Tournament coverage

USA's round 3 opponents at the Chess Olympiad: how Denmark and Cuba play

Jonas Buhl Bjerre opened 1.c4 against Fabiano Caruana, a move that does not appear once in his 56 broadcast classical games with White. A scouting report shows habits. Olympiad preparation is built to break them.

8 min read · Published on 2026-09-18 · Updated on 2026-09-29

Short answer

In round 3 of the 46th Chess Olympiad, on 18 September 2026, the United States face Denmark in the open section and Cuba in the women's section. Using the classical games broadcast on Lichess over the last three years, this report shows each opponent's repertoire, the lines where each scores below their own average and where the engine finds the most mistakes, always with the sample size alongside.

Event: 46th FIDE Chess Olympiad

How this was compiled: Checked on 18 September 2026 with round 3 already under way: it started at 10:15 UTC and data collection finished at 13:30 UTC. Pairings checked on Chess-Results; line-ups, ratings and titles from the official PGN broadcast on Lichess. The repertoire uses over-the-board classical games from the last three years broadcast on Lichess, matched to each player by FIDE ID, excluding round 3 onwards of this Olympiad: 112, 53, 134 and 51 games for the four Danish players and 48, 48, 62 and 12 for the four Cuban players. Error profiles measured with Stockfish 19 at depth 16 on each player's 12 most recent classical games. Online accounts are not included in this version. Final round 3 scores checked against the official PGN at 15:53 UTC.

The essentials in a few lines

  • In round 3 of the Samarkand Olympiad, on 18 September 2026, the United States face Denmark on match 1 of the open section and Cuba on match 5 of the women's section.
  • Round 3 result: the US beat Denmark 2.5-1.5 in the open section and Cuba 3.5-0.5 in the women's section.
  • The US team is higher rated on every board: by 174 to 215 points in the open section and by 98 to 211 in the women's section.
  • Denmark's Mads Andersen scored 70% with White overall but 50% in 6 games after 1.d4 Nf6 2.c4 e6 3.Nf3, the line Hans Niemann chose against him in round 3.
  • Cuba's Yerisbel Miranda Llanes scored 74% with White overall but 50% in 8 games of the Queen's Gambit Declined, which Alice Lee entered in round 3.
  • All figures are from the opponent's point of view and show tendencies: samples range from 12 to 134 games, and error profiles use 12 games per player.

Round 3 at a glance

Who do the United States play in round 3 of the 2026 Chess Olympiad?

In the open section the US play Denmark on match 1, with Fabiano Caruana, Wesley So, Hans Niemann and Awonder Liang; Levon Aronian rests. In the women's section they play Cuba on match 5, with Carissa Yip, Alice Lee, Tatev Abrahamyan and Rose Atwell; Irina Krush rests.

The US team is favoured on all eight boards. In the open section the gaps run from 174 points on board 1 to 215 on board 4. In the women's section the narrowest is board 3, where Tatev Abrahamyan is 98 points above Yaniela Forgas Moreno, and the widest is board 1, where Carissa Yip is 211 above Maritza Arribas Robaina.

The Danish team has large samples, from 51 to 134 games per player. Cuba's board 4, Ineymig Hernández Gil, has only 12, so little is said about her here.

Round 3 line-ups, 18 September 2026 (official PGN). Sample is the opponent's number of over-the-board classical games in the last three years.
Section and boardUnited StatesOpponentSample
Open 1GM Fabiano Caruana (2789), BlackGM Jonas Buhl Bjerre (2615)112
Open 2GM Wesley So (2774), WhiteGM Jesper Søndergaard Thybo (2579)53
Open 3GM Hans Niemann (2725), BlackGM Mads Andersen (2549)134
Open 4GM Awonder Liang (2693), WhiteGM Bjørn Møller Ochsner (2478)51
Women 1IM Carissa Yip (2468), WhiteWGM Maritza Arribas Robaina (2257)48
Women 2IM Alice Lee (2453), BlackWGM Yerisbel Miranda Llanes (2251)48
Women 3IM Tatev Abrahamyan (2347), WhiteWGM Yaniela Forgas Moreno (2249)62
Women 4IM Rose Atwell (2342), BlackWIM Ineymig Hernández Gil (2157)12

Denmark: a 1.e4 player who opened 1.c4

How do the USA's open-section opponents play in round 3?

