Main
Date: 25 Aug 2007 13:33:42
From: help bot
Subject: game annotations
I am currently reading the August 2007 issue of Chess Life
magazine, and an interesting this is that the very same game
is annotated therein by two different people. More accurately,
the first time is credited to a combination of GM Susan Polgar
and FM Paul Truong, while the second annotator is IM John
Watson.

IMO, the Polgar/Truong team did a much better job, as
indicated in both their handling of the computer-aided analysis
and in their interpretations of the often times nonsensical
results computers can produce. For instance, in the position
after:

1.e4 e5
2.Nf3 Nc6
3.c3 f5
4.d4 fe
5.Nxe5 Nf6
6.Bb5 Bd6
7.Nc4 Be7
8.Ba4 d5
9.Ne5 O-O
10.Bxc6 bxc6
11.Nxc6 Qe8
12.Nxe7+ Qxe7
13.O-O Ng4
14.h3 e3
15.Bxe3 Nxe3
16.fxe3 Bxh3
17.Rf3 Bg4
18.Rxf8+ Rxf8
19.Qe1 Rf6
20.Nd2 Rg6
21.Qg3 Qe6
22.Qf4 Bh3
23.g3 h5
24.e4 Rg4

Here a possibility which did not occur in the game was
analyzed by both Polgar/Truon and John Watson, but with
revealingly different results. JW strongly implies that the
line: 25.Qe5 Qxe5 26.dxe5 Rxg3+ 27.Kh2 Rd3 28.exd5
Bg4 29.Nb3 Rxd5 30.Re1 leads to a very slim advantage
for White, and blames a later move for losing the game.

This entire line is just following along computer analysis,
which explains why there is zero difference in the moves
given by both annotators. But the key here is that at the
tail end, JW looks at the number on his computer screen
and concludes nothing is particularly wrong, just a slight
disadvantage for White, while the Polgar/Truong team
notice that Black has something called...

CONNECTED PASSED PAWNS.

Of course, we all know that connected passed pawns
are not just a slight problem, but rather a very severe one.
The Polgar/Truong team call this "a clear advantage"
(not a "winning advantage" because there are still some
difficult problems to solve). Because of this superior
grasp of the position, we have to back up to move 22 in
order to find the real culprit, the real losing move.

Apparently, there is a whole lot of computer-dominated
analysis these days, and this explains why so much of
the annotations look almost exactly alike. But I find that
you can't always go by the numbers on the screen, as
computers don't really "comprehend" strategy nearly as
well as they do tactics. In another game, which I will
attempt to tackle in a separate post, this problem rears
its ugly head again.


-- help bot





 
Date: 26 Aug 2007 09:45:02
From: help bot
Subject: Re: game annotations
On Aug 26, 6:00 am, Sanny <[email protected] > wrote:

> Very hard to grasp

So difficult to calculate, that even the loser of this game,
a grandmaster, could not do it OTB. And it is very likely
that the winner did not intend to allow an escape into a
Rook ending either, but simply overlooked the combination.
I say this because, for instance, he chose to castle on
opposite wings, which usually signals an "all or nothing"
attitude.

When you break it down, there are only a few different
lines to analyze, so it's not hard if you can visualize a
few exchanges and assess the quiet positions which
result after all exchanges. All that was needed was a
quick glance at the ridiculous self-fork line: ...Nxe5,
dxe5 Bxe5 (of course not, because it hangs the Bishop,
right?), and this should be a piece of cake for a GM, I
would think.

Notes by the winner would be very useful, so we
might get a grasp of what he thought during the game,
and hopefully of what the loser said to him afterward.
Many times the two players will reveal what they
believed were the critical errors in a post mortem.
Instead, what we got in Chess Life was Fritz (or
whatever) having no difficulty calculating the exact
tactics, which is great but somehow leaves a hollow
feeling. I seriously wonder if the annotator feels
superior because his chess program found the right
moves here, while the GMs did not. The idea is
grotesque, since he seems to have failed to grasp
the strategic fundamentals of the position; the fact
that White's "normal" plan is to launch a pawn
storm against the fixed weakness at h6, while
Black must generate counterplay to derail this plan
very quickly or lose.


