Lately, I’ve been playing blitz games against MadChess 3.3 (not yet released) to get a feel for how strongly it plays when weakened to my patzer Elo level. Sparring against MadChess helps me calibrate its limit-strength parameters and perhaps glean ideas for classifying more unreasonable inferior moves.
My most recent game begins 1.e4 c5 2.Nf3 d6 3.d4 Nf6 4.Nc3 Nc6 5.Bb5 a5 6.Be3 d5 7.e5 Ng4 8.Qd2 Nxe3 9.Qxe3 cxd4 10.Nxd4 Qd7 11.O-O f5 12.e6 Qd6 13.Nxf5 Qxe6 14.Qxe6 Bxe6 15.Nd4 Bd7 16.Rad1 g6 17.a4 Nb4 18.Rfe1 h5 19.Rd2 b6.
So far I’ve played well enough for a player of my strength in the limited time available in 3m+2s blitz. I’ve developed my pieces, protected my king, and weakened Black’s pawn structure near its king. Material is even. I’ve missed opportunities to advance or attack with my central pawns, but haven’t outright blundered. I spot Black’s weak c7 square and realize my knight on d4 can reach it via e6 because Black’s light squared bishop is pinned to its king by my bishop on b5 and cannot capture on e6. If I can get my knight to c7 it will fork Black’s king and rook on a8.
I play 20.Ne6 and MadChess replies 20…Bxb5. Black trading bishops improves my position as it enables me to place a knight on the b5 outpost, controlling Black territory without fear of harassment from any Black pawns. Black should have developed the dark squared bishop via 20…Bh6 to harass my rook on d2.
I should have played 21.Nxb5 to capture Black’s light squared bishop and place a knight on an outpost. But I was fixated on the fork on c7. Play continues 21.Nc7+ Kf7 22.Nxa8 Bh6.
Damn. By moving its dark squared bishop to a square where it attacks my rook on d2, MadChess also reveals an attack on my adventurous knight on a8. Now I’m under double attack. What to do? I can’t capture either of Black’s attacking pieces, so I should retreat the more valuable of my attacked pieces.
23.Rde2?? Doubling rooks on the e file, but blundering. I move my rook out of the crosshairs of Black’s dark squared bishop but into the crosshairs of the light squared bishop. Doh! I need to improve my “whole board” vision. I was focused on the lower half and upper right portion of the board and failed to see a ranged attack originating from the upper left portion of the board.
While reviewing the game after its conclusion, but before peeking at any engine suggestions, I thought I should have simply retreated my threatened rook via 23.Rdd1 Nxc2 24.axb5 Nxe1 25.Nxb6, which creates a passed pawn for White. However, I have a defensive resource in this position. Can you find it?
I missed this blocking tactic, thinking it was nothing more than a delaying move that loses a pawn. However, we see (in the puzzle above and annotated game below) that Black cannot capture the pawn on f4.
My blunder cost me an exchange and a knight, a net loss of five pawns using traditional material values. I played on, as patzer chess players should, attempting to force MadChess to blunder a piece back to me.
23…Bxe2 24.Rxe2 Rxa8 25.f3 Kf8 26.h3 Bf4 27.Kf2 Ke8 28.g3 Bc1 29.b3 Rd8 30.Re6 Rc8 31.Ne2 Rxc2 32.Rxb6 g5 0-1
But I ran out of time. Overall, a competitive game where MadChess played convincingly like a 635 Elo blitz player. (I’m attempting to calibrate “the feel of” MadChess’ UCI_Elo
parameter to chess.com’s blitz Elo ratings. Other chess sites may differ by 200 Elo points or so. See Chess.com Versus Lichess Ratings.) The world-class Komodo Dragon chess engine believes MadChess played slightly worse than I. But as Grandmaster Savielly Tartakower said many years ago, “The winner of the game is the player who makes the next-to-last mistake.”
“The winner of the game is the player who makes the next-to-last mistake.” -Savielly Tartakower
I capitalized on MadChess’ 11… f5?? blunder with the correct 12.e6! and held the advantage (unevenly due to numerous inaccuracies) until I overlooked my 23.f4! defensive resource and this “last mistake” proved decisive.
I have annotated the game with a few variations and comments, which you may review below. In addition, I gave Komodo Dragon, a world-class chess engine, one minute and 60 CPU threads (I have a powerful home PC) to analyze the game and insert its evaluations and suggested variations.