The Kibitzer 
by Tim Harding

A la Recherche des Echecs Perdus (or Remembrance of Chess
Times Past)

IT IS only occasionally that I write a Kibitzer column based on
some of my own games. I was inspired to do so by reading a copy
of Journal of a Chess Master by Stephan Gerzadowicz which
somebody lent me recently. It reminded me that it can be beneficial
to and interesting to take a critical new look at correspondence
games that one has played in the past, but the difference in this
case is that the games were played a very long time ago.
Gerzadowicz, for those who have not come across his writings, is
an American postal chess master with an extremely original style,
full of quotations and allusions to his favourite writers, such as
Thoreaux and Shakespeare. (I won't be attempting to imitate that
style here.) His books include the aforementioned Journal of a
Chess Master (now out of print and hard to get hold of) and its
sequel, Journal of a Chess Original, which he certainly is.

I was recently looking through a few of my postal games from the
1970s, which I found a rather strange experience. They were
mostly from the ICCF master class group WT/M/GT31, which
began in late 1973. I don't remember for sure but I guess that most
of my games were over by the end of 1975. In those days, I was a
stronger OTB than postal player.

I had to hunt for the tournament announcement and final crosstable
in the ICCF magazine Fernschach as I have long since forgotten
most of what occurred. I only have a few of the game scores
preserved. I found the start-list in the December 1973 issue of the
magazine. Strangely, there were no Russians which meant the
tournament proceeded fairly quickly and smoothly.

There was one curious incident. One of the players in this event
was Claude Bloodgood, the American prisoner who had written a
monograph on The Tactical Grob. However I had White so he
didn't get a chance to play this. After 1 e4 Nf6 2 e5 I never heard
from him again. Eventually I learned that he had escaped from jail,
and been recaptured, and that was the end of him being allowed to
play postal chess for many years. His games were cancelled. I
found the final crosstable of the event in the September 1977 issue
of Fernschach where I see I ended up in seventh place with 7/13,
the Bloodgood game not counting. Campos Moreno of Spain
scored 0 which probably means he dropped out or went silent some
time after Bloodgood. So of the twelve real games, I won three,
lost three and drew six. 

At this stage I was still learning international postal chess
technique and that may explain why I didn't play and score as well
as I should have done. I had been playing CC keenly for about a
year at this stage and the BPCF had agreed (chiefly thanks to Keith
Richardson) that I should start in ICCF master class. Maybe a
mistake. It might have been better to begin with a Higher Class
event that I could win. That's what Alan Ludgate decided to do
when he started CC in the mid-80s and quickly worked his way up
to a 2500+ rating and a World Championship Three-quarter Final.
This was the first tournament in which I had to use ICCF numeric
notation and international CC postcards, a bit of a distraction.

However, I do have a draw, a loss and a win to show you from that
event. I will analyse the win last and in more detail than the others,
because it is the most interesting of the three and the only one of
which I have any notes at all dating from when it was played.

I begin with the game against Ove Kroll of Denmark, who came
third in the tournament. Twenty years later we met again in CC
Olympiad 12 and this time he won, but we both earned IM titles
from the event. In this game, he surprised me in the Nimzowitsch
Defence, with which I was already dabbling myself when Black. At
some point over the years I lost the score of this game, but Kroll
has sent it to me. 

The funny thing is that I recalled I had played a game around this
time with queen against rook and minor piece, but I thought it was
a different opponent in a different event. Memory can certainly
play tricks over such intervals of time.

Tim Harding - Ove Kroll Nimzowitsch Defence WT/M/GT31
1973-1974

1 e4 Nc6 2 Nc3 

Avoiding the main lines and inviting a Vienna.

2...e6 

A disappointment for me; this move was favoured by Nimzowitsch
himself.

3 d4 Bb4!

Maybe better than playing in French Defence style with 3...d5.

4 Nf3 d6 5 Qd3!?

I decided to vary from 5 Bf4 of Maroczy-Nimzowitsch, San Remo
1930. Already we have a fairly rare situation on the board.

5...Nf6 6 Bg5 Bxc3+

I see that in my 1981 book Nimzowitsch Defence this move is
awarded an exclamation mark: White cannot recapture with the
queen because of...Nxe4 which means my fifth was maybe a waste
of time, or was it?

