Brazil's round 8 opponents at the Olympiad: how Japan and Chile play

Short answer

In round 8 of the 46th Chess Olympiad, on 24 September 2026, Brazil meet Japan in the open section, on table 19, and Chile in the women's section, on table 28. Drawing on 595 classical over-the-board games broadcast on Lichess over the past three years, this dossier analyses the opening repertoire of each squad, the specific variations in which opponents score below their historical averages, the phases in which the engine identifies significant inaccuracies, and the supplementary insights provided by online games — accompanied throughout by sample sizes to ensure statistical rigor.

Event: 46th FIDE Chess Olympiad

How this was compiled: Researched on 23 September 2026, with round 7 finished and round 8 still to be played: it starts on 24 September at 3 pm Samarkand time (10:00 UTC). 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 595 classical over-the-board games from the last three years broadcast on Lichess, linked to each player by FIDE ID — 216 from the Japanese squad and 379 from the Chilean. The error profile measures each player's 12 most recent games with Stockfish 19 at depth 16; Kohei Yamada has 9 games in his whole sample, and for Francisca Sanchez Tobar and Rocio Contreras Suazo it was 11, because one game each 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; correspondence games are left out.

The essentials in a few lines

• In round 8, on 24 September 2026, Brazil face Japan in the open section (table 19) and Chile in the women's section (table 28). In the open, both teams enter with 10 match points; in the women's, 8. Brazil hold the rating advantage in both encounters. Because team line-ups are confirmed only in the official PGN at the start of the round, this dossier provides comprehensive coverage of all ten squad members.
• Shou Otsuka registers 57.4% with White across 61 games, but falls to 25.0% with Black in 52 encounters against opposition rated on average approximately 50 points higher than him. Facing 1.e4, he scores 23.2% in 28 games; against 1.d4, he achieves 23.5% in 17.
• Ryosuke Nanjo was defeated in all 3 classical games in the sample when facing the Sicilian Defence with 1.e4 c5 2.Nf3 d6, a variation in which he records 30.8% across 13 blitz and rapid games on Chess.com. With White, he manages a modest 39.5% in 19 classical games, despite facing lower-rated opposition on average.
• The five Chilean players opened with 1.e4 in 178 of their 185 games with White. Conducing the Black pieces, Francisca Sanchez Tobar scores 28.6% in 7 games against 1.d4, whilst Maria Belen Carvajal Mendoza, generally dependable against 1.d4, drops to 28.6% in 7 encounters in the Sicilian against 2.Nc3 e6.
• In round 4, against Italy, Rocio Contreras Suazo twice reached endgames evaluated as winning by the tablebase — each dependent on a solitary winning move (64.g4 and 65.Ke4) — and emerged with a draw. In round 7, against Croatia, Maria Belen Carvajal Mendoza played 22.Rf1 from a +6.13 evaluation; although the engine assessed the position as lost following the move, she ultimately secured victory.

The round 8 picture

Brazil enter the eighth round following a 2.5–1.5 victory over Tajikistan in the open section and a 1–3 defeat against Vietnam in the women's section. In the open standings, Brazil and Japan are level on 10 match points, with Japan maintaining a slight advantage in individual game points (17 to 15.5). In the women's competition, both teams have accumulated 8 match points, with Brazil ahead on tiebreak criteria by 18 game points to Chile's 15.5.

In contrast to round 7, both fixtures confer favourite status upon Brazil according to the initial tournament seeding. In the open section, Brazil are seeded 31st and meet the 76th seeds; in the women's event, Brazil occupy the 44th seed and face the 70th. Japan arrive with considerable momentum: after consecutive 0.5–3.5 defeats to Italy and Kazakhstan, they secured successive victories over Monaco, Saudi Arabia, and most recently Belgium (2.5–1.5), highlighted by Shou Otsuka's decisive win against Elias Ruzhansky (rated 2427). Chile, conversely, arrive from an opposing trajectory: after consecutive 2.5–1.5 victories over Thailand and Slovakia, they were defeated 1–3 by Croatia in round 7.

The sample sizes recorded in the table below establish the empirical basis for each tactical deduction, reflecting marked differences between the two squads. The Chilean dataset comprises 379 games, whereas the Japanese sample contains 216. Shou Otsuka accounts for the largest individual sample across both matches (113 games), owing to his extensive campaign in European open tournaments, with appearances in Budapest, Marienbad, Biel, and London. At the other extreme, Kohei Yamada has just 9 games in the database, all contested in Olympiad events (4 in Budapest 2024 and 5 in Samarkand). As he is the sole player below the 10-game threshold — below which the dossier assigns low statistical confidence —, the text refrains from recommending concrete opening preparations against him.

