Friday, March 28, 2008
Some excellent tactical play
Wellington Open (2) 2008
This game was another sharp one from the Wellington Open.
1.d4 d5 2.c4 c6 3.Nc3 Nf6 4.Nf3 dxc4 5.a4 Bf5 6.Ne5 Nbd7 7.Nxc4 Qc7 8.g3
I realised I was entering a very sharp opening and in fact one I didn't really know anything about but I decided not to let my opponent dictate the terms on which we would play.
8...e5 9.dxe5 Nxe5 10.Bf4 Nbd7 11.Qd4 f6 12.Rad1 g5?!
12...Bc5 would have lead to an equal position with chances for both sides. None of the Black pawns are really weak after 13.Bxe5 fxe5 14.Nd6+ Bxd6 15.Qxd6 Qxd6 16.Rxd6 and a roughly level endgame arises. There are also other possibilities for White but none lead to a clear advantage.
13.Bxe5 fxe5 14.Qd2 h6 15.Bg2 Bb4 16.O-O O-O-O 17.Qc1!
I didn't like the opposition of white queen and black rook on the open d-file and was looking to move my knight preparing b2-b4-b5 with chances to develop an attack on the black king position. It turns out this is the winning move because my opponent over-estimates his position and interprets this move as a mistake.
17...Nc5?
Thinking that the queen is in trouble... We didn't find a clear evaluation after 17...h5 and White can take the pawn with 18.Qxg5 Be6 or continue with the attack via 18.Na2 Be7 both positions offer chances to both sides.
18.Na2 Nb3 19.Qe3 Rxd1 20.Rxd1 Bc2 21.Qxa7!
I had also prepared the bail-out variation 21.Rf1 though after 21...Bc5 22.Qc3 Nd4 23.Re1 the position is very unclear and Black might be able to take a pawn with ...Bxa4
21...Bxd1
21...g4 does not improve Black's predicament 22.Nxb4 Bxd1 23.Nb6+ Kd8 24.Qa8+ Ke7 25.N6d5+ cxd5 26.Nxd5+ Kd6 27.Qxh8! Qc1 28.Qd8+ Kc5 29.Qb6+ Kc4 30.Ne3+ forces Black to give up the queen.
22.Bh3+
Not 22.Nb6+? Qxb6! 23.Qxb6 Bc5 as my opponent pointed out after the game.
22...Kd8 23.Qa8+ Ke7 24.Qxh8 Bxe2
Of course Black is totally lost with or without this move. 24...Bc5 25.Qh7+ Kd8 26.Qd3+ picks up a piece as well.
25.Qg7+ Kd8 26.Qg8+ Ke7 27.Nxb4 g4 28.Qg7+ Kd8 29.Qf8+ Kd7 30.Nd3
This was the point though I stuffed up the execution slightly.
30...gxh3 31.Ncxe5+?!
Instead 31.Ndxe5+ Ke6 32.Qg8+ Kf6 33.Qg6+ Ke7 34.Qf7+ Kd8 35.Qg8# is the most accurate. White could have been forced to retract his steps with 31.Ncxe5+ Kd6 32.Nc4+ Ke7 33.Qg7+ Kd8 34.Qf8+ Kd7 35.Nde5+ reaching the mating setup.
31...Ke6 32.Qg8+ Kf5 33.Qg6# 1-0
It's pleasant when your opponent doesn't see the mate coming.
The only lesson from this game is the value of good accurate calculation. Black had little idea what was going on during this game because he simply didn't see all the critical tactical ideas in time.
Wednesday, March 26, 2008
Fireworks in the last round
Wellington Open (6) 2008
Played during the sixth and final round of the Wellington Open. By winning this game I finished in a share of third place with Russell Dive. It was a very difficult tournament especially over just 6 rounds.
1.d4 d5 2.c4 c6 3.Nc3 Nf6 4.Nf3 e6 5.Bg5 Bb4 6.e3 Qa5
Somewhat as I guessed at the time Alan wants to play the Cambridge Springs variation of the Queen's gambit, however he doesn't know the correct move order to reach it. In the Cambridge Springs White can't double the Black pawns on f6 because Black develops his other knight to d7 instead of bishop to b4. My next move is forced otherwise Black will get easy equality or better simply by following up with Ne4.