7 bxc3 e5 

My book says "7...h6 (idea...e5) =" but I wonder if I was being
pessimistic about White's chances..

8 Be2 h6 9 Bh4 Bg4 

Maybe 9...0-0 is better.

10 Rb1 b6 11 h3 Bxf3 12 Qxf3 0-0 13 Bb5 Nb8 (See Diagram)

Looking at this game today, the opening seems to be a success for
White. He has the bishop pair against the knight pair, a pin at f6
and the black development has just gone retrograde. 

14 Rd1 Qe7 

Not 14...g5?? 15 dxe5.

15 Bxf6 

I suppose I worried about...g5 shutting out the bishop. This does
not seem necessary although the remaining B is clearly superior to
the N and structural weaknesses have been imposed on Black.

15...Qxf6 16 Qxf6 gxf6 17 0-0 a6 18 Be2 Re8 19 f4 Nd7 20 Bc4
Re7 21 Rd3 b5 22 Bd5 Rf8 (See Diagram)

White now decides to go for the win of a pawn - wrong decision? 

23 fxe5 fxe5 24 Bb7 Nb8 25 d5 Nd7 26 Rg3+ Kh8 27 Bxa6 Nc5 

27...Ra8 would get rid of my a-pawn but then he wouldn't pick off
my e-pawn, and I would be left a doubled c-pawn ahead.

28 Bxb5 Nxe4 29 Re3 Nc5 30 Bc6 

The bishop looks silly here but of course the idea is to support the
advance of my outside passed pawn. 

30...f5 31 a4 f4 32 c4 Rg8 

Black gives up the f-pawn to speed up his counterplay. 

33 Rxf4 Reg7 34 Rf2 Rb8 35 Ra3 Rb1+ 36 Kh2 Re1 37 a5 Ne4
38 Rf8+ Kh7 39 g4 h5 (See Diagram)

Things are certainly hotting up. White is close to queening his a-
pawn, but 40 gxh5? allows an immediate draw so White must do
something about Black's counterplay on the kingside where I am
rather short of defenders. I think I must have decided that 40 a6
was too risky, as after 40...hxg4 41 a7 g3+ 42 Kg2 Re2+ 43 Kg1
g2 (43...Re1+ 44 Rf1) 44 Ra1 Ng5 Black threatens to win (45
Rh8+ Kxh8 46 a8Q+ Kh7 47 Qa3 e4-+). White would have to play
the deflecting move 45 Bd7 anyway but after 45...Rxd7 46 Ra3
Re1+ (or first Ne4 soon transposing) 47 Kxg2 Ne4 48 Rg3 Ra1
leads to an ending with few winning chances. However, this is
fairly hasty 2000 AD analysis; I don't have any of my notes from
the time the game was played. Anyway I came up with a problem-
like move which the computers of today agree is best

40 Bd7! White gives up the bishop for a crucial pair of tempi, to
get the black rook off the g-file.

40...Rxd7 41 a6 Kg7 42 a7 Kxf8 43 a8Q+ Kg7 44 Rf3

So White has won queen (and pawn) for rook and knight and ought
to be winning. In a minute we shall see if we can find where I went
wrong. 

44...Rf7 

He cannot allow Qf8+ so further simplification is forced. 

45 Rxf7+ Kxf7 46 Qc8 hxg4 47 hxg4 

You would think that, having a passed pawn, and with c7 about to
drop off, that the win should not be too hard. 

47...Kg6 

He makes it as hard as he can for my king to advance. (See
Diagram)

48 Qxc7? 

This move seems to be the culprit - taking the doomed pawn is not
obligatory. It is much more important to prevent my king being
driven to a bad place. Even then, it seems a difficult technical task
to win this position and I was not equal to it. I now think that 48
Kg2! might win. Black's choice is between 48...Re3 and 48...Nf6.
Let us just look at one line where the defence can make life hard.
48...Re3 49 c5 Rg3+ 50 Kf1 Nxc5 51 Qf5+ (51 Qxc7 Ne4)
51...Kg7 52 g5 and now the passed pawn is getting dangerous.
However, Black may give up his knight in the attempt to create a
fortress52...Ne4 53 Qxe4 Rxg5 54 Ke2 (54 Qc4 Rf5+) 54...Kg8 55
Qh4 Rg7. Now if Black's rook were on c5 with his king around b7
this would be a clear draw. White has to prevent that, improve his
king position and eventually play c4-c5 in the attempt to achieve
Q+P v R, but I don't know if this can be done. Any expert
opinions?

