The threat is simply Nxc8. Although the bishop is twice defended, the defence is insecure. The rook also needs to defend the queen, and the queen can't recapture first and is vulnerable to exchange before it finishes its job. Black can't summon another defender (for bishop or queen) and there is no safe square for the bishop to run away to. 

1.d4 Nf6 2.Bg5 A repeat of the opening from our Wellington Open 2015 game, which also ended in a draw  2...d5 3.c3 Bf5 4.Qb3 b6 5.e3 e6 6.Nf3 Bd6 7.Ne5 O-O 8.Nd2 c5 9.g4 Bg6 10.h4 h6 11.Bxf6 Qxf6 12.g5 hxg5 13.hxg5 Qxg5 14.Ndf3 Qf6 15.Nxg6 Qxg6

(15...fxg6!? ) 16.Qd1!? The Queen needs to be transferred to the Kingside, and it is most efficient to do this before castling  16...c4 (16...Nd7 ) 17.Nh4 Qg5 (17...Qh7 18.Bg2 Nd7 19.Qe2 Be7 20.Nf3 Qg6 = ) 18.Qf3

18...Nd7 (18...Nc6 19.O-O-O (19.Bxc4 Nxd4 ) 19...Rfc8 20.Be2 Qf6 = ) 19.O-O-O Rfd8 20.Be2 Kf8 21.Rdg1 Qf6 22.Qg2 b5 23.Rh3 Ke7

24.Rf3 (24.Qxg7 Qxg7 25.Rxg7 Rg8 ) 24...Qxh4 25.Rxf7+ Kxf7 26.Qxg7+ Ke8 27.Qg6+ Draw agreed, as White will apply perpetual check 1/2-1/2

In preparation for this game, I know there was a big chance of me encountering the Caro Kann. I often opt for the advance variation but haven't had much success. I arrived at the Accelerated Panov variation after seeing several interesting games.   1.e4 c6 2.c4 d5 3.exd5 cxd5 4.cxd5 Nf6 5.Qa4+ (5.Bb5+ is also playable ) 5...Bd7 Some prefer ignoring the d pawn completely with Nbd7 then g6, Bg7 and 0-0   6.Qb3 The point, causing a bit of nuisance on b7 after diverting the bishop to d7   6...Qc7 Despite being the most popular move, it has the worst stats   7.Nc3 g6 8.Bc4 Qc8 White was threatening d6 then Bxf7+   9.Nf3 Bg7

In the chessbase database, there has been 4 games that has reached this position, all resulted with white winning. Although I had my chance, I was unable to add to the statistics   10.O-O O-O 11.d4 Bf5 12.Re1 Re8 13.Bf4 (13.Nb5 a6 14.d6 axb5 15.Bxf7+ Kh8 16.Bxe8 is an interesting position ) 13...Nbd7 14.Rac1 a6 15.Ne5 Nh5 16.Bg3 Nxg3 17.hxg3 b5 I knew this move was coming and felt compelled to exchange on b5 to keep the initiative, Stockfish was not convinced   18.Bxb5 axb5 19.Nxb5 Qd8 20.Nc6 Qb6 21.Nxe7+ Rxe7 Or else white's next move will be Nc7   22.Rxe7 Rb8? Both sides had chances before this move despite black's two pieces for white's rook.  (22...Bf8 was the preferred move by Stockfish ) 23.d6!

23...Be6 (23...Rf8 24.Rc8! Be6 25.Rxe6 Rxc8 26.Re7 Rc1+ 27.Kh2 +- ) 24.d5?? This was the move that haunted me for the rest of the night  (24.Rxe6! fxe6 (24...Qxb5 25.Qxb5 Rxb5 26.Re8+ Bf8 27.Rc7 Nf6 28.d7 decides ) 25.Qxe6+ Kh8 26.Qxd7 Qxb5 27.Rc8+! allows a spectacular finish. Funnily enough I tried to use the same tactical motif of Rc8+ later on but under much inferior circumstances  ) 24...Bf5 25.Rc6? Qxb5 26.Rc8+ Bf8 The move I missed   27.Rxb8 Shocked and filled with various emotions, I played this move without much thinking in the desperate hopes of 27... Nxb8. Had I calculated further, I would've seen that white still has some play after  (27.Qxb5 Rxb5 28.g4! Bxg4 29.f3 Rxd5 30.fxg4 Rxd6 31.Rd8 ) 27...Qxb8 0-1