-- help bot





 
Date: 26 Aug 2007 04:00:50
From: Sanny
Subject: Re: game annotations
On Aug 26, 2:54 am, help bot <[email protected] > wrote:
> In a second example of the issues relating to modern
> computer-dominated annotations, I would like to go over
> a game from the U.S. Championship between GMs
> Kaidanov (as White) and Alexander Ivanov, a Caro-Kann.
>
> After the opening -- all routine or "book" -- moves:
>
> 1.e4 c6
> 2.d4 d5
> 3.Nd2 de
> 4.Nxe4 Bf5
> 5.Ng3 Bg6
> 6.h4 h6
> 7.Nf3 e6
> 8.h5 Bh7
> 9.Bd3 Bxd3
> 10.Qxd3 Nf6
> 11.Bf4 Bb4+
> 12.c3 Bd6
> 13.Ne5 O-O
> 14.O-O-O
>
> ...the annotator, again John Watson, says that "Black
> has no serious weaknesses". IMHO, this is just plain
> wrong, as Black has moved one of the pawns in front
> of his King, and this is a well-known issue addressed in
> numerous books and articles on chess. It may well be a
> common position, but this hardly implies the existence of
> "no serious weaknesses".
>
> One routine method of exploiting such weaknesses as
> this is to advance the pawns up the board and trade off
> the Black King's cover, pawn for pawn. Another idea, not
> seen in this game, is to attack the pawns with pieces in
> front of White's pawns, as with R-h3 or Q-g3, for instance,
> and this nearly always involves the sacrifice of material.
>
> But back to this game: 14. ...N/b-d7. No comment by
> annotator JW, even though it might be pointed out to
> weaker players that the loose Bishop on d6 is "safe" on
> account of, for instance, 15.Nxc6?? Bxf4 being a check.
>
> 15.Ne4
>
> Now this is apparently where the hackneyed openings
> theory comes to a screeching halt, and White introduces
> a "new" move (which is not to say it hasn't ever been tried
> before). Black's decision here is critical to the rest of the
> game, as you might expect.
>
> 15. ...Nxe4
>
> What I believe has happened here is that Black has
> been "taken out of book", and he was unable to properly
> calculate the tactics involved in various exchanges on
> e5; simple as that. The fact that even a grandmaster
> can't do this is reassuring, for otherwise we could all
> just resign at move two and go for lunch, saving energy
> for games where we actually have a ghost of a chance
> to draw or win.
>
> After GM Ivanov's inferior 15. ...Nxe4 White has little
> trouble executing a routine pawn storm which later
> results in a mating attack on the Black King, so I
> want to focus on this, the key moment when he might
> very well have finished the game with a crusher and an
> offer of a grandmaster draw to avoid the agonies of a
> long and boring Rook endgame struggle.
>
> At this point JW points out, in rather odd language,
> that "...Nxe5 would be a normal policy, since the tactics
> work out". Translated, what he means is that *his Fritz
> computer* had no difficulty seeing that in all variations,
> there is an escape route to a tenable Rooks and pawns
> ending. Unfortunately, his analysis seems to be missing
> some key lines. It appears as though the editor zapped
> his work, or else he got very confused and left out the
> most important part, while analyzing on and on in what
> was intended to be a mere sub-variation. Maybe it was
> just the omission of a right parenthesis -- I don't know.
>
> The correct line was 15. ...Nxe5!! 16.dxe5 Bxe5, and
> even though it would seem that Black has forked himself
> or moved his pieces en prise, the thing is that White's
> own Knight on e4 is without any support after any trade
> of Queens, so things even out and the move 15.Ne4 is
> shown to be innocuous. Once again, the fact that the
> move ...Bxf4 *is a check* is key to the outcome, though
> that is never explained. (Heck, Fritz "knows" this at one
> half second per move, so why even mention it?)
>
> At any rate, the fact remains that in this line of the
> Caro-Kann, Black has a serious problem in that he has
> advanced one of the pawns in front of his King, and thus
> he is a sitting duck for the kind of attack that GM
> Kaidanov very competently used to win this game.
>
> The nearly random sprinkling of exclamation ks,
> half-hearted queries (?!), and so forth, is no way to
> annotate a game and if anything, it can lead to the
> lazy over-reliance on computers to assess everything
> properly so the writer need not have any clue what is
> really going on. In particular, I am noticing that the
> annotations in Chess Life are over-sweetened with
> exclams praising the eventual winner's every whim,
> regardless of any objective merit or lack thereof. A
> game's ultimate winner is described as "enterprising",
> while his (equally enterprising) opponent's moves are
> termed risky or worse. This is just plain silly.
>
> What I do like about computer-aided analysis is that,
> right or wrong, a computer is 100% objective, and it
> never plays favorites. In fact, with some work, I wonder
> if Fritz 10 or something like it could not churn out
> some publishable game annotations which are at least
> as good as what I am currently seeing in Chess Life.
>
> -- help bot

Very hard to grasp

Bye
Sanny

Play Chess at: http://www.getclub.com/Chess.html



 
Date: 25 Aug 2007 14:54:16
From: help bot
Subject: Re: game annotations
In a second example of the issues relating to modern
computer-dominated annotations, I would like to go over
a game from the U.S. Championship between GMs
Kaidanov (as White) and Alexander Ivanov, a Caro-Kann.

