The Battle of the Bots: Caruana vs Nakamura
Who would win in a 2021-championship-style match? Would Fabiano Caruana win or would Hikaru Nakamura? According to chessgames.com, Caruana has 8 wins to 7 against Hikaru, with 33 draws.
I put their chess.com bots against each other in a maximum of 14 games like the 2021 WCC, and we will see who would win.
Note: This is not intended to make a point about either of these players. Keep playing chess!
Conditions
These two computers play in a maximum of 14 games.
Wins score 1 point, draws score half each and losses score none.
Caruana starts as White and colors alternate each game.
First to 7.5 points wins!
Game 1: Caruana vs Nakamura, 0-1 (1-0 Nakamura)
chess.com states that Hikaru’s bot is rated 20 points less than Fabiano’s, but somehow Hikaru pulled off a win with the black pieces!
Everything was quite equal in the KID until Caruana’s bot played 29. Bb6?? allowing Rg4! attacking the queen. Later in the game Black’s Kh7? proved ineffective, but then White played 37. Qd8? and Bf8 defended the weak h-pawn. A long 95 move game resulted in Hikaru giving quite random checks in a KQ vs K endgame, but in the end he was able to deliver a checkmate with his queen and king.
Link to analysis: https://www.chess.com/analysis/game/pgn/2oeSXE7WZc?tab=review
Full game:
Hikaru leads by one point, and next game he has a White advantage. Can Caruana turn things around?
Game 2: Nakamura vs Caruana, 1-0 (2-0 Nakamura)
If you thought the last game was long, this 126 move game would only be 10 moves less than the longest WCC game, Carlsen vs Nepomniachtchi, Game 6, 2021. The game opened with yet another KID, but the majority of the game was spent with Hikaru grinding down on a two-pawns-ahead rook endgame. However, he successfully promoted his two passed pawns and checkmated with two major pieces.
Link to analysis: https://www.chess.com/analysis/game/pgn/5av4P19UMc?tab=review
Full game:
And Caruana is down two points.
Game 3: Caruana vs Nakamura, 1/2-1/2 (2.5-0.5 Nakamura)
A draw in yet another KID keeps Nakmura’s lead at 2.
No big chances were present for both players; the chess.com accuracy is 97+ for both bots. The game ended with a repetition.
Link to analysis: https://www.chess.com/analysis/game/pgn/5ZofrZQB82?tab=review
Will we see another KID next round?
Game 4: Nakamura vs Caruana, 0-1 (2.5-1.5 Nakamura)
Answer: NO.
Caruana scored his first victory off from KID ground, going all out in the endgame before successfully pushing a pawn.
Instead, the players set out for a Semi-Slav, and the game was quite equal…until Nakamura’s blunder on move 90. The reasoning behind these long games is that bots can’t draw (at least the bots on chess.com) so then games that are supposed to be drawn are ridiculously long.
Caruana then promoted a pawn in a knight versus bishop and two pawns endgame; Hikaru could’ve drawn immediately on move 77 with Nxf5!! knowing that the resulting endgame is drawn.
Link to analysis: https://www.chess.com/analysis/game/pgn/2Ae2sDTn7x?tab=review
Will game 5 jump off 1. d4 theory?
Game 5: Caruana vs Nakamura, 1/2-1/2 (3-2 Nakamura)
Answer: YES.
A Four Knights Game saw an equalizing with trades and trades and an eventual draw in a 1v2 and then 1v1 rook and pawn endgame.
Link to analysis: https://www.chess.com/analysis/game/pgn/AhfsDS5k2?tab=review
Game 6: Nakamura vs Caruana, 1/2-1/2 (3.5-2.5 Nakamura)
Welcome back, 1. d4!
This game featured the Bogo-Indian Defense, where on the fourth move instead of going for a QID with 4…b6, White checks with Bb4+ immediately.
The game quickly died down with a trade of minor pieces, pawns, and eventually queens, and a perpetual check ended the game early. (Thank you!)