7.Bxf6 Bxc3+ 8.bxc3 Qxc3+ 9.Nd2 gxf6 10.Rb1!?
10...Nd7
If there was some improvement I would be looking for it instead of this move. Actually I think that White is better already in any case because the only place the Black king will be perminantly safe is on the kingside which is clearly a risky venture. If the Black king remains in the centre then White has lots of compensation as the process of realising the extra queenside pawn will be a difficult one. Even if Black captures the second pawn on a2 White should not panic but keep on developing and slowly opening the centre. Eventually the Black king will be exposed there.
11.cxd5 cxd5 12.Rb3 Qa5 13.Bd3 Rg8 14.O-O f5 15.f3 Nb6 16.e4 f4 17.Bb5+Something like this was bound to happen at some point. Maybe 17...Kf8 keeps the king more secure.
17...Bd7 18.Bxd7+ Kxd7 19.exd5 exd5!?
Now with the king sandwiched between two open files the situation looks dire, however the only variation I discovered here leads to a tense situation where most White pieces are hanging and defending each other. 19...Qxd5 20.Ne4 looks wrong on principal and I get the feeling that this was as bad so suggesting it as an alternative just because Black lost with the game continuation would be rather silly.
20.Qc2! Rg5
A check on f5 would be devistating. Self defence by the king with 20...Ke6 is also not sufficient after 21.Rf2 Rac8 22.Qxh7 Rc1+ 23.Nf1 White has the secure king and should soon be material ahead.
21.Qxh7! Ke7
If Black takes a piece with 21...Qxd2 then after 22.Qxf7+ White forces a dark squared check and picks up the lose rook on g5 quite easily. I expected 21...Rag8 22.Qxf7+ Kc8 23.Rfc1+ Kb8 24.Qxf4+ Ka8 25.g3 and White has secured his king position but I did not realise how hopeless this really is. Alan's move is a much better defence.
22.Re1+ Kf6 23.Qh6+ Rg6 24.Qxf4+ Kg7 25.Re7! Rf8 26.h4?!
The position is very tense and not only difficult for Black despite the open Black King and White having an extra pawn. Many of the White pieces are defending each other in long chains and there is a constant problem with a mate on the g2 square in some variations. For example after the careless 26.Nf1 Qxa2 White is forced to reply 27.Nd2 again to meet the dual threats to b3 and g2. Maybe better however was 26.g4! defending against one of the threats. A follow up variation might be 26...Qxa2 27.Rb1 Qc2 28.Re1 and White can start thinking about getting the rooks doubled on the seventh rank. The position remains very sharp however it seems this plan will eventually lead to a break through which I am not sure of a good follow up to the variation mentioned in the next note.
26...Kg8?
Alan flinches and now the h-pawn disrupts his position and it breaks down. Instead 26...Qxa2 27.h5 Rh6 28.Qg5+ Kh7 leads to a sharp position where White might be best to take a perpetual check considering how many White pieces are hanging. I certainly have not found any plan which clearly breaks down the Black position from here.
27.h5 Rg7 28.h6 Rg6 29.h7+
29...Kxh7
29...Kg7 30.Rbe3! cleaverly guards the knight on d2 and prepares something similar to the game because if 30...Qxd2 31.Rxf7+ Rxf7 32.h8=Q+ Kxh8 33.Re8+ Kg7 34.Qxd2 +- White should win. Instead Black can try 30...Rf6 31.Qg5+ Rg6 32.Rxf7+ Rxf7 (32...Kxf7 33.Re7+ Kg8 34.Qxg6+ Kh8 35.Qh7#) 33.h8=Q+ Kxh8 34.Qxg6 Rf8 35.Re7 and White delivers mate shortly.
30.Rxf7+ Rxf7 31.Qxf7+ Rg7 32.Qh5+ Kg8 33.Re3! Rxg2+?!