48...Re2+ 49 Kg1 Rxc2 50 Qc8 Rc1+ 51 Kg2 Rc2+ 52 Kh3 Kg7
53 g5?!

Of course not 53 Kh4?? Rh2 mate.

I must have decided that checks led nowhere but in an OTB game
(with no postage and time factors to consider) I would probably
have tried a few things before giving the pawn away. The player
with the advantage operates on the principle "do not hurry", tries
lots of long checking sequences and king manoeuvres and
eventually the defender is apt to go wrong. In a postal game, these
sort of options are not really available as stamps cost money,
weeks turn into years and the game gets called for adjudication. In
an email game, however, one could approach this situation a bit
more like an OTB game because the response time is quicker and
the moves don't cost money.

53...Nxg5+ 54 Kg4 Nf7 55 Kf5 Rf2+ 56 Ke6 Rf6+ 57 Ke7 Nh6 

So the king march has achieved nothing and I no longer have my
passed pawn. 

58 Qh3 Nf5+ 59 Kd7 Rf7+ 60 Kc6 Rf6 61 Qh2 Kg6 62 c5 

This is the only thing that is left to try. 

62...dxc5+ 

Otherwise I will eventually get in c5-c6 with much improved
winning chances. 

63 Kxc5 e4 64 Qg2+ Kh5 65 Qxe4 (See Diagram)

Now it has all reduced to queen and pawn versus rook and knight
but there are zero winning chances, even if your computer tells you
White is +2. The passed pawn cannot get across the doubly
guarded square d6, and Black's compact arrangement of pieces
means he can maintain this defence. 

65...Kg5 66 Qe5 Ra6 67 Qc7 Rh6 68 Qc8 Rg6 69 Qd8+ Kh5 70
Kc4 Rd6 71 Qg8 Rg6 72 Qh7+ Kg5 73 Kd3 Rh6 

So I decided it was time to call it a day. 

74 Qxh6+ -.

I lost three games in the tournament. One was to American player
Alan Benson, and is published in the book Correspondence Chess
in America by Bryce Avery. I think Benson played rather well. I
also lost to another American named Stone, and the following
game to a German player who was in CC World Championship
Final XIV a few years ago.

Hans Ziewitz-Tim Harding Dutch Defence

1 d4 f5 2 g3 g6 3 Bg2 Bg7 4 Nf3 Nh6 

This eccentric development of the knight is the Basman Variation
of the Leningrad Dutch which I played in a few over-the-board
games but never again in postal.

5 0-0 d6 6 c4 c6 7 Nc3 0-0 8 Bf4 

I also met 8 e4 Na6 9 Re1 Nf7 10 Be3 Kh8 11 Rc1 fxe4 12 Nxe4
Qa5!? (Basman's idea was to play...d5 in such positions, fixing d4.)
13 Nc3 Qh5 14 Ne2? e5 15 dxe5 dxe5 16 Nd2 Ng5! and I won
quite a nice game in N.J.Holloway-Harding, Cambridgeshir. v
Kent, January 1974 (see Informator 17 for the rest). I think this
postal game was hardly begun when that game was played.

8...Nf7 9 e4 Na6 10 Re1 e5 11 Be3 (See Diagram)

11...fxe4? 

This blunder makes the game a good example of how NOT to play
the Leningrad Dutch. Basman later suggested to me 11...f4!? 12
gxf4 exf4 13 Bxf4 Bg4 with interesting play for the pawn, e.g. 14
Qd2 Qd7! with not unfavourable complications. Also 11...Be6 and
11...Kh8 come into consideration.

12 Nxe4 Bf5

If 12...d5 13 cxd5 cxd5 14 Neg5 e4 15 Nxf7 Rxf7 16 Ne5 with
clear advantage to White.

13 Nh4! 

Not 13 Neg5 Nb4!.

13...Bxe4 14 Bxe4 d5 15 cxd5 cxd5 16 Bg2 e4 

Afterwards my opponent suggested 16...Nb4!? here.

17 f3 exf3 

White is much better already. If now 17...Qd7?! 18 fxe4 dxe4 19
Bxe4 Rad8 20 Nf3 Qg4 21 Bxb7.