Link to analysis: https://www.chess.com/analysis/game/pgn/2f7NGJ85wc?tab=review
Game 7: Caruana vs Nakamura, 0-1 (4.5-2.5 Nakamura)
Nakamura wins another game, this time in an endgame with a rook pawn and bishop of opposite color, but successfully wins by preventing access to h1.
In the opening, Caruana…gambits? blunders? a pawn.
If he gambited an unsound pawn, that doesn’t do well against computers with no worry.
If he blundered a pawn, that won’t fare well against a computer.
Link to analysis: https://www.chess.com/analysis/game/pgn/5emQmASsTL?tab=review
Game 8: Nakamura vs Caruana, 1-0 (5.5-2.5 Nakamura)
What a game! Nakamura’s sound play against the Old Indian Defense gave him the win in a game up a piece.
Hikaru started building up an advantage starting move 12, and eventually he was up a knight by using his advanced a- and b- pawns, eventually promoting his passed h-pawn in the endgame and checkmating.
Link to analysis: https://www.chess.com/analysis/game/pgn/32ULtGPWU6?tab=review
Game 9: Caruana vs Nakamura, 1/2-1/2 (6-3 Nakamura)
Game 9 featured Nakamura with a minor advantage but failing to convert, leading to a repetition draw with passed pawns on both sides of the board.
In a QGD Catalan, Nakamura was able to win two pawns with an advantage, but then trades occurred and he found his king in a trapped position, allowing for a draw by repetition Rg7-Rh7 and Kg8-Kh8. Nakamura is only one and a half points from winning.
Link to analysis: https://www.chess.com/analysis/game/pgn/4wtnzcLfkS?tab=review
Game 10: Nakamura vs Caruana, 0-1 (6-4 Nakamura)
Caruana won a pawn in the middlegame and used it well in the endgame, ending up in a knight endgame two pawns up.
At the end of the game, Caruana played a brilliant promotion and deflection tactic, allowing a checkmate with the knight.
Link to analysis: https://www.chess.com/analysis/game/pgn/477UWmLD2v?tab=review
Game 11: Caruana vs Nakamura, 1/2-1/2 (6.5-4.5 Nakamura)
In game 11, the computers traded and traded, and what was left was an endgame where Nakamura was up a pawn, but was a drawn position. On move 50, the third move in the same position was played, and a draw by repetition was established. Hikaru must now win one game to reach 7.5.
Link to analysis: https://www.chess.com/analysis/game/pgn/4VVVcsQvjg?tab=review
Game 12: Nakamura vs Caruana, 1/2-1/2 (7-5 Nakamura)
Game 12 was a quite equal game; at one point Hikaru was up two pawns but failed to convert the advantage into a win in the endgame and eventually drew by repetition. Now Caruana must win twice to tie.
Link to analysis: https://www.chess.com/analysis/game/pgn/5W4pZ9B4ge?tab=review
Game 13: Caruana vs Nakamura, 1-0 (7-6 Nakamura)
A game in the French Defense led an advantage for Caruana.
A long endgame in an imbalanced position between a rook and rook pawn and a bishop. Eventually, Nakamura snapped with 97…Bh6?? 98. Rc7+! followed by 99. Kg6 won the bishop and the game.
Link to analysis: https://www.chess.com/analysis/game/pgn/4TYYakBUkE?tab=review
Caruana must win again in order to tie.
Game 14: Nakamura vs Caruana, 1-0 (8-6 Nakamura)
In a second Larsen Opening, Nakamura gained two pawns in the middlegame, and in the endgame he was up a lot of material and eventually checkmated with two queens, a rook, and a pawn for good measure. Nakamura won the match.
Link to analysis: https://www.chess.com/analysis/game/pgn/Kn7yui2z2?tab=review
Results
Apparently Hikaru won narrowly, but there are some things to consider.
1. Computers don’t have any pressure as they are bots.
2. Computers don’t have to win “on demand” as they have no idea what is going on here.
3. Computers must be easier to beat than the people they’re representing as then it would be nearly impossible for a non-GM to beat them.
Fabiano’s bot would never know that he had to win; the computers can’t play “risky” chess.
In other words, this trial has accomplished nothing, but had interesting results.
