The Kibitzer 
by Tim Harding

Some Openings Topics Revisited

FROM TIME TO TIME, readers respond with queries or
additional material about my Kibitzer articles but the normal
format of this columns doesn't really suit replying. However, I
decided that this might be a good time to look at some of the more
interesting points that have arisen over the two and a half years of
The Kibitzer.

The columns relating to openings topics seem to be the most
popular, or perhaps it is just that they generate more specific
queries. The recent appearance of the book "Nunn's Chess
Openings" (a single-volume encyclopaedia in the old "MCO"
format) also gives an opportunity to check what I said against
"official" theory and perhaps spot cases where that book falls short.
Henceforth that new book will be referred to as "NCO". This
month I will look at a few of the opening variations I have dealt
with and see what both readers and "NCO" have to say about them.
You can download the old Kibitzers by number from the Chess
Caf‚ archives if you want to read these earlier articles.

The Urusov Gambit, which can arise from either the Petroff
Defence (1 e4 e5 2 Nf3 Nf6 3 d4 exd4 4 Bc4) or the Bishop's
Opening (2 Bc4 Nf6 3 d4 exd4 4 Nf3) seems to attract interest on
the Internet out of all proportion to its position in "official" theory.
I dealt with this gambit in the September 1998 The Kibitzer, while
the main line arising from 4...Nxe4 5 Qxd4 Nf6 6 Bg5 Be7 7 Nc3
was analysed in the next month's Kibitzer. NCO has little
worthwhile to say on this Urusov Gambit. Row 1 on page 287
gives the main line 7...Nc6 8 Qh4 d6 (or 8...d5 9 0-0-0 Be6 10
Rhe1) 9 0-0-0 Be6 10 Rhe1 and ends with the "compensation for
the material sacrificed symbol". However the game is only
beginning at this point! A note also deals with 7...c6 8 0-0-0 d5 9
Rhe1 Be6 (postponing castling) but doesn't really make any firm
suggestions, except perhaps concerning move order. Probably this
line with 7...c6 (rather than 7...Nc6) should be considered the
critical line of the Urusov Gambit now. My articles last year dealt
with 9...0-0 10 Qh4 and now various possibilities ending up with
10...Bf5. (See Diagram)

S.R. Gibson wrote in from Britain to say "I found your article very
interesting, but I am puzzled by the final game, Zarske-Schneider.
Why can't White play 11. Bxd5, as he did against 10...h6? Maybe
I'm missing something (more than likely), but I thought that the
whole point of 10...Be6 was to prevent 11 Bxd5."

I replied that in Schlechter v. Neustadtl & Tietz, Carlsbad 1901
(September Kibitzer) after 10...h6 White could play 11 Bxd5
because after 11...cxd5 12 Rxe7! Qxe7 13 Nxd5 Black has no
threat and if for example 13...Qc5 14 Bxf6+- In Zarske-Schneider
(October Kibitzer) 10...Bf5 cannot be met by 11 Bxd5?? because
after 11...cxd5 12 Rxe7 Qxe7 13 Nxd5 Qc5 Black threatens mate
on c2. Also h7 is protected. This gives Black a tempo to organise
his defence (e.g. 14 Ne3 Nbd7) and Black is after all a rook ahead.

A much more cogent criticism came from Francesco Rinaldi and
Nello D'Alessandro, who wrote from Italy about a similar line"We
read with great pleasure your articles about this opening in the
column The Kibitzer, and we would like to participate with our
ideas."

Their analysis concerns the variation (1 e4 e5 2 Bc4 Nf6 3 d4 exd4
4 Nf3 Nxe4 5 Qxd4 Nf6 6 Bg5 Be7) 7 Nc3 c6 8 0-0-0 d5 9 Rhe1
Be6 10 Qh4 Nbd7 11 Bd3 (NCO gives here 11 Nd4 Nc5 12 f4
unclear, citing Mark Tseitlin-Lev, Tel Aviv 1990.) 11...Nc5 (NCO
also gives "11...c5 unclear" without mentioning the game
Timoschenko-Karpov.) 12 Nd4 which is where NCO ends here,
saying "unclear"; 12...Ng8 is one critical line here. My Italian
readers rightly take issue with what I said about the alternative
12...h6 which is perhaps playable after all. They have obviously
examined this in far more detail than I had time to do. (See
Diagram)

"First of all, we don't understand why do you give a (!) to the move
13 Nf5. In fact, it seems that after 13...Kd7! White is in trouble.
For example, 14 Bxf6 Bxf6 (not 14...,Nxd3+ 15 Rxd3 Bxf6 16
Qb4 !? and White has counterplay) 15 Qg3 Bxc3 16 bxc3 Qg5
Black stands better." They are right. "Instead, we think White can
play 13 Bf5, entering by transposition in a variation given in the
last Kibitzer, after the moves 13... Kd7 14 Bd2. At the this point
the variation ends, but we'd like to ask your opinion on our analysis
about the obvious move for Black, 14...Nfe4."