18 Nxf3! Nc7 19 Qb3 b6

19...b5 is met by 20 Bf4.

20 Rac1 Re8 21 Bf4 Ne6 22 Ne5 Nxd4 23 Qxd5! Qxd5 24 Bxd5
Bxe5 (See Diagram)

25 Bxe5! Rxe5 

There is no choice because if 25...Rad8? 26 Bxd4 Rxe1+ 27 Rxe1
Rxd5 28 Re8 mate. I guess I had overlooked this earlier on.

26 Rxe5 

Not 26 Bxa8? Ne2+ playing on for a while, but now if 26...Rb8 27
Bxf7+. So I resigned1-0.

Now we come to the best of my wins: an up and down struggle but
with many interesting turns. I don't claim to have the absolute truth
on this game by any means. Again, I don't have my notes on the
middle game.

Harding,T - Karafiath,D WT/M/GT31 1973-75

1 e4 e5 2 Bc4 

I had been writing a book about this (then almost forgotten)
opening and had won a lot of games with it OTB in 1971-73 but
this was my first chance to play it in a high-level postal event. My
opponent was an experienced Hungarian CC player who beat me
twice in subsequent events. Surprise plays only a small role in the
CC opening, because the clock is not ticking. The opponent can put
the set away and do some research, think about it and come back to
the position refreshed another day. In one of the books by Thomas
Love Peacock, a fashionable gardener describes one of the qualities
of his new design as Surprise. The answer comes back (this is NOT
verbatim) something along these lines: "And what pray do you call
this quality of surprise the second time somebody walks around
your garden?" So in CC, it is not enough that a move be
unexpected: it must also be able to stand up for itself. 

2...Nf6 3 Nc3

GM Bent Larsen had played some games with the Bishop's
Opening but he always chose 3 d3. I experimented with various
moves here as White,. including 3 Qe2 and 3 d4, but usually I
played the knight move which invites Black to play the
Frankenstein-Dracula Variation, 3...Nxe4. My opponent demurs. 

3...Bc5 

A straight-forward move, but one that already restricts Black's
options compared with 3...Nc6. The rule "knights before bishops"
applies here. Then after 4 d3 Black could still play...Bc5 but also
has the options of 4...Na5 and 4...Bb4 which are two of the most
effective lines. When Black plays...Bc5 then a King's Gambit
Declined position can often arise, and I sometimes played the
Bishop's Opening that way. However, there are some lines with an
early...Bc5 in which...Nc6 is deferred and that is what happens in
this game. 

4 d3 

White is definitely not interested in playing 4 Nf3 and 4 f4 will
either lead to something sharp (if Black takes up the challenge) or
the King's Gambit Declined if he does not. 

4...d6 

After 4...Nc6 White has 5 Bg5, 5 Be3 or 5 f4 heading for the KGD.
With the black knight still at home, the next move becomes an
option because Black cannot harass the queen and queenside
by...Nd4. 

5 Qf3!? (See Diagram)

White gives up the characteristic Bishop's Opening idea of
advancing the f-pawn early on and instead tries to build up some
pressure against f6/f7. Note that after plays Bg5 the pin cannot be
broken by ...Be7 because Black has played ...d6.White isn't trying
to smash his opponent up quickly in this line but to force the
concession ...h6 which weakened Black's kingside and costs him a
development tempo. 

5...h6 

I have seen one game with 5...c6, but it looks artificial and slow.
So Black takes a tempo to stop Bg5a moral victory for White but
he has no choice. I was following a game by Larsen (see below)
which went 5...Nc6 6 Nge2 (to rule out...Nd4) 6...h6 reaching the
same position as occurs in a moment. In his book that came out in
Denmark in 1969 and England in 1970, Larsen gave a few of his
games with the Bishop's Opening but not this one. By the way,
instead of the "selected games" title of those editions, I think the
German publisher came out with a much better title for his edition
- Ich Spiel Auf Sieg ("I Play to Win"). That exactly expresses the
Larsen philosophy. 

After 5...Nc6 6 Nge2 Black could play 6...Na5 but Larsen's games
in other lines of the Bishop's Opening tended to show that the
exchanges after 7 Bb3 Nxb3 8 axb3 could work to White's
advantage: this structure soon arises in my own game.