None of the ten players has featured in all seven rounds. In the seventh round, Japan fielded Tran, Otsuka, Kojima, and Yamada, resting Nanjo; Chile deployed Gomez, Sanchez, Contreras, and Carvajal, resting Morales. Squad rotation is standard practice across demanding Olympiad schedules, which underscores the necessity of evaluating all five players on each roster rather than relying on speculative predictions regarding four boards.

Japan: contrasting performances by colour and two opening approaches

The central reference point in preparing against Japan lies in the numbers of Shou Otsuka. Conducting the White pieces, he attains a solid 57.4% in 61 games; with Black, his score contracts sharply to 25.0% over 52 encounters, comprising 5 wins, 16 draws, and 31 defeats. This disparity cannot be explained by tournament schedule: his opponents with Black held an average rating of 2372, barely 50 points higher than his sample average of 2323 — a rating difference insufficient to account for a 25% performance. This structural vulnerability manifests against both primary opening moves: against 1.e4, Otsuka scores 23.2% in 28 games; against 1.d4, he records 23.5% in 17. Facing 1.e4, he achieved 22.7% in 11 games with 1...e5, 38.9% in 9 with the French Defence, and a meagre 7.1% in 7 with the Sicilian. In Samarkand, playing Black, he was defeated by Italian Grandmaster Francesco Sonis in round 2.

With the White pieces, Otsuka presents an entirely different proposition, displaying near-total commitment to 1.e4: he chose it in 51 of 61 games, scoring 57.8%. The line in which he encounters the greatest resistance is the classical main line: following 1.e4 e5 2.Nf3 Nc6, his score drops to 38.5% across 13 games, a variation classified as high confidence by the dossier. Within this sequence, 3.Bb5 (the Ruy Lopez) yielded only 10.0% in 5 games, whereas 3.Bc4 (the Italian Game) produced 56.2% in 8. Against the Sicilian Defence, he remains close to his overall average, scoring 52.3% in 22 games after 2.Nf3, but experiences greater difficulty against 2...Nc6 (44.4% in 9) and 2...e6 (41.7% in 6).

Tran Thanh Tu is Japan's highest-rated player at 2406, with 28 games in the sample, equally divided between 14 appearances with each colour. He represented Japan at the Budapest Olympiad in 2024 and qualified for the 2025 World Cup, where he was eliminated in the opening round by GM A.R. Saleh Salem of the United Arab Emirates, 2–0. With White, he scores a solid 60.7%, selecting 1.d4 in 9 of 14 games with a 72.2% return; in the 3 games where he chose 1.e4, he scored 33.3%. With Black, he maintains 57.1% overall and exhibits a corresponding divergence: 65.0% in 10 games against 1.d4, but 33.3% in 3 games against 1.e4, which he meets with the Sicilian Defence. Although three games do not form an exhaustive sample, the parallelism across both colours points to an unmistakable strategic preference.

Ryosuke Nanjo exhibits the inverse dynamic to Otsuka. With White, he manages only 39.5% in 19 games against opposition averaging 2264 — approximately 70 points below his sample average of 2334. He opened 1.e4 in 15 of those 19 games, returning 33.3%. Against the Sicilian with 2.Nf3, he scored 37.5% in 8 encounters, and when Black opted for 2...d6, he lost all 3 classical games in the sample. While three classical encounters might represent merely an isolated indicator, the exact pattern recurs in his verified Chess.com account, where he scores 30.8% in 13 games — as detailed in the online analysis. With Black, Nanjo performs substantially better, scoring 52.4% in 21 games, albeit against opposition averaging 2200 (134 points below his rating): against 1.d4 he attains 65.0% in 10 games, dropping to 38.9% in 9 against 1.e4.