Russell Dive suggested the defensive move 33...Nd7 with the idea of 34.Re8+ Nf8 and it looks like Black has defended and is only a pawn behind however after 34.Ne4! Black remains in trouble for example 34...Qxa2 35.Qe8+! (Not 35.Ng5?! Qb1+ 36.Kh2 Qf5 and with the queen involved everything is covered.) 35...Kh7 (35...Nf8 36.Nf6+ wins) 36.Ng5+! Kh6 (36...Rxg5 37.Qxd7+ Rg7 38.Qh3+ Kg6 39.Re6+ Kg5 40.Qh6+ Kf5 41.Rf6# or something similar.) 37.Nf7+ wins the exchange because of 37...Kh7 38.Qh8+ Kg6 39.Ne5+ winning even more.
34.Kh1
Of course not 34.Kxg2?? Qxd2+ and Black picks up the rook at e3 with an extra piece.
34...Qxd2
I didn't calculate the mating sequence before reaching this position. It is obvious that there is going to be a forced mate with queen and rook against the lone king. If necessary I could simply have won the rook on g2 however.
35.Re8+ Kg7 36.Qh8+ Kg6 37.Rg8+ Kf7 38.Rf8+ Ke7
A shorter alternative was 38...Kg6 39.Rf6+ Kg5 40.Qh6# and 38...Kf5 39.Rf8+ Kg5 40.Qf6+ Kh5 41.Rh8+ Qh6 42.Rxh6# is also mate.
39.Qf6+ Kd7 40.Rd8+
Alan chooses to resign, one move before I deliver mate with 40...Kc7 41.Qd6#
1-0
Lessons from this game
a) The solutions to tactical problems often have a logical basis. During the final assult it could be noted that the Black king lacks defenders. The only defender is the rook (which vacates it's post shortly before the end) in the position after 33.Re3 the key move brings in another attacker and better defence was offered after 33...Nd7 bringing in another defender. Also earlier with 26...Kg8 it would be possible to note that the queen is isolated from the defence and the maneuvre 26...Qa5xa2-c2 prepares to bring it back into the game. Of course this is usually much easier to see in hindsight but these sorts of evaluations of tactical problems will often save much time finding the best moves where this is possible.
b) Some positions are simply too sharp, complicated or difficult to evaluate or calculate accurately. In these situations you should simply play with the principals you know and have some faith in them. Of course I could not guarantee that I had compensation for the pawn or for two pawns which Alan might have taken later but it was clear that he was going to have to play well in order to survive the opening and middlegame and emerge victorious.
c) Often in a sharp situation with many potential threats a move which stops one or more of these threats is the strongest. With 26.g4 I stopped Alans principal threat of a checkmate on g2. This was the most promising continuation, allowing me to solve my current tactical problems and bring the knight into the attack. The move I played, 26.h4 was less good because although it eventually lead to a tactical breakthrough with best play Alan might have made his own threats count.
Saturday, February 23, 2008
The advantage of being the stronger player
Justin Davis - Nicolas Croad
1.d4 Nf6 2.c4 e6 3.Nf3 b6 4.a3 Bb7 5.Nc3 d5 6.cxd5 Nxd5 7.Bg5?!
This move must be criticised and blamed for giving away any opening advantage available to the White player. Justin probably decided he would have more chance of keeping pace with me in a simplified position with fewer pieces on the board but on positional grounds White has a slight space advantage and exchanging the dark squared bishops affords the Black queen a nice development square on e7.
Some people would probably argue that because White can be somewhat more satisfied with a drawn result he can use this to his advantage by simplifying the game and trying to force Black to take risks to avoid a draw but I do not really agree. As this game shows even in a simplified situation there are often plenty of mistakes to be made and if I am actually stronger than my opponent then I should expect I will exploit my chances better. In other words by playing inferior moves for the purpose of exchanging pieces White is gradually heading towards a situation where I can put pressure on him not one where he will pressure me.
7...Nxc3
I decided to close the c-file before challenging the bishop. After this exchange there is no way for a rook to pressure me along that file so I should have no trouble executing c7-c5.
If 8.Bxd8 Nxd1 9.Bxc7 Nxb2 White loses a piece.