They suggest the very interesting queen sacrifice 15 Nxe4! arguing
that it is not good for Black to decline it, in view of 15...Nxe4 16
Qg4 Bxf5 17 Qxf5+ Ke8 18 f3 for example 18...Nd6 (18...Nxd2?
19 Rxd2 +-) 19 Qe5 and now if 19...Kf8 20 Bb4, and if 19...Kd7
20 Nf5. The critical reply seems to be 15...Nd3+ since on the
immediate 15...Bxh4 we reach the same positions as the main line
with the d-file open for White.

After 15...Nd3+ 16 cxd3 Bxh4 17 Nc5+ Black has a difficult
choice. I don't entirely agree with the details of their analysis but
they are on the right track, I think. The white rook and minor
pieces generate tremendous activity while Black is underdeveloped
and his queen and king exposed. In some lines Black must return
the queen to reach an endgame; in other lines, even that is not
possible. They say that 17...Kd6 and 17...Kc7 lose easily so the
defences to be considered are 17...Ke7, 17...Ke8 and 17...Kc8.
Taking the main lines in reverse order to the analysis from Messrs.
Rinaldi and D'Alessandro, they say that 17...Kc8 is "perhaps the
best for Black" and after 18 Ndxe6 fxe6 they consider both 19
Rxe6 and 19 Nxe6. In my opinion, 19 Rxe6 can be discarded as 19
Nxe6! refutes 17...Kc8. On 18...Qf6 19 Nxg7+ Kc7 the move 20
Bf4+ wins at once for White, while if 18...fxe6 (which they didn't
mention) 19 Rxe6 Kc7 White wins by 20 Rg6! (not 20 Ba5+ b6 21
Na6+ Kb7 22 Nc5+=) 20...Kb6 21 Ba5+!. Therefore 17...Kc8 is
unplayable. If instead 17...Ke8 they give the plausible variation 18
Ndxe6 fxe6 19 Bg6+ Ke7 20 Rxe6+ Kf8 21 Nxb7 Qd7 22 Rde1
Be7 23 Rxe7 Qxe7 24 Rxe7 Kxe7 25 d4+-. Therefore the main line
seems to me to be their variation 17...Ke7 18 Bb4! Qb6 19 a3 Bxf2
20 Ndxe6 Bxe1 21 Rxe1 fxe6 22 Na4+ Qxb4 23 Rxe6+ Kf7 24
axb4. White has two pieces for a rook but the game still has to be
won. Finally, I observe that Black is not forced to enter this line
after 12...h6 13 Bf5 Kd7 as 13...Qd7 is also possible with a
completely different line for readers to investigate.

Not all the openings articles in The Kibitzer have been about
gambits. For example number the June 1997 Kibitzer dealt with
the Franco-Benoni, 1 e4 e6 2 d4 c5 3 d5, a line which I had to meet
in a high-level correspondence game. Geoff McAuliffe PhD.
supplied me a with reference I didn't know apparently the book
"Barcza-Larsen Defense" by Paul H. Fields (Chess Enterprises,
1991) apparently contains 75 annotated games with 1. e4 e6 2. d4
c5 including games by Karpov (as white) and Kasparov (as black).
McAuliffe informed me that "The list of references is impressive.
Fields is a National Master here in the States."

Unfortunately, so far I have not been able to obtain a copy of this
monograph in Europe but if anybody plays this defence against me
again, I guess I'll need it in a hurry. Does any reader have a copy to
spare? I will consider exchanging for a copy of one of my own
books.

We also dealt with the Two Knights Defence, in one of the earliest
Kibitzers, October 1996. "Two Knights' Defense, the Amazing
Counterattack" which dealt with 1 e4 e5 2 Nf3 Nc6 3 Bc4 Nf6 4
Ng5 and now 4...Nxe4!? (See Diagram). 

