The Kibitzer 
by Tim Harding

A Difficult Postal Game

This month's Kibitzer is devoted to just one game that was crucial
for my achieving the International Master title, awarded to me at
this month's congress of the International Correspondence Chess
Federation, ICCF. There are several points about the course of this
game, which took three whole years to complete, that particularly
well illustrate the difference between over-the-board and
correspondence play.

The Olympiad 12 preliminaries which began five years ago. My
Romanian opponent had no published rating but I could be sure
that any player who made that country's national team would be
master strength. Last year he appeared on the rating list for the first
time at 2415 but he has now slipped back. The latter stages of this
game perhaps show why.

Tim Harding (Ireland)- Adrian Parau (Romania) 12th CC
Olympiad preliminaries, board 4, 1992-95

1 e4 c5 

Of my six games with White in this event, two had already begun
1 c5 2 f4 so I decided to have some variety.

2 Nf3 d6 3 Bb5+ Nc6 4 0 0 Bd7 5 Re1 a6 6 Bxc6 Bxc6 7 c3 Nf6 8
d4 Bxe4 9 Bg5 

White gambits his e-pawn to open a central file and build up
pressure while Black is underdeveloped. I had studied this quite a
bit and thought it a difficult line about which to form firm
conclusions. Does White really have enough? 

In an OTB game, the deciding factors would firstly be memory of
known examples, then judgment and the willingness to take risks
against the clock. In CC, the important thing is to research the
opening variation thoroughly, base critical analysis on that
research, and maintain an objectively sound position at all times.
Then the more complicated it gets, the more likely the better
analyst is to win.

With access to a good library and databases, I thought I was
probably in a better position to do research on the line than a player
in Romania so it seemed a good choice for postal play. My
opponent chooses the main reply.

Here are a couple of examples of how Black can go wrong:

a) 9 d5 10 dxc5! e6 11 Bxf6! gxf6 (11...Qxf6? 12 Qa4+) 12 Nbd2
Bxf3?! 13 Qxf3 Bxc5 14 c4! dxc4 15 Rad1! Qe7 16 Ne4 f5
(16...0-0? 17. Rd7+-) 17 Qc3! fxe4 18 Qxh8+ Qf8 19 Qxh7 e3 20.
fxe3 Qe7 21 b3 Bb4 22 Rf1 c3 23 a3 Bxa3 24 Qh8+ Qf8 25 Qxc3
1-0 S.Arkhipov- M. Muhutdinov, Nabereznye Chelny 1993. Going
for a closed centre can be just as bad:

b) 9 Bxb1 10 Rxb1 e6 11 Bxf6 gxf6 12 d5 e5 13 Qd3 Qd7 14 Nh4
Qg4 15 g3 h5 16 Re4 Qd7 17 Qf3 Be7 18 Nf5 (J.Barle-Quinteros,
Lone Pine 1979); the good knight versus bad bishop gives
tremendous compensation for Black's extra doubled pawn.

9...Bd5 10 Nbd2 b5 

Here 10 e6 was more normal, while White had done well in
games where 10 cxd4 was played. In the few games where 10 b5
had previously been played, White had opted for 11 c4 e.g.
11...bxc4 12 dxc5 e6 13 Rc1 (or 13 Bxf6 gxf6 14 cxd6 Bxd6 15
Ne4 unclear in Veroeci-Ghinda, Brno 1989) 13...dxc5 14 Ne5 Be7
15 Qa4+ Kf8 16 Nexc4 Ra7 17 Be3 Qe8 unclear
(Ostrovsky-Chechelian, USSR 1979).

11 b3 

New; now after 7 days thought he opted to transpose to the 10...c4
11 b3 b5 line. If 11...cxd4 I thought that 12 c4!? looked interesting,
while if 11...e6 my analysis went 12 c4 Bc6 13 Ne4!? Be7 14 d5
Bd7 (14...Nxe4 15 Bxe7 unclear) 15 Bxf6 gxf6 16 dxe6 Bxe6 17
Qe2 bxc4 18 bxc4 d5? 19 cxd5 Qxd5 20 Qb2 Qf5 21 Qb7 with an
attack.

11...c4 12 bxc4 bxc4 13 Nh4

An alternative is 13 Nf1 as in Fette-Ruban, Miskolc/Tapolca 1990,
which was featured in a 1991 database survey in 'New In Chess
Yearbook 21'. I spent four days on this choice.

13...Qd7 

If 13...e6 14 Nf5 (Cifuentes) is unclear and gives good scope for
the CC analyst.

