Since I've finished finals at UTD and gone back to Kansas now, I am finally going to write a blog about the chess.com deathmatch. I played it on April 28 against Eric Hansen, a fellow member of the UTD chess team. After a couple of weeks the games had become only a dim memory, but I was able to refresh it by looking on a page called Game Archive.
The match was divided into three segments. The first time control was 5 1. 5-minute is my favorite time control, so I hoped to do well in that one. But I didn't manage to get more than an even score. I kept losing what seemed like promising positions against the Gruenfeld. I had to make up for that by winning some games in the French. Looking back at the games, it seems like I had the most problems with time trouble in the 5 1, despite it being the longest time control.
In the next time control (3 1) I scored +2. The most memorable game was one where I had to defend rook vs. rook and bishop for 50 moves on the 1-second increment. I've always thought that this endgame looked really tough to defend, but it was not as difficult as I expected. I did get a lost position right away at the beginning though. At one point I could have been mated in 2 moves. But then I escaped from the corner with a stalemate trick and never had any more problems.
I was most worried about the final time control because I don't play 1-minute very often, while Eric does and is highly rated in it. Fortunately for me, the time control was not actually 1-minute; instead it was 1 1, which turned out to be much slower. Eric moved very quickly as for a 1 0 game, but was unable to flag me due to the pesky increment. This resulted in too many unnecessary mistakes for him, and I scored +3 in the final third of the match.
I am going to focus on another game from the 3 1:
Black lost a knight with 21...Nxe5 after which the game ended very quickly. But instead of taking e5, he could have taken on f3. White can then sacrifice a rook to destroy black's kingside, with attacking chances For this blog I decided to try studying this one position without using a computer. I haven't tried analyzing without a computer very much before, because it's too easy to get tired after a couple of minutes and turn on the engine to find out the correct evaluation. I am posting the analysis below, but I wouldn't recommend trying to look at it. It is a big mess and probably almost impossible to navigate.
After I was done analyzing I turned on the engine. Since it was a tactically complicated position, I expected to have a lot of mistakes in my lines even though I spent several hours on them. Indeed, there were many errors. Of the two main lines I studied, 23. fxg7 and 23. Bh7+, I reached a completely wrong conclusion in one of them. Bh7+ I correctly evaluated as being lost for white. But while I thought best play after fxg7 would lead to a repetition draw, it is in fact no better than Bh7+ and totally lost for white. I changed a few lines below to reflect some of the more serious corrections which were pointed out by the engine. Unfortunately, I was too lazy to delete any of the nonsensical lines, so it is an even bigger mess than before.
The biggest surprise from the computer was not in the lines I analyzed, but in the game continuation. It turns out that black is still winning even after the "blunder" 21...Nxe5, in two different ways, no less. I hadn't even thought about checking that position.
I'm looking forward to my next tournament which starts in nine days: the Chicago Open. It will be a good opportunity to meet some guys from chess.com, plus I have good memories of getting my first GM norm there last year.