NCO does not mention this possibility and there are only 22 games
with this move in my database of over one and a half million
games so it is certainly not that common. Note that White cannot
win a piece by 5 Nxe4 because of the fork 5...d5.

One correspondent, who just signed himself "Jr" wrote "Your
article in Chess Caf‚'s The Kibitzer on this not so well known
variation of the Two Knights has become my only defence lately
against 4 Ng5... I know that it may not be the soundest but I have
surprised many opponents with it. Some good results, some bad,
but always wild... Just the look in some of my foes' faces is worth
playing it." I have to confess, though, that I also received a couple
of messages from masters (which I now cannot find in my archive)
informing me of their opinion that 4...Nxe4 is "rubbish" or "a very
bad move". Of course such matters are often relative to the level of
opposition one meets and also whether one is playing serious over-
the-board games at normal time limits, Internet blitz, or
email/postal correspondence chess at three days per move. What
will work brilliantly as a surprise weapon in one context may fail if
the opponent has plenty of time for thought or research.

So what should White play in this line? I wrote in the October
1996 Kibitzer that 5 Bxf7+ Ke7 6 d4 must be critical, because after
6...d5 (the reply usually cited) White has the very strong reply 7
Nc3! but R. Maric's move 6...h6 is better, and after 7 Nxe4 Kxf7 8
d5 Ne7 9 Qh5+ Kg8!!. I suggested that White should instead play 8
dxe5 as in Krgin-Maric, Novi Sad 1950, which continued 8...Qe8 9
Qd5+ Kg6 10 f4 Nb4?! and Black eventually won but wrote later
that he should have played 10...d6.

I have since found an earlier game in this line on the excellent New
In Chess CD about 50 years of the Hoogovens tournaments. H.
Van Steenis - C. Vlagsma, Beverwijk 1942, went 8 dxe5 Qe8 9 f4!
d6 10 0-0 Kg8 11 Nbc3 dxe5 12 f5! Qf7 13 Nd5 Bd7? 14 f6 g6 15
Ne7+ Nxe7 16 fxe7 Qxe7 17 Nf6+ Kg7 18 Qxd7 Qxd7 19 Nxd7 1-
0. Black can try to improve on this by 13...Nd4 but after 14 f6!? g6
15 Ne3 and 16 c3 things look bleak for him, and 14...Qxd5 15 f7+
Kh7 is not a line that many would willingly enter. There is an
amusing continuation 16 Qd3 (threatening double check and mate,
to which the only answer is a "charge of the light brigade" by
16...Kg6! 17 Nc3+ (17 Qg3+ Kh7 repeats.) 17...e4 18 Nxd5 Ne2+
reaching an ending which looks terrible (c7 hangs and the f7 pawn
looks strong) although Black may have some traps. If you don't
like this, simply 14 Ne3 (followed by c2-c3) looks effective despite
Black's bishop pair.

This variation is more convincing than most for White so I cannot
honestly recommend 4...Nxe4 anymore unless you are willing to
take a gamble. Here is another line in the Two Knights about
which I have received correspondence lately.

Miles Scull wrote that after 1 e4 e5 2 Nf3 Nc6 3 Bc4 Nf6 4 Ng5 d5
5 exd5 Nd4 (See Diagram) "6 d6! is very good for White, but I
don't find this line anywhere; only the 6. c3 line". This was
intriguing and I had to challenge him about it. I wrote, after a little
research, "If you believe 6 d6 is very good, please tell me how
White improves on Bogoljubow-Rubinstein, Stockholm 1921,
which is quoted in several books in my library. 6...Qxd6 is the
correct reply when 7 Nxf7? loses to 7...Qc6 so Bogo tried 7 Bxf7+
Ke7 8 Bb3 but stood worse, and Estrin also won a game as Black
in this line." NCO only gives 7 Bxf7+ Ke7 with the 'edge to Black'
symbol and doesn't mention 7 Nxf7 at all. However, Mr Scull soon
pointed out that the "?" often given to 7 Nxf7 is an exaggerated
criticism.

"Regarding 6 d6, don't follow Bogo-Rubin., instead wait,
preserving tension, say with7 d3 Nd5 8 0-0 f6 9 Ne4 Qc6 10 Bxd5
Qxd5 11 a4 c5 12 c3... And White stands somewhat better, not
equal as in mainline. Or, have I missed something?"