I have found a couple of games in my database where Black
refused to play...h6. For example, 5...Nc6 6 Nge2 Bg4 7 Qg3 h5
(Now we see what can happen when the pin is allowed) 8 Bg5 Na5
9 Bb3 Nxb3 10 axb3 c6 11 Qh4 Qe7 12 h3 Bd7 13 f4 0-0-0 14 Rf1
Rde8 15 0-0-0 exf4 16 d4 Bb6 17 Nxf4 Bd8 18 Rde1 Qf8 19 e5!
dxe5 20 dxe5 Nd5 21 Ncxd5 Bxg5 22 Qxg5 cxd5 23 Ng6 and
White won the exchange (and ultimately the game) in Rogers-
Martinovsky, Hawaii 1998.

6 Nge2 Nc6 7 Be3

Denied g5, the bishop naturally goes here. 

7...Bg4 

This may be better than what Larsen's opponent played. Black
develops with tempo, though the queen is quite happy to go to g3
and the bishop may not be on its best square (could be e6 is
better?). After 7...Bxe3 8 fxe3 Na5 9 Bb3 Nxb3 10 axb3 c6 11 0-0
0-0 12 Ng3 d5 13 exd5 cxd5 14 d4 exd4 15 exd4 Be6 16 Rae1
White had all the active chances. He won quickly in Larsen-Finn
Petersen, Danish Ch, Holstebro 196416...Rc8 17 Re5 b5 18 Qd3 a6
19 Nf5 Rc6 20 h3 Kh8 21 b4 Ng8 22 Qg3 g6 23 Nh4 Kh7 24 Ne2
Rxc2? (A blunder in a very difficult, if not lost position.) 25 Rxe6
1-0.

8 Qg3 Bxe3 9 fxe3

Of course, just like Larsen! In my database I found an ICCF email
game where White recaptured with the queen but that is the wrong
idea. In the slow variations of the Bishop's Opening, where there is
no early tactical melee, White wants open files (especially the a-
and f-files) and he is not worried about doubled pawns. See how
the e-pawn controls e4 and f4, denying his knights good squares
and usually at least one of the doubled pawns can be dissolved
later. 

9...Na5 10 Bb3 Nxb3 11 axb3 0-0 12 0-0 

So now I have achieved the first objective: control of three of the
four central squares and both the half-open files. Black does not
see a future for his bishop so decides to simplify further. He had to
do something as I threatened Rxf6 and if...Qxf6 the Qxg4, or
if...Bxe2?? then Rxh6. Of course it is quite possible that Black
thought he stood better here. A player over-influenced by static
positional elements could suppose that my structural "weaknesses",
the doubled pawns were the main feature of the position. On the
other hand, White does need a plan of action for his "more active"
pieces and I perhaps don't find the right one. 

12...Bxe2 13 Nxe2 (See Diagram)

13...Nh7!

I cannot remember now what I thought almost 30 years ago, but
probably I would have liked to play something like Qh4 and N-g3-
f5, or maybe Rf3, Qf2 and Ng3, so Black takes steps to secure
some counterplay. Probably I did not expect this move and did not
form the right plan. 

14 Nc3?! 

Deciding to work against b5 and d5 though I suppose I could still
have tried to manoeuvre this piece to f5 which is where the knight
really belongs. The problem in that case is what to do with the
queen? I think I was in too much of a hurry to undertake something
on the queenside where I have no advantage, instead of trying to
make progress on the kingside or just improve my position slightly.
If I am careful it is very hard for Black to get his game in order.

Looking at this game in 2000 A.D., I don't like 14 Nc3 as I cannot
see what this knight can accomplish on its own. One possibility is
to play 14 c4 so that ...b5 and ...d5 are harder for him, and if he
wants to play ...Qg5 and ...Qxg3 that is OK by me as I can capture
on g3 with the knight. Another idea is 14 Rf5. Although it blocks
the square where the Knight eventually wants to go, it is an
annoying move for him to meet and does not commit any pawns to
new squares.

14...Qd7 15 b4!? a6 

He has to play this otherwise his rook cannot move, but now I have
the potential lever b4-b5 to dissolve the doubled pawn and open
the file. 

16 Nd5 f6 17 c4 

White has gained a bit of space. However, Black is preparing some
counterplay that I did not really judge correctly. 