14 Rb1 Ng8!

This paradoxical retreat was probably first played in this game and
it casts serious doubt on the variation for White. Leaving his
Bishop dominant, Black threatens to drive back my minor pieces.
The example I knew was Ricardi-Cifuentes Parada, Buenos Aires
1991, which went 14 h6 15 Bxf6 gxf6 16 Qh5 Qc6 (16...Be6 17
Qf3 Rc8 18 d5 Bg4 19 Qxf6) 17 Rb6! Qxb6 18 Qxd5 Rd8
(18...Rc8 19 Nxc4 Qc6 20 Nxd6+ Kd7 21 Qxf7! Qxd6 22 Ng6+-)
19 Nf5 Qb5 20 Nxd6+ Rxd6 21 Qxd6 e6 22 Qc7 Qd7 23 Qb8+
Qd8 24 Qxd8+ Kxd8 25 Nxc4 and White won.

After 14 Ng8 I had my longest think of the game: 11 days. As I
recall, I recognised that it was no longer possible to make anything
happen on the kingside and that therefore I must use my only asset,
the b-file, and try to recover the gambit pawn. (See Diagram)

15 Rb4!

A game that began later but finished earlier was published in
"Correspondence Chess Yearbook 11". J.Elburg-R. Hendriks, 22nd
Dutch CC Ch 1993-4, went instead 15 Qe2 h6 16 Be3 g5 17 Nhf3
Rc8 18 Rb6 f5 19 Nf1 f4 20 Bd2 Qf5? 21 Rxd6 Bg7 22 Rxa6 h5
23 Ra7 and White eventually won, but according to John Elburg,
20...Bg7 would have given Black some advantage.

15...h6 

If 15...a5 I planned an exchange sacrifice by 16 Rxc4 Bxc4 17
Nxc4 Qc6 18 Qd3 with compensation since he is very far behind
on development, e.g. 18 h6 19 Bf4 e6 20 Nf5.

16 Be3 Rc8

I also spent a lot of time on 16...g5 17 Nhf3.

17 Qa4 

The threats to c4 and a6 regain the pawn. The alternative was 17
Qe2 when the exchange sacrifice is still in the air but there is
nothing clear: more suited to OTB time-scrambles than postal.

17...Qxa4

If 17...e6 I intended 18 Reb1!? 

18 Rxa4 g5

If 18 Rc6 then 19 Rb1 seems to regain the material.

19 Nf5 e6 20 Ng3 f5 21 f4 Be7 22 Rxa6 

Material is level again but the position is unclear at best for White.
He has the bishop pair, my bishop is very bad and my passed
a-pawn almost worthless. I offered a draw but he ignored it. 

22...Nf6!? (See Diagram)

This gives a very doubled-edged position requiring deep analysis
before reply but I still managed to answer in three days so I must
have been prepared for this. Essentially the boot is now on the
other foot: Black is the one sacrificing a pawn for the initiative and
White must be very careful. The questions now are not about
pawns but who will achieve coordination of his pieces and, later,
whose king is in the more danger?

He was now moving much slower than me, and the fairly slow post
to and from Romania meant I could do a lot of work on the game
without spending "clock" time. White would be more likely to go
wrong in an email game.

23 fxg5 

I decided there was nothing better.

23...Ng4 24 Bf4 hxg5 25 Bxd6 Bxd6 

If 25...Rxh2 26 Nf3 Bxf3 27 gxf3 Bxd6 28 Rxd6 Rh3 29 Nxf5
Black has nothing clear. Instead he will keep pieces on, castle, and
seek to invade my position down the queenside files that seemed to
be in my possession earlier on. Against that, the frontal assault on
my kingside is over and I have chances to infiltrate there.

26 Rxd6 0-0 27 Nf3 Ra8 28 Nxg5 

This was a hard decision but I had my holidays at this time. It was
now August 1994 and the game had been going for 21 months. I
thought this more active than 28 Re2 f4! Which gives Black
winning chances. I had to avoid 28 h3? Rxa2! 29 Re2 (29 hxg4?
fxg4 30 Nxg5 Rff2 +) 29...Ra1+ 30 Re1 Rxe1+ 31 Nxe1 Ne3 32
Rd7 f4 33 Ne2 Rb8 34 Kf2 Nd1+ 35 Kf1 Rb2 36 Rc7 Be4 37 g3
Ne3+ 38 Kf2 Bd3 +.

28...Rxa2 29 Ne2 

This move cost me 10 days, mostly spent convincing myself that
29 Re2?! Rxe2 30 Nxe2 Ra8!? Gave him too much counterplay.

29...Rb8!? 

I was very lucky here. My notes show that when deciding on my
28th, I seem to have overlooked this move, threatening 30 Rxe2.
Although 29 Rb8 is his most dangerous try, it also involves risk
for him. I had concentrated on 29 Ne3 30 Nxe6 Nxg2 31 Rxd5
Nxe1 32 Nxf8 Rxe2 33 Rxf5, leading to a draw, and 29 Rfa8
when there is a choice between 30 Nf4!? and 30 Nf3. (See Diagram)

30 Nf4

Neither 30 h3 nor 30 Nxe6 Rxe2 31 Rd8+ Kf7 is playable, so my
reply is forced and in general my choices became easier from now
on, exactly where they would probably have been too hard in an
over-the-board game!