I am unsure of the merits of that sequence, which seems to include
several inferior moves. Before I could answer, another broadside
arrived "You said 6...Qxd6 is correct because 7 Nxf7? loses to
7...Qc6. But, on second glance, is that really true in light of 8 0-0,
Qxc4 9 Nxh8 Nxc2 10 d3 (not 10 Na3? Bxa3) 10...Qc6 11.Nc3
Nxa1 12 Bg5 Bf5 13 Re1 0-0-0 14 Nf7... seems to lead to good
play for White." 

I was suffering from flu at the time but threw out a few ideas at
random "Your 8 0-0 is some improvement on the lines normally
cited but not good enough to revive a bad idea at master level I
think. When I was quite delirious a couple of days ago I fancied
that 8 0-0 would be smashed by 8...Bh3!? but now that an
antibiotic is getting my temperature down to the level where I can
check my musings against Fritz, I see that White only gets mated if
he is foolish and White is able to bail out for a repetition draw in a
couple of ways." The variations after 8...Bh3 are quite amusing and
instructive for Kibitzer readers find for yourself. 

Shots like 8...Bh3!? are typical of this kind of line; you may find
advantages for White if you give Black the same kind of
materialistic moves that White is making. Instead of making the
a1-R his main goal, Black can combine central piece pressure with
the goal of picking up the h8-N. I think most masters who play
open games a lot would agree that after 9 Nxh8 White is way
behind on development and has several vulnerable spots, for the
sake of a material advantage that he is unlikely to be able to
maintain. He has shot his bolt - the B has gone, the d-pawn has
gone, the N has exhausted its options and the whole queenside is
undeveloped. Black can exchange queens by ...Qxc2 so long as he
picks up the h8-N soon afterwards. Therefore I would be inclined
to play 9...Bg4 and see if White is willing to weaken himself
further by 10 f3. Also in your main line Black can play 12...g6 and
there are probably other tries. (Indeed, my correspondent later
wrote to indicate the improvement 13...Qe6.)

However, one point that did emerge from this little debate is that 6
d6, while not as strong as Mr Scull first thought, is also not as bad
as the books suggest. His move 8 0-0 is harder to refute than other
moves previously tried, while with his shrewd 7 d3 White isn't
worse. Then Black doesn't have to play his 7...Nd5 but after the
safe 7...Ne6 a draw is very likely between evenly-matched
opponents; maybe a player with the subtlety of an Ulf Andersson
might even generate a slight nagging initiative for White. 

Not to worry, Black! Hans Berliner reckons 5...Nd4 is not even
best since Black need not commit the knight yet; he played 5...b5 6
Bf1 Nd4 in his famous win against Estrin. I think (depending on
your strength) if you played Scull's line against ordinary players
you would get a few wins, especially in blitz, but above a certain
level White will get carved up. Mr Scull also wrote that he was
acquiring one of my Evans Gambit books and proposed to look at
some of the main lines with a chess analysis program on his new
Pentium II 400. I warned him"Typically computers will judge any
Evans Gambit position as good for Black unless White has a forced
short-range continuation leading to advantage. Humans are able to
judge positional compensation and sometimes to calculate beyond
the computer's horizon, so don't be too reliant on your electronic
toys."

Finally, let us take a quick look at the Marshall Attack in the
Spanish (Ruy Lopez) which was featured in the May 1998
Kibitzer. There I discussed the important improvement for White
found by American amateur Daniel J. Quigley in the line 1 e4 e5 2
Nf3 Nc6 3 Bb5 a6 4 Ba4 Nf6 5 0-0 Be7 6 Re1 b5 7 Bb3 0-0 8 c3
d5 9 exd5 Nxd5 10 Nxe5 Nxe5 11 Rxe5 c6 12 d4 Bd6 13 Re1 Qh4
14 g3 Qh3 15 Be3 Bg4 16 Qd3 Rae8 17 Nd2 Re6 18 a4 f5. (See
Diagram)

Now "NCO" devotes most of its page 346 to analysis of the
"official" main line 19 Qf1 with a small note dismissing 19 f4? but
there is no mention of the move which seems to refute the whole
line   widely publicised last year on the Internet and in my
magazine "Chess Mail". Grandmaster John Nunn, who wrote this
section of the book himself, must have been too busy with his new
baby to keep up with developments, or maybe this is an example of
his attitude to the medium expressed on page 168 of his popular
book Secrets Of Practical Chess, where he writes that apart from
email and a few other specific uses, the Internet "strikes me as
being a great time-waster". 