17...Rab8 

Of course he cannot play...c6 yet because of Nb6 but after this
move Black is better placed for action on the queenside than
White. 

18 Rad1 

Moving the rook from the a-file means he can play....b5 without
his a-pawn dropping off, but I soon won't be able to stop...c6
and/or...b5 anyway and I want the rook on the d-file which I expect
to become open soon. I didn't want to move the king's rook there
because of...f5 counterplay. If 18 Qe1 (protecting b4, with the idea
Rf3 or Qc3) he unfortunately has 18...c6!.

18...b5 19 Rd2!? 

I see I gave this an exclamation mark in my old records. The rook
protects b2 and prepares doubling on a choice of files, as well as
the second rank. Looking at the game There is problem no better
move. Now, however, it seems that Black has found a plan (from
move 17 onwards) and White has not prepared correctly to meet it. 

19...c6 20 Nc3 (See Diagram)

20...Rb7?! 

After this, Black loses the opportunity to open the b-file.
Something must have put him off the idea. What was it? In my
1976 follow-up monograph, Vienna Opening, I see that I wrote "If
20...bxc4 21 dxc4 Rxb4 22 c5" but how correct is this for White?
No doubt I had lots of analysis in my long-lost notebook, but was it
any good? What should really happen after 22...Qb7!, for example
(a) 23 Rxd6 Rxb2 24 Rfd1 (24 Qg4 Rf7) 24...Qb4; (b) 23 cxd6
Rxb2 looks as if it should be good for Black. I obviously thought I
had worked something out here, and my opponent believed me,
since I suppose the passed d-pawn looked dangerous. However, 26
years later in the cold light of Fritz, it is far from obvious how
White continues here.

21 cxb5 axb5 22 d4 

"And Black is under a lot of pressure" I said in Vienna Opening.
Now my b-pawns are doubled and isolated but on a closed file,
which means the weakness is not too serious. If he captures on d4,
I obviously have a lovely game so now he must constantly contend
with the possibilities of d4xe5 or (more likely) d4-d5 when my
heavy pieces are correctly placed. His b5-pawn could then be a
weakness and if I could win it, then lo and behold, passed pawns
can march to glory! Obviously I don't have anything much to worry
about on the kingside, unless he is able to play...f5 under
favourable circumstances. However, if his queen gets to c4 then
my pawns become vulnerable.

22...Ng5 23 h4 Nh7! 24 Kh2? 

Incomprehensible. White doesn't want to play 24 dxe5 unless it
definitely achieves something by forced piece play, as it leaves all
White's pawns weak. The line 24...fxe5 25 Rfd1 Rf6 26 Qxe5 dxe5
27 Rxd7 Rxd7 28 Rxd7 probably looked drawish to me. White is a
pawn ahead, but with two pairs of doubled isolated pawns this is
going to be hard to prosecute. Instead 24 Rfd1! looks good,
retaining options of d4-d5 as well as dxe5.

24...Qf7 25 Qg4!?

25 d5 looks safer but less likely to get winning chances. I must
have underestimated his counterplay at move 26 when I allowed
his next move. - or maybe I correctly judged my opponent's lack of
ambition and comprehension of the position's resources! 

25...Qc4 

Awkward. It looks like I am losing a pawn.

26 Rfd1 (See Diagram)

 26...Rbf7! 

I see I gave him an ! for this but I have not preserved the analysis
on which that judgment was based. Looking at the game now, it
seems Black two or three times was afraid to take tactical
opportunities that presented themselves only briefly. However, in
this case White does seem to have enough activity and therefore I
think Rbf7 is the right move. I suppose he rejected 26...Qxb4
because of 27 dxe5! (27 Qe6+ Rbf7 is unclear) when Black must
definitely not play 27...dxe5? 28 Rd7! (28 Qe6+ regains the pawn
with maybe a slight advantage, but the rook invasion looks
stronger.) 28...Rxd7 29 Rxd7 Rf7 30 Qe6! and Black loses his
queenside. So the critical line is 27...fxe5! 28 Rxd6 Rbf7 29 Rxc6
(not 29 Rxh6 Nf6) when it's complicated but White may well have
some advantage. 

27 d5 

Thus I get in my lever and ensure play against his pawns which
now become targets. However, in view of what might have
happened in a few moves from now, this was a questionable
decision.

27...c5 

Capturing on d5 obviously gives me something for nothing. 