8...Be7 9.Bxe7 Qxe7 10.e3
10...O-O 11.Be2 Nd7?!
12.O-O c5 13.Qd2 Rfd8 14.Rfd1 Rac8 15.Qb2 Nb8!
Both sides have completed their development and now I needed to find a plan. It is still relatively difficult for White to improve his position so I felt that I was better placed but there are no obvious weaknesses in the White camp. Maybe another reasonable square for the knight was on e4 after 15...Nf6 but it seems that the reply 16.Ne5 with the intention of exchanging the light squared bishops was quite strong here.
16.Rac1
Maybe 16.Nd2 Nc6 17.Bf3 trying to exchange black's best minor piece was a little better.
16...Nc6 17.e4 Na5
A rare case of the knight being well placed on the edge of the board. While it never occupies the c4 square during the game this may become possible at some point. It stops White playing a4-a5 and breaking up the Black queenside and this combination allows Black to develop the remainder of his position while the knight stops much of White's counterplay. In lots of variations it also helps by guarding the bishop on b7.
18.Nd2
There were still some useful moves for Black to play before carrying on with the decisive offensive. After 18...h6 19.f3 Qg5! 20.Nf1 cxd4 21.cxd4 Rxc1 22.Rxc1 Rxd4 wins a pawn and shows how difficult to hold together the White position has become. Notice if the knight was not guarding the bishop on b7 now 23.Rc7 would win material here.
19.cxd4 Qd7
Instead 19...Qd6 20.e5 Qxd4? 21.Rxc8 Bxc8 22.Nf1 Qxd1 23.Bxd1 Rxd1 24.Qc2! Rd8 25.Qc7 Rf8 26.Qxa7 White is clearly winning now.
An excellent move which shuts down the long diagonal to the extent that Black's light squared bishop becomes cramped for space.
21...Qc7
This was part of my plan when I played 18...cxd4 I control the c-file and that can not be challenged. It turns out however that even though I can stop the passed pawn that White creates in the centre White gains enough counterplay with it to hold the balance.
22.Nb3 Nxb3 23.Qxb3 Qc3
I also considered 23...Kf8 but Black has gained nothing after 24.Qb4+ Qe7 25.Qb3 anyway and further pushing in this direction will only make Black's position more difficult.
Forced by the variation 28...Kd8 29.Re1 with the threat of Re8#
The decisive mistake allowing me to create two connected passed pawns on the queenside. Instead White needed to more accurately exploit the same idea with 29.Bxc6+ Rxc6 30.Re1+ Kd7 31.Re7+ Kxd6 32.Rxa7 Rc7 with an endgame which offers no advantage and very few winning or losing chances to either side.
29...Kd7 30.Bxc6+ Kxc6!
Instead 30...Rxc6? 31.Re7+ Kxd6 32.Rxa7 Rc7 leads to the same position as White should have forced in the note above.
31.Re7 Rxa3
White was threatening Rc7+ anyway. Now the two protected passed pawns on the queenside are more dangerous than their White counterparts on the kingside.
32.Rxf7 b5 33.d7
The variation 33.Rxg7 b4 34.Rg4 (34.Rxh7 b3 35.d7 this is the most inconvenient moment White can arrange (35.Rh4 b2 36.Rb4 Ra1+ 37.Kf2 b1=Q 38.Rxb1 Rxb1 -+ is the sort of position a player might avoid because of the three white passed pawns against a rook. Converting this will not present a particular challenge because Black has a passed pawn of his own which rather forces White's hand 39.h4 Kxd6 40.h5 Ke6 41.g4 Kf6 42.f4 and since the pawns can't allow themselves to be blockaded this is as far as they are going to go.) 35...Ra1+ 36.Kf2 Rd1 37.d8=Q Rxd8 38.Rxa7 b2 transposes into a position where White can no-longer stop the Black pawns promotion.) 34...a5! 35.Rd4 Kd7 36.f4 b3 37.Rd1 b2 shows which passed pawns are the most dangerous.
There are no alternative defences now anyway.
33...Rd3!
