Patrick O'Sullivan (1750) vs Dan Raats (1822) (Active)

This game was played in the first round of the Kitchener-Waterloo Spring Active tournament.  It is particularly notable because I was offered a draw by my expert-level opponent (standard rating 2034) which I declined because there was no possibility of losing for me if the position hadn't changed character.  My plan was to slowly build my position and if I could break through, I would, and if I couldn't, I would have offered the draw back.  I ended up losing this game by playing a bold Knight sacrifice that didn't work over the board, but actually turns out to draw.  I blundered in the follow-up because I was still committed to winning the game.

Patrick O'Sullivan (1750) vs Dan Raats (1822)
1.e4 e6 2.d4 d5 3.Nc3 Nf6 4.e5 Nfd7 5.f4 c5 6.Nf3 Nc6 7.Be3 cxd4- An odd choice.  I am sure this is the most familiar system for my opponent, but he is a higher rated player than me and this system leads to a very even slow game.  However, with that said, I would not deviate from my Caro-Kann based on my opponent's rating.
8.Nxd4 Bc5 9.Qd2 Nxd4 10.Bxd4 Bxd4 11.Qxd4 Qb6 12.O-O-O- I should point out that 12.Nb5 is the main line because the Knight seems better on d4 than the Rook, but my d4 Rook allows for a tactic later that I unfortunately failed to capitalize on.
12...Qxd4 13.Rxd4 a6 14.g3 Ke7 15.Bg2 Nb8 16.Rhd1 Nc6 17.R4d2 b5?-  There has not been much to writ about in this slow positional battle until now.  The Queenside thrust 17...b5 is a blunder because it allows me the shot 18.Bxd5!.  During the game, I looked at this possibility and concluded that I was winning the a8-Rook if 18...exd5 19.Nxd5+ and king anywhere but f8.  When I looked at 19...Kf8 I glanced at 20.Nb6 Rb8 21.Rd8+ but saw that the black Knight covers the square.  However, I did not look deeply enough, as 21...Nxd8 is met by 22.Rxd8+ Ke7 23.Rxh8 Rxb6 24.Rxc8 and white is winning as he is up two pawns with more to come.  It is interesting that at the critical moment I failed to look deeply enough into the position, just like in the game below when I considered the line "ending" with 36.g4.  In both cases had I looked one move deeper, I would have avoided a serious oversight.  I am not too upset though, as these games were played under active time control; discouraging a re-checking of calculations.

P

Position after 18.Bxd5! (Analysis Diagram)
18.Ne2- 18.Bxd5! is much better as explained above.  The game now enters a maneuvering phase.
18...Bb7 19.Nd4 Rac8 20.a3- Stopping the Queenside attack that begins with b4.
20...g6 21.g4 h5 22.g5 Rc7 23.h4- Always fix the opponent's pawns on the same colour of his bishop.
23...Rhc8 24.c3 Na5 25.Bf1 Nc4 26.Bxc4- The Knights are much better than the Bishops in this position.
26...Rxc4- bxc4 is better because the long-term pressure on b2 would be hard to handle.  The black Rooks have no play because of the white pawn on c3.
27.Kc2- Beginning a King walk that seals white's advantage
27...R8c7 28.Kd3 Bc8 29.Ke3 Bd7 30.Nb3?- I thought that this stops the Queenside pawns, but Black is now slightly better after 30...b4! 31.axb4 Ba4 and black wins the exchange.  I did not see this possibility over the board.
30...Re4+ 31.Kf3 Rcc4 32.Rd4 Rexd4 33.Rxd4 Rxd4 34.Nxd4- I survived my mistake and now hold an edge because my Knight is much better than the Bishop.
34...Kd8 35.b4- Now the King can never penetrate my position.
35...Kc7 36.Nb3 Kb6- This move came with a draw offer which I did not accept because there is no threat to my position.  However, I had a strong feeling that the game was drawn.  I did not know that my opponent was such a strong player, so I felt I could trudge on and take the draw in the later stages.
37. Ke3 Be8 38.Kd4 Bd7 39.Nc5 Bc6 40.Nxa6?!- The computer is harsh with this move, giving the position -1.10 but white turns out to be fine.  My real mistake came later on in the game.  The plan behind Nxa6 was to crash through with the King and destroy the Black Kingside pawns.  I did not see that the bishop was in time to defend.

Position after 40.Nxa6?!.  The game is still a draw here, as the black King cannot enter the white camp.
 40...Kxa6 41.Kc5 Kb7 42.Kd6 d4 43.cxd4 Be4 44.Kc5 Bd3 45.d5??- The losing move.  The black King has no way in if the white King swings back and forth between d6 and d7, because the Bishop cannot hinder the King's movements.  Now the Bishop has the key e6 square, which decides the game.
45...exd5 46.Kxd5 Kb6 47.Kd6 Bc4 48.Kd7 Bd5 49.Kd6 Be6- No more d7, so the game is lost.
50.Ke7 Kc6 51.Kf6 Kd5 52.f5 gxf5 53.g6 fxg6 54.Kxg6 f4 55.Kxh5 f3 0-1

This game demonstrates the importance of re-checking all variations and of devising endgame "schemes".  I found the Kd6 Kd7 Kd6 idea one move too late, when I really should have seen it as a fallback option on 40.Nxa6.  Even in benign positions, you must be mindful of your opponent's threats to avoid catastrophe.  Both of us missed chances to grab an edge, but that is the nature of active chess.  The goal is not always to see the deepest into things, but to find the best ideas.