Shinya Kojima is the most predictable player on the Japanese roster. With White, he opened with 1.d4 in all 12 games in the sample and scored 58.3% against opponents averaging 2158 (146 points below him) — a figure reflecting the level of competition rather than opening advantage. With Black, he met 1.e4 with the Caro-Kann in all 9 games, scoring 55.6%; after 1.e4 c6 2.d4 d5, his score declines to 42.9% in 7 games. His games tend towards brevity: the median duration is 71 plies, with only 26.9% progressing beyond move 40. Finally, fifth board Kohei Yamada scored 90.0% in 5 White games and 62.5% in 4 Black games, all contested at Olympiads. Nine games illustrate his tournament record rather than a deeply established repertoire.

Chile: five 1.e4 practitioners and a challenging first move

In their initial move with White, the Chilean squad displays remarkable uniformity. Gomez selected 1.e4 in all 54 of her White games, Sanchez in all 26, Morales in 30 of 31, Carvajal in 46 of 48, and Contreras in 22 of 26. Brazil therefore enjoy the strategic clarity of preparing their Black defences strictly against 1.e4 on all boards, leaving the choice of defensive system as the primary consideration. The empirical evidence provides clear guidance: against the classical reply 1...e5, three Chilean players perform significantly worse than against the Sicilian. Gomez records 55.3% in 19 games against 1...e5 compared to 66.7% in 24 against 1...c5; Sanchez scores 41.7% in 6 against 1...e5 versus 62.5% in 12 against the Sicilian; and Carvajal achieves 59.1% in 11 against 1...e5 compared to 73.5% in 17 against the Sicilian. Morales is the sole exception, scoring 80.0% in 5 games against 1...e5 and 50.0% in 13 against the Sicilian.

With the Black pieces, the clearest tactical opportunity lies against 1.d4. Facing the queen's pawn, Sanchez scores only 28.6% in 7 games, compared to 50.0% in 14 against 1.e4; after 1.d4 Nf6 2.c4 e6, her score drops to 10.0% in 5 games. Contreras scores 50.0% in 7 games against 1.d4 compared to 60.7% in 14 against 1.e4; Morales registers 50.0% in 14 against 1.d4 versus 55.0% in 20 against 1.e4. Gomez displays near-perfect parity between both openings (44.2% in 26 against 1.d4 versus 45.2% in 21 against 1.e4). Carvajal is the only player to favour 1.d4, posting 70.0% in 15 games against 1.d4 compared to 59.5% in 37 against 1.e4. Despite the modest sample sizes in certain lines, the consistency of these trends provides meaningful strategic guidance.

Javiera Gomez Barrera, Chile's top board at 2181, holds the largest sample on the team (108 games) and displays a marked performance asymmetry. With White she scores 63.9% in 54 games; with Black, she declines to 45.4% across the same number of games against opponents of comparable strength. Against 1.e4, she replies 1...e5 in 19 of 21 games, scoring 42.1%; following 2.Nf3 Nc6, her score settles at 38.9% in 18 games. Against 1.d4 d5 2.c4, she plays 2...e6 and achieves 40.0% in 10 games, falling to 20.0% in 5 games when White continues with 3.Nf3.

Monserrat Morales Flores and Francisca Sanchez Tobar post comparable figures with White (58.1% in 31 games and 61.5% in 26, respectively), but adopt divergent approaches with Black. Morales employs the Sicilian in 13 of her 20 games against 1.e4, scoring 42.3%; in the Kan Variation (1.e4 c5 2.Nf3 e6 3.d4), she achieves 42.9% in 7 games. By contrast, when playing 1...e5, her score rises to 78.6% in 7 games. Against 1.d4 she invariably selects 1...Nf6, scoring 71.4% in 7 games if White continues 2.c4, against 40.0% in 5 with 2.Nf3. Sanchez meets 1.e4 with 1...e5 in 8 of 14 games (37.5%), reserving the Caro-Kann Defence for 6 encounters in which she achieved an impressive 66.7%.

Rocio Contreras Suazo boasts the highest percentages on the roster — 75.0% in 26 games with White and 60.0% in 25 with Black —, figures that require contextualisation: her opponents when playing Black averaged a rating of 1786, approximately 90 points below her level, leaving no glaring weaknesses in her sample. Maria Belen Carvajal Mendoza, born in 2010, accounts for 105 games, nearly matching Gomez's volume. She records 70.8% with White and 62.3% with Black against opposition averaging around 1720 (roughly 100 points below her rating of 1824). She plays the Sicilian in 36 of 37 games against 1.e4, scoring 61.5% in 26 encounters after 2.Nf3. Against 2.Nc3, however, her score drops to 37.5% in 8 games, and falls further to 28.6% in 7 when the game continues 2.Nc3 e6.