After 33...Kc7 34.d8=Q+ Kxd8 35.Rxg7 b4 36.Rxh7 b3 (or 36...Kc8 37.Rh4 a5 38.Rh5 b3 39.Rb5 a4 40.h4 with the active rook going to White again.) 37.Rb7 a5 38.h4 the win comes into doubt. The difference between these positions and the game is which side has the active rook. Actually this pawn race has some nice tactical twists to it. After 38...Ke8 39.h5 Kf8 40.h6 Kg8 41.g4 a4 42.g5 b2! 43.Rxb2 (43.g6 Rb3! -+) 43...Rxf3 = the game should be drawn. The game can also end with a draw after 41.f4 Ra1+ 42.Kf2 (42.Kh2 Rf1 43.Kg3 a4 = with ...Rh1 to follow) 42...Ra2+ 43.Kf3 b2 44.f5 Ra3+ 45.Kf2 (45.Kf4? Ra4+ 46.Kg5 Rb4 -+) 45...Ra2 = and Black must take the repetition because of 46.Kf3 a4? 47.f6 Ra3+ 48.Kf2 Rb3 49.f7+ Kf8 50.h7! when White wins.
There are no longer any moves which stop the b-pawn from becoming a queen.
0-1
Wednesday, February 20, 2008
Winning with the Caro-Kann
Summer Cup, Wellington 2008
1.e4 c6
Some people think the Caro-Kann is a drawing opening however their judgement is often based on elite level chess which is quite different from games between non-master players. I think that the Caro-Kann is excellent for use against lower rated opposition (who might be trying to dull the game down) because there are no symmetric pawn structures and there are several different ways to play the opening. Mark also plays the Caro-Kann so he was also facing the challenge of creating pressure against his own defence though he would typically choose the Larsen variation 4...Nf6 5.Nxf6+ gxf6 if he was black.
2.d4 d5 3.Nc3 dxe4 4.Nxe4 Bf5 5.Ng3 Bg6 6.h4 h6 7.Nf3 Nd7 8.Bd3 Bxd3 9.Qxd3 e6
10.Bf4 Qa5+ 11.Qd2!?
After the game we also discussed the possibility to play 11.Kf1 (or even 11.Ke2) with the idea of artificially securing the king with Kg1 and keeping an attacking rook on the h-file. Still this plan looks rather too artificial to cause Black serious problems.
11...Bb4!?
Maybe the most simple reply was 11...Qxd2+ 12.Kxd2 Ngf6 and the position is completely level as neither position has any significant weaknesses or clear plans. I was trying to discourage Mark from castling queenside and I didn't feel that the position could be too bad as otherwise 11.c3 would be a strong reply to 10...Qa5+ forcing a similar situation.
12.c3 Be7 13.b4 Qa6 14.Qe2
This move is not very incisive, it was possible to put significant pressure on the Black position with 14.Nh5! Ndf6! 15.Nxg7+ (15.Ng3 Nd5 16.Nh5 Ndf6! =) 15...Kf8 16.Nxe6+ fxe6 17.Ne5 Rh7 18.Rh3 which can only be judged as unclear, because it is almost impossible to see how Black is going to break free or White is going to develop the initiative.
14...Qxe2+ 15.Kxe2 Ngf6 16.Kd3?!
It was probably better to correct the earlier error and play 17.Ke2.
17...O-O 18.Ne4 Rfd8 19.Bc7 f5!? 20.Nc5?
This loses a pawn which Mark simply missed. Much better is 20.Ned2 when the position remains roughly level though Black can take the initiative. An interesting follow up is 20...e5 21.Bxd8 e4+ 22.Kc2 (22.Nxe4 fxe4+ 23.Kxe4 Rxd8 is excellent for Black.) 22...Rxd8 and Black has excellent compensation for the exchange. This might be refuted by 22.Ke2! Rxd8 23.Ng1 Bxh4 24.g3 when Black has a pawn for the exchange but no immediate threats.