1.d4 c5 I play this move quite often. I don't know any theory at all, I just know that at lower levels it tends to get the game out of familiar paths quickly so that both players have to start playing chess at an early stage.  2.dxc5 A case in point, if White does this Black can already dream of taking over the initiative when playing his second move.  2...e5 3.Nf3 Nc6 4.Nc3 Bxc5 5.Ne4? During the opening phase White neglects his development a little too much. What's the old rule, "Don't move a piece twice until all pieces are developed  5...d5! A nice, although obvious trick.  6.Nxc5 Qa5+ 7.c3 Qxc5

Black has a development advantage and an ideal centre. I was very happy here.  8.Be3 Qd6 9.Qb3 Nf6 10.Bg5 O-O!? 11.e3 ( If 11.Bxf6 I intended sac'ing a pawn with  11...Qxf6 12.Qxd5 and Black will gain further tempi chasing the Queen ) 11...Ne4 12.Bh4 f6 13.Rd1 Be6

14.Qc2 ( When I played 12...f6 I calculated the variation 14.Bc4 dxc4 15.Rxd6 cxb3 16.Rxe6 bxa2 and somehow I imagined (in advance) that White couldn't stop me queening. In my minds eye I suppose there was still a bishop on f1 preventing the Rh1 from coming to the rescue. This is not the first gross calculation error I made in this game. In fact... ) 14...Rac8?? I assumed White intended freeing his game with Bd3 after which I can now play ...Nb4  15.Qxe4! Whoops. For the second week in a row my opponent steals a piece from under my nose in broad daylight. I've always liked mixed metaphors. Both times I would have been spared this misery if I had followed the old rule about always looking out for all checks and captures. Against Layla the rest of the game was a case of merciless exploitation of the advantage. Things worked out differently here. One factor that is different is that there is more chance of mixing things up this time  15...g5?! My idea is to exploit the fact that ...f5 comes with tempo in order to trap the Bishop and get my piece back. Of course this will cost some pawns, but it beats resigning right away.  16.Bxg5 Peter trusts me too much and avoids the bishop trapping line ( better is simply 16.Bg3 f5 17.Qd3 f4 18.Nxg5 Bf5 19.Qxd5+ Qxd5 20.Rxd5 fxg3 21.hxg3 with three extra pawns and the initiative ) 16...dxe4 17.Rxd6 exf3 18.Bh6 (18.Rxe6 fxg2 19.Bxg2 fxg5 is two extra pawns and the initiative ) 18...Rfd8?! I thought this was a very cunning trap. (18...Rfe8 would be sensible, with a two pawn deficit but a reasonable position to fight on with ) 19.Rxe6

19...Nb4!? The knight threatens Nc2 mate and steps away from the Re6 so that if I get in ...Kf7 I will win a whole rook. I am a piece and pawns down, but winning a whole rook will definitely bring me roaring back into the game. Meanwhile c3xb4 allows Rc1 mate. How cool is this? (19...Kf7 is the obvious move but two bishops against a rook is already a huge material advantage - and White has extra pawns as well ) 20.gxf3 Bugger. Not only does this secure e2 for the King, it allows Bf1-h3 protecting the rook (I missed this detail when playing 18...Rfd8). I almost resigned here, but I thought I saw one more remote hope...  20...Kf7 21.Bh3 Nd3+ 22.Ke2 Nc5

Hope dwells eternal. I thought that I was at least winning a piece back. Not a whole rook because unfortunately the Bh3 xrays my Rc8.  23.Rxf6+? Peter agrees with me. ( Neither player spotted the nice little trick 23.Rd6! after which White keeps all his material. It really would have been time to resign then  ) 23...Kxf6 24.Bxc8 Rxc8 The smoke has cleared and we've reached an ending with me three pawns down. Stockfish rates this position as about +2.99 so about 0.01 pawns of comp! Maybe. Is it worth playing on? I remembered a Lukey-Pomeroy post-mortem at the Wellington Club many years ago. Stephen made the observation that a many-pawns material advantage can be annoying because of the need to protect all the pawns! Maybe Peter will tire of this burden and let me take some pawns back!  25.e4 Na4 26.Be3 b6 27.Rb1

