I just recorded my highest rated ever chess win by far. Here’s how the game played out.
The position transposed into the Ruy Lopez and as it turns out I played the old Steinitz defense (I kind of just winged it).
Up to move 12 the position was approximately equal. I made a sizeable error by elevating my rook to b7 instead of castling. Castling was the more sensible option as my king was a bit exposed but I was eyeing the open file to white’s king.
It turns out that the evaluation depended on a very precise sequence of moves, and so when white developed rc1 to e1 on move 16, which looks very natural, the position was back to equal. Then I finally castled :/
“You’ve broken a pin and allowed your bishop to provide defense. There was only one move that worked and you played it! This was a game-changing move! You no longer have a losing position.” -chess.com analysis engine
The next move (18) I made I hung my bishop by moving my queen to attack. I got my move order confused (I think I was planning to trade pawns then dark squared bishops) but luckily my opponent did not capture the bishop and instead blocked my queen. The game was then equal again.
Move 18 I attacked the knight, and white blundered by allowing me to trade my weak light squared bishop. The recapture also deflects his queen from the d4 square and I seize my chance. This run of accurate play by myself paves the way…
Taking the victory royale by delivering a checkmate
Commentary resumes at move 24: I forced white’s king on a run after a strong sequence of moves. The position was once again equal after I missed a rook lift. Eventually I shepherded white’s king to the edge of the board. It turns out after a computer analysis, the position was still OK for white at this point. However, white missed the one move to stay alive. King to h5. Instead king h3 was played and with less than 30 seconds left on the clock, I slid my rook across to deliver mate.
Post match thoughts:
I’m happy with the result as players rated around 2554 are usually FM, IM or GM titled players. Beating a player 1000 points higher is statistically extremely unlikely (see below table). Unfortunately the game was unrated so although the game is recorded on both profiles, it didn’t give me any elo or show up as my best win. It does look like the opponent plays most of their games unrated and wins a very high proportion of them. It wasn’t a perfect game but I made the game difficult for my opponent.
This is a probability table that shows the predicted outcomes based on our ratings.
|Orangestem win (Black, 1542)
|player 2 win (White, 2554)
I beat a grandmaster once before, but it was blitz horde chess which probably doesn’t count. That one was tricky to convert though.