Better though less incisive is 20...Rc8 21.Bf4 (21.Bg3 g5!) 21...a5 and now if 22.a3 axb4 23.axb4 Nxf2+ wins a pawn so Black has achieved a reasonable amount of play to compensate for the slight weakness of the e6 pawn.
20...Nxc5+ 21.bxc5 Rd7 22.Bg3 Bxc5 23.Rae1 Kf7
however this would be completely unreasonable. There is no danger of this tactic succeeding because there is no way to simultaniously oust the knight from g4 and setup the threat. In the absence of this tactic Black should use the piece with the least potential (the king not the rook) to defend the weakness on e6.
24.Nd2 b5!
Stopping the white knight from reaching c4 and creating an excellent post for my own knight on d5.
25.Nb3 Bb6 26.f3 Nf6 27.Re5
27...Nd5 28.Re2?!
28...Rc8 29.c4
30.Kxc4 Rb7 31.Rfe1?!
31...c5 32.Rxe6
34...cxb3 35.R1xe3 bxa2
a) The principal of two weaknesses is a significant component in the selection of a winning plan. Here I felt that my position was near to winning before playing 29...bxc4+ but I didn't use this principal to decide on a winning plan following my chosen move and this lead to an over-estimation of my position and the selection of an inferior move.
b) It is preferable not to defend against ghosts. I was able to make a good defensive move 23...Kf7 not restricting the mobility of one of my rooks because I was being concrete about the threat of a knight fork on e5. This paid off in the game because my rook was never forced to defend the weakness on e6.
c) Sometimes a pawn should not be captured when it is blocking the lines for the opponents pieces and/or forces them to defend it. In the line beginning 29...c5 this principal turns a marginal sacrifice of a piece for three pawns into a powerfull one of a piece for two pawns.
Friday, February 8, 2008
Anonymous comments enabled
Wednesday, February 6, 2008
Trying to learn from your games...
Martin Hill - Nicolas Croad
Summer Cup, Wellington 2008
1.f4 c5 2.Nf3 g6 3.e4 Bg7 4.Nc3 Nc6 5.g3 d6 6.Bg2 e6 7.O-O Nge7 8.d3 O-O 9.Be3 Nd4 10.Qd2 Rb8 11.Nd1 b6!?
Up till this point we have been following the main-line of the Closed Sicilian. I was happy with this transposition from the Birds opening as I know this opening reasonably well and considered it gives Black good prospects to play for a win without taking considerable risk. My last move however perhapse requires some comment. I did spend some minutes trying to find a less passive looking alternative. The move itself is prompted by the variation 11...b5 12.c3 Nxf3+ 13.Bxf3 b4 14.c4 and here I thought Black lacks counterplay having closed the queenside while the d6-d5 break is made difficult by the poorly defended c5 pawn. Now if White plays 11.c3 then after 11...Nxf3+ 12.Bxf3 d5 13.e5 d4 Black is certainly no worse. This was my plan however since Black could also play 12...f5 here with some pressure on the White centre it may be that 11...b5 12.c3 Nxf3+ 13.Bxf3 f5 was a quite reasonable alternative though here I would still prefer to have the pawn back on b6. I certainly feel that with White playing moves such as 11.Nd1 b6 can't be a poor move despite it's relative lack of speed.
12.h3?
Martin goes wrong immediately. Maybe he was trying to follow up using ideas with which he was familiar but 12.g4 is possible immediately, so why not get on with it. Black will respond with 12...f5 to be followed by Bb7, Qd7 and Rbe8 as necessary putting more pressure on the White centre.
Another decent alternative would have been to play along the lines of 12.Ne1 Bb7 13.c3 Nc6 14.d4 cxd4 15.cxd4 d5 16.e5 when Black is equal but probably no better.
12...f5 13.Nh2 e5 14.c3 Ne6 15.Nf2
15...d5?
I was trying for too much from the position and ended up throwing away most of the advantage. Instead 15...exf4 16.Bxf4 Nxf4 17.Qxf4 fxe4 18.Qxe4 d5 gives Black a slight advantage because of the Bishop pair.