28 bxc5 Qxc5 

Black attacks the e-pawn and now White's rooks, doubled on a
closed file, look a bit silly. The knight is about to be driven back
too. 28...dxc5 would give me a protected passed pawn with the
option of playing d5-d6 at some time.

29 Qe2 

White was hanging been hanging on (ingeniously or desperately,
depending on how you calculate the critical moments just passed.)
This move might have allowed a powerful counter-attack but it is
hard to find anything better. If 29 Rf1 Qxe3 or 29 Rd3 Qb4.

29...b4 30 Na4 Qa7 31 b3 f5 32 Nb2 fxe4? 

This move is salvation for White. Black was guilty of very sloppy
thinking: he saw a pawn and an open file and played the obvious
move. But now e4 is a dead square and I can once more coordinate
my pieces effectively. However, there is a big hole in White's
recent play. The correct 32...Nf6! would start to give Black
winning chances on the kingside (because of those moves h4 and
Kh2 which he provoked much earlier). If then 33 exf5 Ne4 looks
nasty. 

33 Nc4 

Now I have a strong and secure square for the knight.

33...Qe7 34 Qg4 Nf6 35 Qe6 

...And his extra pawn is pretty meaningless as he has to hang on to
d6 and b4. Clearly exchanging queens on e6 is not a good option. 

35...Ne8 

35...Rd8 looks much the same. I suppose a draw is the most likely
result, whatever computers may say about the position.

36 Qxe7 Rxe7 37 Ra1 

At last the open a-file means somethingthe rook can use it to attack
his weaknesses from the side - or from behind!. In the final phase,
it seems as if Black realises too late that his position has
deteriorated and he cannot find a way to hold it together.

37...Rb7 38 Ra8 

With the idea of Rd8. 

38...Nf6 

He tries to tempt me to simplify (39 Rxf8+? Kxf8 40 Nxd6 Rd7). 

39 Ra6! 

In this semi-endgame, only White has winning chances despite his
pawn minus. This is not the sort of position that computers judge
very accurately because they overvalue the pawn on e4. 

39...Rd8 40 Kg1! (See Diagram)

White will improve his king position, and guard the e-pawn in
particular, before trying to win back the sacrificed pawn. There is
also a psychological trap, into which he falls. When one player
marches his king to the centre, the opponent tends to follow suit. 

40...Kf8 41 Kf2 Ke7? 

Blunder in a deteriorating position! Now White plays a
"randspringer" move and switches the focus to c6, a square which
was previously of no significance. 

42 Na5 

Black loses the exchange; one rook or other must be given up for
the knight. 

42...Rdd7 43 Nxb7 Rxb7 44 Ra5 

To prevent...Rb5. 

44...Nd7 45 Rda2 

White must now exchange rooks to make progress. He cannot stop
this. 

45...Nc5 46 Ra7 Rxa7 47 Rxa7+ Kf6 48 Ke2 (See Diagram)

His next move helps me but probably he is lost anyway. 

48...Nxb3 49 Rd7! 

Much stronger than Rb7. 

49...Nc1+ 50 Kd2 Nd3 51 Rxd6+ Kf7 52 Rd7+ Kf6 53 Rb7 (See
Diagram)

Here Black resigned. This resignation was slightly premature but
he is going to lose in the end. 53...g6 54 g3 h5 55 Kc2 g5 56 Rb6+
Kf5 57 d6 Nc5 (57...Ke6 58 hxg5) 58 Rxb4 is probably what he
reckoned on.

This was a game of many twists and turns and reversals of fortune.
Maybe my determination to score the point with "my" opening was
somehow the decisive factor, but clearly Black recovered from an
early minimal disadvantage and missed some opportunities to
make life much harder for me. Looking back at this tournament in
an objective way, I can only say it was a learning experience. I did
not make the most of my chances in some games (e.g. the one with
Kroll), while opening experiments unsuitable to correspondence
play did cost me some points and half points too. This game with
Dr Karafiath, however, was strangely like an over the board game
in that psychological factors seemed to play a role. I was
determined to win, even when my position was shaky, and I
thought I knew what I was doing even when I didn't.

But if my younger self came forward in a time machine and read
these words and checked these analytical variations, he might say
I've got it all wrong.

The past is another country, as someone said. They play chess
differently there - even if "they" is one's past self!