Anatoly Karpov was once quoted as saying that one of his interests in chess was in paradoxical positions (he was a fan of endgame studies). This game got like that in the end, when I was unfortunately in time-pressure, literally playing on the increment.
In the 2…Nf6 Scandinavian Opening, the quietest choice of lines for White may have simply been 3.Bb5+ Bd7, 4.Bc4 when White faces little danger of losing, and should be on top for a while.
Instead, I tried a line I had never played before but always kind of wanted to see, and that is after 3…Bg4, 4.Be2. I had tried 4.f3 before, and Black got a great attack; in fact it was against Teah where she lost her queen, but it was a tough game. OTB, I almost played 4.Nf3 in order to transpose into normal lines, but then realized that it could be met by 4….BxNf3, 5.QxB Qxd5, trading down early.
Even though I had seen her moves coming, I was still startled at how tactical that this opening became, and so it unfortunately consumed a lot of time on my clock – would not have wanted to see this in a rapid game for the first time.
9.a3, My longest move of the game. My original plan had been to play either 9.Ng3 or 9.Nf4, but Black will not trade queens but rather post his queen on g6 or f5 in response, to keep control over the light-squares.
9…e5. I was happy to see this move, and quickly played 10.d5 without even bothering to analyze it properly, since I didn’t really believe in her …e5 move on a gut/surface level.
OTB, 13…Nxc3 seemed, and is strongest, yet I quickly sort of dismissed it as being likely to be played as (besides my being forced into this variation) doubted that she would give up her queen for the two rooks, even though it is indeed good for Black.
16.Qa8+ Me and Alex both thought that the exchange sac on d7 was probably winning for White, but in fact it is bad with best play and better for Black. The biggest problem here was that I kept looking for a tactic, hence why I played the easily defended against short-term Nb5 move, instead of the long-term Nd5 move. Simply put, I was looking for cheapo tactics and mates. I considered 18.Nd5, but dismissed it due to 18….Bc5, and that is really short-term thinking, since the Nd5 can be supported by c4, and b4-b5 would kick the Bc5. It’s funny, chess is a balance where you can’t get too focused on short-term tactics nor on long-term strategy. You have to find the times when the position is calling for a tactic versus when it is calling for a simple, strategic, positional improvement. 18.Red1, attacking the unprotected e-pawn, is also a strong move.
23…f3? It appears that I am strengthening e4, but in fact am weakening the king’s position and dark squares surrounding it.
24…f5! I was ready for this, OTB, but now White is losing despite what an engine’s initial evaluation of this position may be. She took longer on this move than I thought she would, so by the time she played it I had seen that 25.Nc3 Qe3+, 24.Kh1 Rd8 was not looking good for me, and quickly finished finding the mate, after the game, where White continues with 25.RxR+?? NxR is mate in seven for Black, according to Stockfish (I found the cute B + Q + pawn mate right after the game). Instead 25.Nc3-d5 is the formation I was hoping to achieve during the game, but 25…Qe2, 26.Nbc3 Qf2, 27.Nb5 Rd7, and the engine disproves of 25.c4? after a long think (-1.4), and even 26.h3 is calling it -.96 after a while. Still, this would have been best.
25.g3 This move was understandable in terms of both kicking the queen, and giving the king some breathing room on the light-square g2, but now after 25…Qh6(forced) it makes playing 26.Nc3 impossible as Black has e4 where 26.fxe and 26.f4 are both met by 26…Ne5!! which I had found against the first line with Alex, but didn’t find against the second line until the computer suggested it, and then I was able to quickly calculate why that is.
27.h4 During the game, I didn’t trust the more normal looking 27.Rf1? because for one thing I saw that 27…Qh3 was annoying (-.76 or so), but after 27…Bb6, 28.Kg2 Rd8, the position is already over -3 according to Stockfish (even though I fed it the move 28…Rd8).
28.Rd5?? It is here, in the most paradoxical stage of the game, that I was playing with under half a minute on my clock. I considered first to play 28.Kg2, where if she plays 28….fxg3?! instead of 28…BxN! I can use that g3 pawn as a shield for my king. Instead, I am probing with my rook to get her to make an error, which she did.
28…Be3? Black had here the lovely continuation 28…fxg3!, 29.RxBc5 Qxh4!! letting the knight continue to hang instead of taking it, as the pawn on g3 is far, far more important than the Nf2.
30.Ng4?? It’s odd that I wanted to make a positional move here, instinctively, rather than a tactic, but with only half a minute didn’t have the time to figure out the flaws in neither the position nor the moves. I was attracted to 30.Ne4 initially, but played this tactic since it was easy to see that it wins a piece. The best move is 30.Nh4! preventing the future …Qh2+ in the most economical manner possible. The funny thing about this position is that it is given the eval of -2.75 but is one where Black could make some natural move and have the eval quickly drift toward positions where White could still hope to draw it.
31.QxBe3?? I saw 31…h5 coming, but was hoping that I could figure something out after playing this move, even though I continued to see nothing saving White the whole time. Instead, the paradoxical looking 31.Nxc7 (winning only a pawn instead of a piece, but weakening Black’s kingside much as she has been weakening mine). Now White has great swindling chances after 31…KxNc7, 32.Rb5! 32….Rb8??, 33.Qf7+ is equal. 32…b6??, 33.QxBe3 is also equal as is 32…Nd8??, 33.Qc3+ or 33.QxBe3. After 32…Ba7?? 33.Rxb7+ is mating for White! In fact, even after the correct 32…Rd8!, 33.QxBe3 h5??, 34.Qb6+! would draw, and thus Black needs to make one more move with foresight such as 33…Kc8! since 34.Qb6 could be now met by 34…Rd2+, which is crushing, and other moves allow 34…h5, winning the Ng4 at long last. Curiously, if Black does not take the Nc7 on move 31, then 32.Na6+ is either a perpetual or mating for White. So, there were still hard problems for Black to solve after the blunder 30.Ng4?? and also on the move before, but I didn’t pose them to my opponent in my time-pressure.