The best move however was 15...Bb7 after which the White centre is close to collapsing while Black has nothing to stop him adding more pressure with ...Qd7 and ...Rbe8 as necessary. For example 16.fxe5 Bxe5 leaves White with problems over his g3 pawn because 17.g4 f4 wins a piece for Black while 17.Nf3 Bxg3 18.Nh1 fxe4 19.dxe4 Bf4 leaves White with more weaknesses than Black's d6 pawn can compensate for. Meanwhile after 16.Nf3 exf4 17.gxf4 Qd7 black piles on the pressure and it is not clear how White is going to resist.
This is an excellent example of why it is a good idea to include all your pieces in an attack whenever possible.
16.fxe5 Bxe5 17.Bh6 Rf7
The exchange sacrifice 17...Bxg3 18.Bxf8 Nxf8 provides Black with quite a lot of compensation for a minimal material investment but can hardly be justified from a practical viewpoint.
18.g4! f4
Also quite reasonable was 18...dxe4 19.dxe4 Qxd2 20.Bxd2 f4 with a nice endgame for Black however with only an advantage on the board it is probably better to keep as many pieces as possible on the board for the moment.
I was not entirely happy about the move 18...f4 because though it does gain a strong pawn on f4 it also shuts down most of my play but it does have one major advantage. White is forced to undertake something here or simply be left with a bishop shut out of play at h6. Another alternative was 18...fxe4 19.dxe4 Bf4 20.Bxf4 Nxf4 21.Bh1 but here I thought Black has no tangible advantages and so it is not clear what to play for.
A friend has pointed out to me that activity counts for something and that is exactly what Black
has here, a whole pile of activity. He took this last variation a little further and discovered that my intuition was right. I should have been suspicious of the move f4. It just didn't feel like my advantage was being increased by it. His variations 18...fxe4 19.dxe4 Bf4 20.Bxf4 Nxf4 21.Bh1 Ba6! 22.Re1 dxe4 23.Qxd8+ Rxd8 24.Bxe4 Rd2 25.Rb1 Ne2+ 26.Kg2 Rxf2+ 27.Kxf2 Nxc3+ and Black comes out a pawn ahead. The alternative 22.Rd1 Ne2+ 23.Kg2 Qc7! 24.Nf1 Rbf8 seems to be winning as well.
I was in fact looking at the start of this variation during the game but didn't look further than the move 21.Bh1 which is not nearly far enough to decide between this and the game move. Of course with this new information in mind it is possible to improve the defence for White as well.
Instead of 21.Bh1 White should play the bettern 21.Bf3 when the bishop covers the knight check on e2 so 21...Ba6 22.Rfd1. The white pieces are still precariously placed but I don't see a way to increase the activity ratio of the Black to White pieces so it looks like White will consolidate eventually.
19.Nf3 Bc7 20.d4!
Definitely not 20.Ng5 Nxg5 21.Bxg5 f3 22.Bh1 Qd6 and the weakness of the White king position rears it's ugly head suddenly. If White does nothing however Black will play d4 with the simple follow up Nc6 and g5 so White must act sharply.
20...cxd4 21.cxd4 dxe4 22.Nxe4 Nd5
23.Nfg5 Nxg5 24.Bxg5
24.Nxg5 Re7 25.Ne4 Qd7 was an equally decent alternative. It is clear that the knight sitting on d5 and the threats produced by ...Ne3 will eventually provide Black with an advantage so Black will simply look to quell the White initiative waiting for that moment to arrive. It should be mentioned that I was on the verge of time-trouble here with my opponent having 40+ minutes still on his clock. I had been using my time and he had not been thinking too long to this point in the game.
24...Qd7 25.Rae1?!
White needs to increase or maintain the initiative to stay in the game at the moment and this move doesn't create any more threats but allows Black to complete his development. Better was 25.Nc3 Ne3 26.d5! Bb7! 27.Be4 and the game remains highly unclear and interesting.
25...Bb7
Finally I develop the light squared bishop. Note that 25...Ne3 26.Nf6+ Rxf6 27.Bxf6 Nxf1 is not a very good exchange as I lose my blockader on d5. I noticed here that there was virtually no danger of White setting up a fork (with Nf6+) because in order to oust my knight on d5 he must first move his own from e4.