Jonas Buhl Bjerre opened 1.e4 in 48 of his 56 games with White and never 1.c4, yet played 1.c4 against Caruana. Mads Andersen scored 70% with White but 50% in 6 games after 1.d4 Nf6 2.c4 e6 3.Nf3. Bjørn Møller Ochsner scored 27% against 1.e4 in 11 games with Black. Jesper Søndergaard Thybo scored 73% against 1.e4 in 13.

Jonas Buhl Bjerre, board 1, had White against Caruana. In 56 games with White he opened 1.e4 in 48, 1.Nf3 in 4 and 1.d4 in 3. There is not a single 1.c4 in the sample, and that is what he played in round 3. It is the clearest case in this round of what a scouting report cannot do: it describes what a player has done, and a player facing Caruana on board 1 of the Olympiad prepares precisely not to repeat it.

Jesper Søndergaard Thybo, board 2, scored 61% with Black in 27 games, and he is at his most comfortable against 1.e4: 13 games, 8 wins, 3 draws, 2 losses, almost always with 1...e5. The move he scored worst against was 1.Nf3, 40% in 5 games. Wesley So opened 1.e4 anyway, reached a Giuoco Piano, and won.

Mads Andersen, board 3, scored 70% in 67 games with White. The exception is 1.d4 Nf6 2.c4 e6 3.Nf3, with 1 win, 4 draws and 1 loss in 6 games, 50%. Hans Niemann went into exactly that line, with 3...d5 4.g3 Bb4+, and won. With Black, Andersen's weakest line is 1.e4 e5 2.Nf3 Nc6: 30% in 10 games, with no wins.

Bjørn Møller Ochsner, board 4, scored 75% with White and 48% with Black. Against 1.e4 he has 11 games with Black and 27%: 2 wins, 2 draws, 7 losses. Awonder Liang opened 1.c4, a move Ochsner has faced three times in the sample.

Filip Boe Olsen, who played rounds 1 and 2, rests.

Cuba: strong with White, softer against 1.e4

How do the USA's women's-section opponents play in round 3?

Maritza Arribas Robaina scored 33% against 1.e4 in 12 games with Black, and 25% in 6 games of the French with 2.d4. Yerisbel Miranda Llanes opened 1.d4 in 24 of her 25 games with White and scored 50% in 8 games of the Queen's Gambit Declined. Yaniela Forgas Moreno answers 1.e4 with 1...c5 in 14 of 16 games. Ineymig Hernández Gil has only 12 games.

Maritza Arribas Robaina, board 1, is steady with White, 70% in 23 games, and far weaker against 1.e4 with Black: 33% in 12 games, with 1 win, 6 draws and 5 losses. Her French with 2.d4 d5 went 0 wins, 3 draws and 3 losses. Carissa Yip opened 1.e4, got the French, sidestepped 2.d4 with 2.d3, and won.

Yerisbel Miranda Llanes, board 2, opened 1.d4 in 24 of 25 games with White and scored 74% overall. In the Queen's Gambit Declined, 1.d4 d5 2.c4 e6, she scored 50% in 8 games. Alice Lee entered it, with 3.Nc3 Bb4.

Yaniela Forgas Moreno, board 3, scored 66% with Black in 29 games. Against 1.e4 she played the Sicilian in 14 of 16 games. Against Tatev Abrahamyan she chose 1...c6, a reply that appears at most twice in the sample. With White she is a Bird player: 1.f4 in 17 of her 33 games.

Ineymig Hernández Gil, board 4, has 12 classical games in the sample, eight of them with White. It is too little for a weak line or a game plan. Roxangel Obregón García, who played round 1, rests.

Where the engine finds the mistakes

In which phase of the game do the USA's opponents make most mistakes?

