I played 11…Nd4 looking for a draw, as I hadn’t seen him face or have to react to this move before. In reply. he played 12.Ne4, which I had, in return, never seen before while remembering it shouldn’t be best move. Everything looked good for him, and I spent half an hour before deciding to sac the pawn.
To get an idea of what I was looking at. 12…Bg4?? loses, 12…Bf5?? loses, 12…Bd5? loses, and I saw 12…Bc4 (which looks great at first), and now instead of 13.Rf1-e1, I correctly saw that White plays 13.NxN BxN, 14.Be4 +3, and is winning here too. It’s really startling OTB to see a move you did not expect, where the majority of reasonable replies can lose easily. In the post mortem, I played 12… dxc3 13. Nxc5 Bxc5 14.Be4 Qd7 15. Qxd7+ Bxd7 16. e6 fxe6 17. Ne5 Nxe5 18. Bxa8 O-O 19. Be4 Ng4 20.Bf3 and here wondered if I could play 20… Nxf2 21. Rxf2 e5, which is -1.12 in Black’s favor, according to Stockfish. Imre thought that White was winning after 20.Bf3 because I didn’t risk taking on f2, and his tactics were rather pretty, and he was playing them at bullet-speed, which is rather impressive at his age, that he can still do it (at around 80).
After 14.QxQd8, which I had expected, I was down on myself, thinking I had missed the answer to the previous position. Actually, Stockfish says my play was best after his 12.Ne4 move, but that here is where I make the mistake. I was going to play the correct 15…RxQd8, when I noticed that my a-pawn was not only hanging, but that after 15…Rxd8 Nxa6??, I couldn’t play 16…Rd8-a8 because of 17.Nxc7+. Well, I had spent so much time on the previous combination that I decided to move quickly here with 18.Nc6-d8?
I had seen 15…RxQd8, 16.Nxa6 Bc4, which is what I would have played, but didn’t see that he is busted after simply 17.Nxc7+ Kd7! (I mis-visualized it, thinking the bishop would still be on e6 here). That is the thing, I can’t exactly blitz visualize. After 17.Rfe1 Rd8-d1, 18.Be3 Be2 (I had seen this idea, to remove the Nf3 defender of e5 this way, but not that it also is not hanging because his Ra1 would hang, and meanwhile Be2 protects the Rd1. It’s interesting even here how White can play 19.Nxc7+ Kd7, 20.Nxb5 RxRa8, 21.RxR Be2xNb5, 22.Ne1 Nxe5, 23.Nxc2, and now White has three connected passed pawns for the piece -.86. That line may sound long, but at 30 seconds increment, you just gained four minutes if you play that, and your opponent will spend time on such a sequence as well.
If this sounds too wild for your tastes, then you have 16.Nxa6 0-0, 17.Be3 (Bg5 is better), and now 17.Nxc7+?? would be a mistake. So, after 17.Be3 Rd7, 18.Nc5 BxN, 19.BxB Rd1, 20.Be3 Ra8, and Black is better here =+, where …Bf5 could still protect the c2 pawn in some lines, so 21.Nd4 and 22.Nxc2 is likely, when Black will try to take the e5 and a2 pawns.
16.Be3. 16.Ne1 (idea of Nxc2, was most accurate, according to Stockfish).
16…0-0? 16…Bc5 played now is most accurate. Not only can it not be defended by the Nf1 move here, but after 17.Rfe1 BxB, 18.RxB?? 0-0-0 with …Nf4 idea is -+.
19…Rad8? I saw this was a mistake as soon as I played it because now he can simply double-rooks and force me to trade off both of them. The obvious 19…RxR was best.
23…Kf8. This move might seem obvious, but 23…g5 is best according to Stockfish, and should be.
24…f4. I was expecting 24…g2 here, and saw his blunder right away (as did he). After the game, he said that 24…f3 was best, and Alex liked this as well (they both liked it better than g3). 24.g3 is +.88, whereas 24.f3 is +.32. 24.f3 BxBe3+, 25.NxB Nf4 (well 24. f3 Bxe3+ 25. Nxe3 Nf4 26. Kf1 Ke8 27. b3 Nd3 28. Nd5 Kd7 29. f4 Nc1 30. Ke1 Nxa2 31. Kd2 a5 32. Kc2 b4 33. c4 c6 34. Nb6+ Ke6 35. Na4 Kf5 36. g3 g5 37. fxg5 is = according to Stockfish). So, in that line 26.Nc2 (26. Nc2 Nd3 27.Nb4 Nxb2 28. Nxa6 Nd1 29. Nxc7 Nxc3 30. Kf2 b4 31. Na6 Nxa2 32. Nc5 h6 33. Ke3 Ke7 34. Kd3 g5 35. Kc4 Nc3 36. g3 Ne2 37. Kxb4 Nd4 38. f4 Nf3 39. Kc4 Nxh2 40.Ne4 Ke6 41. Kd3 gxf4 42. gxf4 Ng4) is equal and a draw.
The 24…g3 line is a lot trickier for Black, and OTB it’s almost surely a win for White.
27.Nd5. I decided that if 27.Ke2, that …Nxe5 is correct (told this to Alex), and …Nxb2? is a mistake, and Stockfish confirms this.
28.Nc7 28.Ke2 is still stronger according to Stockfish, and this makes sense.
I could have played 28…Nxb2, saw this, and he wins a-pawn and I e-pawn, but I played the slightly more accurate 28…a5 to control the b4 square, and also to avoid uneccessary trades.
29.e6? I had seen this much earlier, and had planned to play as I did, so it was old news and a bit surprising to see OTB.
30…Ke7. I had seen that 30…Nc1 wins a pawn, but didn’t want to give up control of b4 yet, and already could see that this position was a win.
34.Na6? Now I knew it was really over, as I could see that 34.Na8 was the only way out of this fortress for the knight. I had seen before this move that I could trap the knight after 34…Kd7, 35.Kd3 Kc8, 36.Kd4 g6 and now Black rounds up the knight (Stockfish confirms this line). I didn’t play it because of any possible counterplay, and I simply didn’t need the knight, saw that I could force trade it into a won king and pawn endgame. Plus, I wanted anyone watching to know that I could win this pawn ending, which I did.
The biggest take-away from this game is that I used my clock woefully in the middlegame, but then used his time well in the endgame (had 12 minutes remaining at the end). I can calculate well in the endgame when I am not under 5 minutes, and I have a 30 second increment. Where my skill-set is really lacking is in calculating quickly in the middle-game. Imre is one player who can calculate more quickly in the middlegame than I, and can be more deadly instinctively. If I have to make quite a few middlegame calculations, by contrast, then I can become busted on the clock. It would be great if I could improve this skill one day.