26.Rf2 Kg7 27.Bh4 Ne3 28.Nf6 Qe6
There are many traps around here.
28...Qd6?? 29.Bxb7 Rxb7 30.Ne8+ wins the queen.
28...Qd8?? 29.Nh5+ wins the queen.
28...Rxf6?? 29.Bxf6+ Kxf6 30.Rxe3 wins the exchange.
29.d5 Qd6 30.Qd4 Qe5 31.Qxe5 Bxe5 32.g5
Throwing away lots of the advantage. If I had more time here I might have acted more calmly and realised that the knight on e3 and weakness of the white isolated pawn are not temporary factors. Black can take some time and play 32...Rd8 when 33.Rd2 Bd6 34.Ne4 Bb4 35.Nc3 Nc4 is a good indication of how difficult the White position is.
33.Rd2?!
Much better was 33.Rxf4 Nc2+ 34.Rxd4 Nxd4 when Black has the exchange but for a pawn less than during the game. I got the impression that Martin missed by reply here.
33...Nc2+ 34.Rxd4 Nxd4 35.Bf2 Nf5 36.Be4 Nd6?
Here it was a good time to relieve the pressure by returning the exchange. 36...Rxf6 37.gxf6+ Kxf6 38.Bxf5 Kxf5 though there are basically no winning chances provided by Black's extra pawn here.
37.Bd4 Kf8 38.Bf3 Rd8 39.Bg4 Ne8 40.Be6 Re7 41.Bc3 a5 42.b4 Nc7
43.Re5?
43.bxa5 Nxd5 44.Nxd5 Bxd5 45.Bb4! is winning for White so Black is forced to play 44...Rxe6 45.Rxe6 Bxd5 46.Rf6+ Kg8 47.axb6 with a very difficult endgame, probably losing for Black.
43...Nxe6 44.dxe6 Rd3 45.Nd7+ Ke8 46.Nf6+ Kd8 47.Be1 Re3
Exchanging White's last rook leaves him with no hope of saving the game. Notice how the un-opposed light squared bishop does multiple 'defensive' chores. It stops both the White queenside passed pawns and limits the knight's choice of movement. The opposite colored bishops actually make the exchange much easier for Black to realise here.
48.Rxe3 fxe3 49.bxa5 Rxe6 50.axb6 Ke7
Of course not 50...Rxb6 51.Ba5
51.Bb4+
The move 51.Ba5 is no better because of 51...Re5 winning the pawn on g5.
51...Kf7 52.Ng4 Rxb6 53.a3 Re6 54.Nh6+ Ke8 55.Ng4 Ba6 56.Nf6+ Kd8 57.Nxh7 Rc6 58.Nf6 Rc1+ 59.Kg2 e2 60.Ba5+ Kc8 61.Nd5 Bb7
This is even more clear than the alternative 61...e1=Q 62.Bxe1 Rxe1 63.Nf4 Re5 64.h4 and White captures Black's last pawn.
62.Kf2 Bxd5 63.Kxe2 Rh1 0-1
Lessons from this game:
a) If your attack is not quite there yet and one piece is not participating then it is a good idea to find a way to introduce that piece as well. It doesn't really matter what the target of the attack is (queenside, kingside, attack on the centre). This is of course a variation on the rule, 'Include all your pieces in the attack'. Try giving Black the move 15...Bb7 and then playing through the variations which follow to see multiple examples of this being true.
b) Unless there is some way to drive them away well placed pieces constitute a significant and durable advantage. During this game I under-estimated and then over-estimated my position with the moves 32...Bd4 and 36...Nd6 because I failed to appreciate this. There was no need to rush in the first position because of these advantages and I should have acted immediately in the second because of White's increasing threats.
c) A well placed and un-challenged bishop can perform several duties at once. Here the light squared bishop acted as both defender and attacker in the endgame making it extremely easy for me to realise my exchange advantage (after 47...Re3). In this particular situation my rook was free to roam the board creating threats all over because the bishop easily handled all the defensive duties leaving the rook with none.