Across each player's 12 most recent classical games, measured with Stockfish 19 at depth 16, Bjørn Møller Ochsner's average loss per move rises from 27 in the middlegame to 45 in the endgame, over 243 endgame moves. Cuba's top two boards, Maritza Arribas Robaina and Yerisbel Miranda Llanes, also lose more per move in the endgame than in the middlegame.

The table shows average centipawn loss per move, ACPL, in each phase. In the opening everyone sits between 4 and 9, which says more about preparation than about play. Ochsner's endgame figure rests on 243 moves, the largest endgame sample in the table. Yaniela Forgas Moreno's 22 in the endgame rests on only 40 moves, and Ineymig Hernández Gil's 58 comes from the only 12 games she has in the sample.

With 12 games per player, ACPL compares one player's phases with each other; it does not rank players, and a five-point gap between two names proves nothing.

Blunder counts are left out. When a player who is already winning picks a slower mate, the blunder threshold records a huge loss for a move that changes nothing. ACPL caps every evaluation at a thousand centipawns and does not suffer from this.

ACPL by phase in the 12 most recent classical games (Stockfish 19, depth 16). For Ineymig Hernández Gil these are all 12 games in her sample.
PlayerOpeningMiddlegameEndgameOverall
Jonas Buhl Bjerre4242219
Jesper Søndergaard Thybo5111511
Mads Andersen4222018
Bjørn Møller Ochsner5274530
Maritza Arribas Robaina8202820
Yerisbel Miranda Llanes6222921
Yaniela Forgas Moreno8312224
Ineymig Hernández Gil9325837

What round 3 showed

Did the US players go into their opponents' weak lines in round 3?

On two boards, yes, and in different ways. Hans Niemann entered Mads Andersen's weakest line with White and won; Alice Lee entered Yerisbel Miranda Llanes's Queen's Gambit Declined. On board 1 the sample had nothing to offer: Jonas Buhl Bjerre opened 1.c4 for the first time in it.

The US won both matches. In the open section they beat Denmark 2.5-1.5: Wesley So and Hans Niemann won, Fabiano Caruana drew with Jonas Buhl Bjerre, and Awonder Liang lost to Bjørn Møller Ochsner after opening 1.c4. In the women's section they beat Cuba 3.5-0.5: Carissa Yip, Tatev Abrahamyan and Alice Lee won, Lee in Yerisbel Miranda Llanes's Queen's Gambit Declined, and Rose Atwell drew.

The two cases pull in opposite directions, and both are worth keeping. Niemann's win came in Andersen's weak line, but one game does not confirm a 6-game statistic any more than Bjerre's 1.c4 refutes a 56-game one. What the report measures is where a player has scored less. What the player does on the day depends on who is sitting opposite.

Frequently asked questions

Where does the data on the USA's opponents come from?

From over-the-board classical games broadcast on Lichess over the last three years, matched to each player by FIDE ID. The repertoire by colour, the lines where the opponent scores below their own average and, on the 12 most recent games, an error profile measured with Stockfish 19 at depth 16 are all built from them.

Are the percentages the USA's or the opponent's?

The opponent's. When the report says Mads Andersen scored 50% in 6 games in a line, those are his points in those games. A low percentage is good news for whoever faces him.

Why are online blitz and rapid games not included?

Only online accounts proven to belong to the player are used, and that proof was not ready in time for this version. When the online repertoire is added it will sit in its own section, separate from the classical games.

Why did the report not predict Bjerre's 1.c4?

Because it describes past games, and there is no 1.c4 among Bjerre's 56 broadcast classical games with White. Preparation for a specific opponent at an Olympiad often aims at exactly that: leaving the lines the other side has studied.

Update log

  • 2026-09-18 — First published with round 3 in progress: repertoire from the classical games broadcast on Lichess and error profiles with Stockfish 19 at depth 16, without online accounts.
  • 2026-09-18 — Republished with the final round 3 scores and an interactive report for each opponent in the line-up, comparing the round 3 game with the known repertoire.
  • 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

Related guides

Your next opponent also has a line where they score less

Personal Chess Coach's opponent report reads the public games of whoever you are about to face and shows their repertoire by colour, the lines where they score below their own average and how many games sit behind each figure.

Create a free account