27...Rg8 ( I saw that after the tempting 27...Rxc3 28.bxc3 Nxc3+ 29.Kd3 Nxb1 30.Bc1! I lose my knight ) 28.f4! A good plan, liquidating the doubled pawns and setting up connected passers  28...Rg2 I'd love a passer of my own...  29.fxe5+ Kxe5 30.Kf3! ...nope bugger off  30...Rg6 (30...Rxh2?? 31.Bf4+ ) 31.Bf4+ Ke6 32.Bd2 White has placed both his pieces passively, this was encouraging  32...Nc5 33.b4 Nd3 34.Ke3 Ne5 35.h3 Rg2 Trying to get at the h pawn again. Stockfish actually wants White to play Rh1 to prevent this, but Peter decides that is just too passive.  36.Rd1 Rh2 37.Be1 Rxh3+ Woo hoo, now just two pawns down  38.Ke2 h5 I am not going to die wondering  39.f4 Rh2+ 40.Kf1 Nf3 This looks good but...  41.f5+ (41.Bf2! is very strong, eg if  41...h4 42.Ke2 wins the knight ) 41...Ke7 42.a4? This actually gives away the last of White's advantage.  42...h4!

This pawn is now hard to stop. Stockfish says 0.00  43.Bf2 The h pawn is now one square too advanced for this.  43...Rh1+ 44.Ke2 Rxd1 And the h-pawn is clearly going to cost White his bishop.  45.Bxh4+

This took me by surprise. I actually didn't notice it was check, and I thought Peter was trying something very risky giving me opportunities. Do I have a useful rook move that lets me keep a whole extra rook? I couldn't see one for some reason (neither player had time now), so I meekly captured the bishop. Next morning I sat bolt upright in bed. I should have played Rh1! I announced to my bemused wife. I was thinking that it is bizarre that I often see the obvious in chess when I am not playing and indeed not even really thinking about the position. It turns out I didn't exactly perceive the truth in my dreams and in fact it would have been really embarrassing to (try to) play 45.Rh1 at the board.   45...Nxh4 46.Kxd1 I now have nominal material equality, but I still prefer White. In fact I threw in a draw offer around here which Peter quite rightly declined.  46...Nf3 47.Ke2 Ne5 48.Ke3 Kd6 49.Kd4 Nf3+ 50.Kc4 Nd2+ 51.Kd3 Nf3 52.c4 Ne5+ 53.Kc3

I am holding White at bay and a draw is apparently the only outcome  53...Nd7?? Yuk, what a lemon, as soon as I released the knight I realised that I had given the square d4 for no reason at all.  54.Kd4! Stockfish goes from 0.00 to more than 5!  54...a5?! Pure desperation...  55.b5?? ...that yields a result. Instead either check wins ( eg 55.e5+ Kc6 56.e6 Nf6 57.bxa5 bxa5 58.Ke5 and the White king penetrates ) 55...Nc5? (55...Ne5! With a zugzwang was better ) 56.e5+ Ke7 57.Kd5 Nxa4 58.f6+ Ke8

I still thought I was in trouble here but actually White only has one drawing move.  59.c5?? Incomprehensible. Peter realised immediately what he had done. Presumably he forgot the knight was also covering c5. (59.e6? Looks good but  59...Nc5! holds both advanced pawns and wins  60.Kd6 a4 61.f7+ Kf8 and the knight keeps the White king from d7 ) (59.Kc6! Draws because Black can't afford to jettison b6 in this situation ) 59...Nxc5 Now there is nothing White can do.  60.Kc4 (60.Kc6 a4 61.Kxb6 Ne6 62.Kb7 a3 63.b6 a2 64.Kc8 a1=Q 65.b7 Qc3+ ) 60...Kf7 61.Kd5 a4 62.Kc4 Ke6 63.Kd4 Nd7 64.f7 Kxf7 65.Kd5 Nxe5 66.Kxe5 a3 67.Kd6 a2 68.Kc6 a1=Q 69.Kxb6 Qa8 70.Kc7 Qa7+ 71.Kc6 Qb8 Putting the queen on the queening square is the easiest way to insure against all accidents.  72.b6 Ke6 73.b7 Ke5 74.Kb6 Kd6 0-1