Where the engine finds the errors

The accompanying table outlines the average evaluation loss per move across each phase of play, alongside the volume of moves analysed. In the opening, nine of the ten competitors remain within a controlled band of 5 to 10 centipawns; Contreras, averaging 15 across 86 moves, represents the sole outlier. The middlegame establishes a distinct dividing line between the squads: the Japanese players range from 21 (Yamada) to 37 (Tran, who features the largest sample in this phase with 311 moves), whereas the Chileans span from 22 (Gomez) to 50 (Carvajal) — with Gomez, the team's top board, standing out as the most precise Chilean player across all three phases.

The most instructive episode in the sample involves Rocio Contreras Suazo in round 4 against Italian representative Kamilla Rubinshtein (rated 2108). Playing White, Contreras held a +5.13 advantage when she played 48.Ra6, after which the evaluation dropped to −0.02. Two moves later, 50.h4 left the position at −6.23, as 50...Nxd5+ won decisively for Black; Rubinshtein chose 50...Rxf2+, allowing the contest to continue. On move 64, in an endgame confirmed as won for White by the Lichess tablebase, only 64.g4 preserved the win; Contreras played 64.Rh8, which allows a theoretical draw. Her opponent returned the opportunity with 64...d4, restoring White's winning position, once again conditioned upon a single precise move: 65.Ke4. Contreras played 65.Rd8, concluding in a draw on move 70.

Maria Belen Carvajal Mendoza experienced the sharpest turnaround of round 7 against Croatia's Lara Zagorac (rated 2178). Holding a +6.13 advantage on move 22, she played 22.Rf1, a move which dedicated re-measurement at depth 24 evaluates at −6.25: the response 22...e3 was immediately winning for Black. Zagorac opted for 22...g5, allowing Carvajal to regain the initiative and secure Chile's solitary point in the match. The squandering of large advantages represents a recurring tendency in her recent record: at the Lebu Open in July, 40.Rb8 from a +6.89 evaluation reduced her position to −0.03 against Juan Carlos Rugero Gomez (1646) in a game she ultimately lost, and 40.a7 at +6.67 levelled the evaluation to 0.00 against Pedro Antonio Bravo Ortiz (1667), resulting in a draw.

Among the Japanese squad, the most damaging inaccuracies occur during the middlegame, having directly influenced match results at this Olympiad. In round 5, against Riccardo Costalonga of Monaco (rated 2184), Shou Otsuka stood at +0.68 before playing 19.Rxd5, after which the evaluation plummeted to −5.91, leading to defeat. In round 2, against Francesco Sonis (rated 2539), he held an equal position before playing 52...Rf8 instead of 52...f3+, dropping to −8.26. Tran Thanh Tu, in round 6 against Ahmed Alrehaili of Saudi Arabia (2106), held +7.29 when he played 33.Rd5, which the engine evaluates at −0.13, though he subsequently prevailed. In the same round, Shinya Kojima held +6.40 against Yazeed Alshnaiber (1866) and played 43.Rc1, reducing his advantage to +2.00 before winning on move 75. At the 2026 Asian Championship, Ryosuke Nanjo drew with Zhang Di (2446) following 33.Rhf1, which shifted the evaluation from −0.26 to −5.66.

Elevated figures do not always reflect substantive strategic errors, as tablebase verification resolves two notable discrepancies. At depth 16, Kojima's 75.Rb7 against Alshnaiber appears as an apparent loss of more than 35 pawns, and Carvajal's 67.Kc6 against Zagorac is likewise flagged with a substantial loss. The tablebase confirms that both moves retain forced wins: shallow analysis misread positions that were already fundamentally decided. This represents a known pattern in automated analysis: in settled endgames, low-depth calculations frequently assign enormous penalties to moves that do not alter the theoretical result. By contrast, Contreras's endgame average of 60 centipawns across 108 moves reflects genuine instability, as demonstrated by the drawn result against Italy.

Blunder counts are intentionally excluded from the table for this reason: the classification applies uncapped evaluation changes, meaning a player choosing a longer path to an inevitable checkmate registers multiple blunders without deteriorating the position. The average loss per move, by capping individual evaluations at one thousand centipawns, eliminates this distortion.