After the opening -- all routine or "book" -- moves:

1.e4 c6
2.d4 d5
3.Nd2 de
4.Nxe4 Bf5
5.Ng3 Bg6
6.h4 h6
7.Nf3 e6
8.h5 Bh7
9.Bd3 Bxd3
10.Qxd3 Nf6
11.Bf4 Bb4+
12.c3 Bd6
13.Ne5 O-O
14.O-O-O

...the annotator, again John Watson, says that "Black
has no serious weaknesses". IMHO, this is just plain
wrong, as Black has moved one of the pawns in front
of his King, and this is a well-known issue addressed in
numerous books and articles on chess. It may well be a
common position, but this hardly implies the existence of
"no serious weaknesses".

One routine method of exploiting such weaknesses as
this is to advance the pawns up the board and trade off
the Black King's cover, pawn for pawn. Another idea, not
seen in this game, is to attack the pawns with pieces in
front of White's pawns, as with R-h3 or Q-g3, for instance,
and this nearly always involves the sacrifice of material.

But back to this game: 14. ...N/b-d7. No comment by
annotator JW, even though it might be pointed out to
weaker players that the loose Bishop on d6 is "safe" on
account of, for instance, 15.Nxc6?? Bxf4 being a check.

15.Ne4

Now this is apparently where the hackneyed openings
theory comes to a screeching halt, and White introduces
a "new" move (which is not to say it hasn't ever been tried
before). Black's decision here is critical to the rest of the
game, as you might expect.

15. ...Nxe4

What I believe has happened here is that Black has
been "taken out of book", and he was unable to properly
calculate the tactics involved in various exchanges on
e5; simple as that. The fact that even a grandmaster
can't do this is reassuring, for otherwise we could all
just resign at move two and go for lunch, saving energy
for games where we actually have a ghost of a chance
to draw or win.

After GM Ivanov's inferior 15. ...Nxe4 White has little
trouble executing a routine pawn storm which later
results in a mating attack on the Black King, so I
want to focus on this, the key moment when he might
very well have finished the game with a crusher and an
offer of a grandmaster draw to avoid the agonies of a
long and boring Rook endgame struggle.

At this point JW points out, in rather odd language,
that "...Nxe5 would be a normal policy, since the tactics
work out". Translated, what he means is that *his Fritz
computer* had no difficulty seeing that in all variations,
there is an escape route to a tenable Rooks and pawns
ending. Unfortunately, his analysis seems to be missing
some key lines. It appears as though the editor zapped
his work, or else he got very confused and left out the
most important part, while analyzing on and on in what
was intended to be a mere sub-variation. Maybe it was
just the omission of a right parenthesis -- I don't know.

The correct line was 15. ...Nxe5!! 16.dxe5 Bxe5, and
even though it would seem that Black has forked himself
or moved his pieces en prise, the thing is that White's
own Knight on e4 is without any support after any trade
of Queens, so things even out and the move 15.Ne4 is
shown to be innocuous. Once again, the fact that the
move ...Bxf4 *is a check* is key to the outcome, though
that is never explained. (Heck, Fritz "knows" this at one
half second per move, so why even mention it?)

At any rate, the fact remains that in this line of the
Caro-Kann, Black has a serious problem in that he has
advanced one of the pawns in front of his King, and thus
he is a sitting duck for the kind of attack that GM
Kaidanov very competently used to win this game.

The nearly random sprinkling of exclamation ks,
half-hearted queries (?!), and so forth, is no way to
annotate a game and if anything, it can lead to the
lazy over-reliance on computers to assess everything
properly so the writer need not have any clue what is
really going on. In particular, I am noticing that the
annotations in Chess Life are over-sweetened with
exclams praising the eventual winner's every whim,
regardless of any objective merit or lack thereof. A
game's ultimate winner is described as "enterprising",
while his (equally enterprising) opponent's moves are
termed risky or worse. This is just plain silly.

What I do like about computer-aided analysis is that,
right or wrong, a computer is 100% objective, and it
never plays favorites. In fact, with some work, I wonder
if Fritz 10 or something like it could not churn out
some publishable game annotations which are at least
as good as what I am currently seeing in Chess Life.


-- help bot