OTB, the clock and nerves would probably have been major
factors and experience shows that the player with the initiative
normally comes out on top in such situations. Playing by post,
however, I was able to find a line of defence that held and make
detailed notes to which I could refer when my opponent's
subsequent moves arrived, at roughly monthly intervals. 

30...Ne3 

If 30...Rbb2 31 Rd8+ Kg7 32 Ngxe6+ White has at least a draw.

31 g3! Ng2!? 

If 31...Rbb2 32 Rd8+ Kg7 33 Ngxe6+ Bxe6 34 Nxe6+ Kf7 35 Nf4
Rc2 36 Rd7+ Kf8 37 Rb1 and he has to allow a pair of rooks to be
exchanged.

32 Re2 

Probably the only move: if 32 Nxg2?! Rxg2+ 33 Kf1 Rxh2! with
strong mating threats. However, I was no longer worried by this
game and was virtually playing it on autopilot. Since he was
committed to a definite line of attack and I had been unable to find
a winning line for him, there were fewer alternatives to consider. I
don't give all the possible variations below but they were in my
notebook in case I needed them.

The main line seemed to peter out into a an ending that he could
probably draw, but no more than that. Black took eight days on move
31, six days on move 32, six days on move 34: all encouraging signs
that he could not force a breakthrough.

32...Rxe2 33 Nxe2 Ne3 

Threatening 34...Rb1+ 35 Kf2 Ng4 mate so I need a flight square
on e2. 

34 Nf4 Rb1+ 35 Kf2 Nd1+? (See Diagram)

This position, which I had foreseen when playing my 30th move, is
the point where he should have realised his attack had failed. It is
the ultimate turning point in a game where the advantage seemed
to ebb and flow dramatically.

The safest line for him is probably 35...Ng4+ 36 Ke2 Rb2+ when I
have not found a win for White. A possible continuation is 37 Ke1
Nxh2 38 Nxd5!? exd5 39 Rxd5 Rc2 40 Rxf5 Rxc3 41 Rc5 Rxg3
followed by him, if necessary, sacrificing his knight for my last
pawn to force a book draw.

36 Ke2 Nxc3+ 37 Kd2 Ne4+

Simplifying to an endgame where he stands worse, as I shall
explain later.

37...Na2!? was the move I half-expected, trying to shepherd his
Pawn forward to create counterplay. However, after 38 Nh5!! c3+
39 Kd3 it seems to fall into a mating net on the kingside. I was
analysing this but had not worked it all out when his 37th move
arrived. 

For example, 39...Kf8 40 Rd7 Rd1+ 41 Ke2 Rd2+ 42 Ke3 Nb4
fails to 43 Nh7+ Kg8 44 Rg7+ (or 44 N5f6+ Kh8 45 Nf8!!)
44 Kh8 45 N7f6. A similar finish follows 39...Rd1+!? 40 Ke2
Rd2+ (40...c2?? 41 Rd8 mate) 41 Ke3 Kf8 42 Rd7 Nb4 43 Nh7+
etc. This explains why he had to exchange a pair of knights. In
those variations his bishop was an idle bystander. This piece,
which seemed so strong in the middle-game, was technically a
"bad" bishop because of the e6- and c4-pawns fixed on its own
colour. Now at last, and in the endgame that follows, this becomes
the deciding factor!

Even so, had Black fully realised the danger he was in, he could
maybe have held on and could certainly have prolonged the
struggle and put more technical difficulties in my way.

38 Nxe4 Bxe4 39 Rxe6 Kf7?! 

The best move, I thought at first, making it hard for my Rook to go
to the c-file. He actually offered a draw here, and also overstepped
the time limit for the first time. (In ICCF games, you have to
overstep the 10-moves-in-30-days limit twice to lose a game on
time.) Maybe, he could still have held the game by 39 Rb3,
however.

40 Re5! Rb3 

Also if 40...Rb2+ 41 Kc3 Rc2+ (41...Rxh2 42 Kxc4 Rc2+ 43 Kb5
Rc3 44 g4 Bc6+ 45 Kb4) 42 Kb4 he has some losing chances. 

41 Rc5 

Obviously White cannot allow ...c3+; now with an extra pawn and
all my pieces better than their opposite numbers, the chances of a
win are becoming apparent.

41...Bd3

This was half-expected. He hopes to bail out into a rook ending
with good drawing chances. Objectively he should have tried
41...Rb2+ 42 Kc3 Rxh2 43 d5 Ra2 although this is unpleasant for
him after what has gone before. 

It was now July 1995. Several of my opponents faltered in the third
and fourth year of play in this long tournament. Patience and
consistency of effort over a long period are part of the qualities
required to achieve mastery in CC. The will to fight on for a draw
after going wrong in a game you expected to win is a difficult
requirement in both forms of chess.

42 d5! Ke7 43 Ke3 Rb2 44 Kd4 1-0

I sent the conditional 44...Rxh2 45 Ke5 but he resigned. His rook
and bishop are ineffectual in the face of my threats to advance the
d-pawn and to mate his king.