What the online accounts show, and what they do not

This section addresses a distinct analytical objective from the preceding discussion and is therefore treated independently. Repertoires in rapid and blitz controls are naturally more fluid and experimental than in classical tournaments, as decisions executed under a three-minute clock differ fundamentally from deep preparation for multi-hour contests. These records serve to indicate a player's familiarity with specific tactical themes and identify areas of practical discomfort, rather than providing definitive forecasts for upcoming games.

Under the account rnanjo, verified with his International Master title on Chess.com, Ryosuke Nanjo contested 300 games within 90 days, scoring 54.0% with White and 54.7% with Black across exactly 150 games with each colour. His repertoire with White is broader than over the board: he chose 1.e4 in 81 games and 1.c4 in 39, contrasting with the near-exclusive use of 1.e4 (15 of 19) in his classical record. Against the Sicilian with 1.e4 c5 2.Nf3 d6, he scored 30.8% across 13 games — a benchmark classified as high confidence —, with a sharper decline in the Open Sicilian, where he managed just 10.0% in 5 games following 3.d4, matching his classical defeats. Against the Caro-Kann with 2.d4 d5, he scored 16.7% in 6 games.

With the Black pieces, Nanjo's profile records 42.9% in 7 games in the Open Sicilian with 1.e4 c5 2.Nf3 Nc6 3.d4. Owing to the limited sample size, this metric remains categorized as an initial indicator rather than an established trend.

Tran Thanh Tu and Javiera Gomez Barrera possess verified accounts, but their activity during the 90-day window was restricted to correspondence chess — a discipline with multi-day move horizons that falls outside the criteria of this study. Shou Otsuka played no games in the relevant period; Shinya Kojima logged 3 games, Kohei Yamada 4, and Monserrat Morales Flores just 1, none of which provide statistically meaningful samples. Francisca Sanchez Tobar, Rocio Contreras Suazo, and Maria Belen Carvajal Mendoza have no Chess.com accounts linked to their names, in accordance with the dossier's strict policy of excluding unverified profiles.

What to watch in round 8

Strategic preparation in the open section depends upon Japan's final team selection, with the dossier providing tactical recommendations for all five candidates. With White, facing Otsuka offers the most solid empirical foundation, backed by 52 games demonstrating vulnerabilities against primary opening moves; against Tran, 1.e4 stems from his modest score in 3 games and should be treated cautiously. Playing Black against Otsuka, the continuation 1...e5 followed by 2...Nc6 steers the struggle into the line where he manages only 38.5% in 13 games; against Nanjo, 1...c5 with 2...d6 replicates the structure in which he suffered classical defeats and scores only 30.8% online. Kojima's profile warrants prudence: his 58.3% score with White was constructed against opponents rated nearly 150 points lower, while his Caro-Kann drops to 42.9% in 7 games after 2.d4 d5.

In the women's fixture, Brazil hold the rating advantage and possess distinct guidelines for both colours. Conducting the Black pieces, the Chilean first move is virtually certain: 1.e4 featured in 178 of 185 instances. The classical reply 1...e5 emerges as a sound strategic choice, as Gomez, Sanchez, and Carvajal score lower against it than against the Sicilian. Playing White, introducing 1.d4 confronts Sanchez, Contreras, and Morales with lower performance figures than they achieve against 1.e4; against Carvajal, who defends well against the queen's pawn, the most effective route is the Sicilian against 2.Nc3, whereas against Gomez the variation 1.d4 d5 2.c4 e6 3.Nf3 stands out. These findings rest on samples ranging from five to nineteen games — sufficient to inform preparation without offering dogmatic guarantees.

Once the round commences and the official PGN confirms the board pairings, the accompanying dossier files will display the four active boards and their respective games in real time. Until then, the assessment presented here covers both squads in their entirety: ten players analysed, 595 games examined, and the distinctive profile of a Japanese master whose performance alters dramatically depending on the colour of the pieces.

Frequently asked questions

Update log

• 2026-09-23 — First published, with the round 8 pairings checked on chess-results and the round still to be played: repertoire from 595 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 8 pairings — Chess-Results
• 46th Chess Olympiad Samarkand 2026, Women — round 8 pairings — Chess-Results
• Maria Belen Carvajal Mendoza v Lara Zagorac, round 7 — broadcast game — Lichess
• Rocio Contreras Suazo v Kamilla Rubinshtein, round 4 — broadcast game — Lichess
• Ryosuke Nanjo (rnanjo) — 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