Evidently Nunn chiefly examined previously published analysis
and grandmaster games, disregarding games in databases if they
were not played by well-known players. For example, in his
introduction he makes much of the fact that a footnote on page 344
demolishes analysis (in the 11...Bb7 line of the Marshall) given by
Informator, ECO and Anand. However, two pages later he
overlooks one of the most important theoretical discoveries of
1998! It was certainly first published early enough in the year
(March) to be included in NCO.

So the important page 346 in this much-hyped book is out of date
even before it was printed. If Nunn had kept up with the newsgroup
rec.games.chess.analysis or even read The Kibitzer or Chess Mail
regularly, then this would not have happened! 

I don't want to go into the whole line in great detail again.
Quigley's main point is that White can play 19 axb5!! which was
supposed to be bad on account of 19...f4, because then the piece
sacrifice 20 Bxf4 is very strong because of White's queenside
counterplay." If Nunn had noted the discussion, I think he would
certainly have mentioned the possibility so I think it is very
probable that he overlooked the idea altogether.

It is very hard to see a good continuation after 20 Bxf4. For
example, if 20...Nxf4 21 Qf1! (not 21 Bxe6+), or 20...Rh6 21 Bf1!
(also analysed in the May 1998 Kibitzer).

Black has tried instead 20...Rxf4 when if 21 Rxe6 Black forces a
draw by 21...Rxf2! 22 Re8+ (22 Kxf2 Qxh2+ also draws.) 22...Kf7
23 Bxd5+ cxd5 24 Re7+ Kf6 25 Re6+ Kf7 26 Re7+ «-« G.
Gorges-K. Kling, Germany corr. 1987. It is too risky for either side
to vary from this, except by other lines that lead to perpetual check.

Quigley's improvement in this line is again based on 21 Qf1! e.g.
he analysed 21...Qh5 22 bxc6 Rff6 23 Rxe6 Bxe6 24 Ra5! Rh6 25
h4+-

The other critical line is 20...Bxf4. In this case 21 Rxe6 is correct
and after White has two possible methods

b1) 22 bxc6 Rb8 23 Qe4 Bxd2 24 Bxd5 Bxd5 25 Qxd5+ Kh8 26
Rxa6 Re8 27 Ra1 Qg4 28 Qd7 Qe2 29 c7 Rf8 30 Qf7 1 0 A.
Uhlig-J.Guzman, Compuserve casual corr. 1996.

b2) Quigley's preference 22 bxa6 Bxd2 when a 1998 IECG master
class game A. Bell (2270) - J. Simmelink, continued 23 Qxd2 Ra8
24 Qe2 Bf7 25 Qe5 Qd7 26 Bc4 Nf6 27 Bxf7+ Qxf7 28 Ra5 h5 29
Qf5 Qe7 30 h3 Qe2 31 a7 Kh8 32 Ra2 Ne4 0-1. However,
although it is interesting to see a player as strong as Simmelink
(editor of the Dutch correspondence chess magazine
Schaakschakeringen) willing to test this line, his opponent did not
really follow Quigley's recommendation to the end, which is why
he came unstuck.

I wonder how Simmelink intended to meet 23 a7 when Quigley
analysed 23...Ra8 24 Qxd2 Bf7 25 c4 Bf6 26 Ra3 intending Bd1 to
protect against ...Bg4 and then to activate the Queen. Somebody
then suggested the improvement 23...Bc7 24 Bxe6+ Qxe6 25 Qxd2
after which Black has a knight for 3 pawns. Possibly that is what
Simmelink intended but I would be surprised if it were correct.

Simmelink's game seems to be the only one so far published that
tests the Quigley plan at all. Martin Bennedik organised a small
Scheveningen-system email thematic tournament at the end of last
year, dealing with the 11...c6 Marshall, but so far as I am aware no
completed games featured the move 18...f5. Black has preferred to
play 18...bxa4 or to diverge at an earlier point. You can follow the
progress at this URL http//www.hlp.de/bennedik/marshall.html. 

The answer to last year's question "Has the Marshall Gambit been
refuted on the Internet" seems to be "not quite". The move 18...f5
is looking sick, but other lines are still playable. 

Martin will soon have an article out about the Marshall in the
German chess history/theory magazine "Kaissiber" with an
updated version to follow in Chess Mail's gambit issue in May. If
any readers have suggestions or interesting Marshall games, please
let me know!