This short game I believe is of some interest to all correspondence players as well as those that play the King's Indian. It is a classic demonstration of the fact that correspondence chess is not perfect chess but the errors in it are of a different nature than those OTB.
1.d4
Nf6
2.c4
g6
3.Nc3
Bg7
4.e4
d6
5.Nf3
0-0
6.h3
e5
7.d5
Na6
8.Bg5
h6
9.Be3
Nc5
10.Nd2
a5
11.Be2
Bd7
12.Rb1
This deviates from my two books on the King's Indian by Panczyk and by Joe Gallagher but I followed Gallagher's advice and tried to get in the early thematic f5. Hence
12...Nh7
13.0-0
f5
14.exf5
gxf5
15.Qc2
playing the King's Indian in a serious game for the first time I wanted something spicey and was already considering a pawn sac for a king's side attack. The thinking behind this was play something sharp, test the sac against best analysis, which correspondence chess should be about, and finally learn the basics of a new defence. So e4, but in retrospect Nf6 would have been more sensible.
15...e4
16.f3
f4
17.Bxc5
dxc5
18.fxe4?!
This move is probably an error in that a knight should be on this square. Stuart Knox, the expert on this system in the Andover Club was rather scathing about this move; I find it understandable in that white wants to clear the f3 square for a knight. Admittedly there is overload of possible white piece moves on the f3 square which was a large part of the idea of the sac.
18...Qg5
quite basic but still needs answering.
19.Kh1!
A nice simple side stepping move which cuts across all the sacs on h3
19...Ra6!
I really congratulated myself on this obvious move bringing another heavy piece into the attack and had visions of a sparkling win in about 10 more moves
20.Bd3
This stops the obvious Rg6, so by a process of elimination, improve your pieces, it was
20...Nf6
21.Nf3
Qh5
22.Kg1!
he spotted the obvious N manouevre to land on e3 and now he is threatening e5 in a lot of lines. So backwards
22...Ne8
23.Rf2?!
?! But he decides to go slow, gang up to win the f pawn after which black is surely lost. e5 with best play after about 30 more moves is a win for white or at best a draw for black. So play active with
23...Kh8!
24.Rbf1
Bxh3
Here I thought e5 and I should be able to squirm my way towards a draw but he took
25.gxh3
Rg6+
26.Rg2
[26.Kh1
]
26...Rxg2+
27.Kxg2
Rg8
now I thought he was busted but I had missed his response, the simple
28.Ne2!
Oh dear! Well play on a few moves, keep attacking, and resign when the f pawn is gone
28...Be5+
I played this rather than the more obvious Bd4; this is only good if he attempts to escape across to the queen side but Kh2 stops all that nonsense
29.Kh2
Nf6
Bluff, expecting 30. NxBe5 f3 31.Rxf3 QxNe5+ 32. Nf4 Nh5 33. Qc3! And resign but my lucky day. I sent my postcard off with the following three move forcing sequence
30.Nxf4
Bxf4+
31.Kh1
Qxh3+
32.Nh2
Nh5
It is white to move.33. RxBf4 loses to Ng3+
34. Kg1 Ne2+
35. Kf2 QxNh2
36. Ke1 QxRf4
37. any capure of the knight on e2 loses
33. Qc3+ is countered by Rg7
Whilst
33. Rg1 or Rf3 loses to Ng3+
34. RxN QxR
and black threatens mate in one on three squares.
My opponent resigned and in his letter he wrote "I accidentally played 28 Bd4+ on my board when you actually played Be5 and of course, now that I've corrected the error my position is lost."
I can fully understand how my opponent did this; he was anticipating what he thought was the stronger double check move so only gave a cursory look at my postcard. I made another similar correspondence mistake a couple of seasons back by playing a knight sacrifice in a won position without even checking it on the board first. Given that a correspondence match can last 9 months, as this one has done one would expect near perfection. Who said in chess it is the last mistake that counts the most?
33.Rxf4
[33.Qc3+
Rg7
; 33.Rg1
Ng3+
; 33.Rf3
Ng3+
]
33...Ng3+
34.Kg1
Ne2+
35.Kf2
Qxh2+
36.Ke1
Qxf4
37.Kxe2
Rg2+
38.Kd1
0-1