included on the strength of a matching username alone.

What to watch in round 6

The open match has a clear shape. Against Christiansen, Urkedal and Amar the repertoire route is mostly closed: the samples are large, the percentages are high and the weak lines that exist have five or seven games behind them. Against Kvaloy the dossier suggests 1.Nf3 and notes that he draws more than anybody else here. Against Kaasen it offers something rarer — a recommendation with 38 games behind it, backed by a second measurement in the same direction.

The women's match is the one the tournament will be watching, and the preparation question is not which line to play but which reading of the numbers to trust. India's grandmasters look beatable in the percentage column and are not: their endgames are the cleanest measured this round, and their samples are the Candidates and Norway Chess. The two Indians the dossier can genuinely attack are the two IMs, and the sharpest single number is 1.d4 against Vantika Agrawal.

There is also a file that both teams already have. Alice Lee beat Humpy Koneru in the Cairns Cup six weeks ago, in a game Koneru was winning until move 30. Whether that pairing repeats depends on two line-ups nobody has seen yet.

Once the round starts and the official PGN carries the line-ups, the dossier file attached to this article switches to showing the four boards and each board's game, with no change to the text. Until then, what is here is the full squad — which is what there is to prepare on the eve of a round.

Frequently asked questions

Update log

• 2026-09-20 — First publication, with the round 6 pairings checked on chess-results and the round not yet played: repertoire from the 1,097 classical games broadcast by Lichess, mistake profile with Stockfish 19 at depth 16, and the blitz and rapid repertoire of the verified accounts — eight of the ten opponents had online games in the window, the best coverage of the series. The dossiers cover both full squads; once the line-ups are published, the same file will show the four boards and each board'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 6 pairings — Chess-Results
• 46th Chess Olympiad Samarkand 2026, Women — round 6 pairings — Chess-Results
• Humpy Koneru — broadcasts under FIDE ID 5008123 — Lichess
• Elham Amar (twitch_elhamblitz05) — 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
• An open games repertoire: meeting 1.e4 with 1...e5 without memorising Ruy Lopez theory