| |
Main
Date: 13 Sep 2007 23:34:14
From: Sanny
Subject: Can you tell me about pawn structure?
|
Till now GetClub Chess only avoid double pawn formation and isolated
pawns.
Play Chess at: http://www.GetClub.com/Chess.html
What type of Pawn structures are good and what type of pawn structure
are bad?
1. If Pawns are on same rows not protecting each other {in front of
King}. (Is it good)
King has no place to move???
2. All pawns protecting other making a long diagonal Chain. (Is it
good)
A pawn Exchange can bring it down to Isolated Pawns???
3. Two pawns on side protecting middle pawn. (Is it good)
There is a hole in Center???
How to find weak spots? Whenever a Pawn is played we have toi leave a
diagonal hole for bishop or a place for opponent Knight.
Can you tell me how to tell which pawn structure is better.
I went through lots of sites to know about Pawn Formations but could
not find any concrete example.
How to tell whether a Pawn formation is weak or strong?
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
|
| |
Date: 19 Sep 2007 02:23:05
From: Sanny
Subject: Re: Can you tell me about pawn structure?
|
> > Every month prize is given to toppers. Zebediah plays with master
> > level. By winning 12 games he gets a score of 3*12=36
>
> Yes, but perhaps he is not disconnected five hundred
> times during those games. Right now I am playing the
> Easy level and your program flatly refuses to make a
> move. Here is the displayed time so far: 4130 seconds
> and counting. I already tried restarting in a fresh
> window.
Yes there was a bug, now removed which was telling it to think longer.
Please refresh your page ort better restart your computer to reload
the new program which will never think so long.
So now on higher levels will not take much long time. Try the game
again with easy level and let me know the problem has gone or not.
Bye
Sanny
Play Chess at:http://www.getclub.com/Chess.html
|
| |
Date: 19 Sep 2007 00:52:11
From: help bot
Subject: Re: Can you tell me about pawn structure?
|
On Sep 19, 2:02 am, Sanny <[email protected] > wrote:
> > > Now I see you have to struggle quite hard to beat even the beginner
> > > level. Play 5 games with Easy Level and lets see how many of them you
> > > are able to win.
>
> > Okay, I admit it: in addition to the fact that it takes
> > half the night to play one game at the decent levels,
> > there is also the fact that your fastest levels have
> > higher ratings by far than the much slower ones, so
> > there is no incentive for me to torture myself very often.
>
> Every month prize is given to toppers. Zebediah plays with master
> level. By winning 12 games he gets a score of 3*12=36
Yes, but perhaps he is not disconnected five hundred
times during those games. Right now I am playing the
Easy level and your program flatly refuses to make a
move. Here is the displayed time so far: 4130 seconds
and counting. I already tried restarting in a fresh
window.
> While you play with beginner level, Even if you win 30 games with
> beginner your score for the month will be 1/3 * 30 =10.
Right. But at least I am able to finish games against
the Beginner level; unfinished games net zero points,
and if I die of old age waiting for the program to move
I will lose rating points. ; >D
> So every month Highest score will be of person who beats higher level.
>
> Beginner : 0.3 for each game you win
> Easy: 1.0 for each game you win
> Normal 2.0 for each game you win
> Master: 3.0 for each game you win
>
> Bye
> Sanny
>
> Play Chess at:http://www.getclub.com/Chess.html
In reality, beating the higher levels takes a lot of
time, so it is probably simpler to beat one of the
middle levels repeatedly instead. For example, if
Zebediah were to start a game against the Master
level at the same time as I started one against the
Easy level, he would still be in the opening when I
deliver checkmate and begin game #2.
-- help bot
|
| |
Date: 19 Sep 2007 00:02:22
From: Sanny
Subject: Re: Can you tell me about pawn structure?
|
> > Now I see you have to struggle quite hard to beat even the beginner
> > level. Play 5 games with Easy Level and lets see how many of them you
> > are able to win.
>
> Okay, I admit it: in addition to the fact that it takes
> half the night to play one game at the decent levels,
> there is also the fact that your fastest levels have
> higher ratings by far than the much slower ones, so
> there is no incentive for me to torture myself very often.
Every month prize is given to toppers. Zebediah plays with master
level. By winning 12 games he gets a score of 3*12=36
While you play with beginner level, Even if you win 30 games with
beginner your score for the month will be 1/3 * 30 =10.
So every month Highest score will be of person who beats higher level.
Beginner : 0.3 for each game you win
Easy: 1.0 for each game you win
Normal 2.0 for each game you win
Master: 3.0 for each game you win
Bye
Sanny
Play Chess at: http://www.getclub.com/Chess.html
|
| |
Date: 18 Sep 2007 15:48:01
From: help bot
Subject: Re: Can you tell me about pawn structure?
|
On Sep 18, 11:52 am, Sanny <[email protected] > wrote:
> Out of Last 10 games you played 9 games with beginner and only 1 with
> easy level. If you had played with Easy & Normal levels there would
> have been a few more defeats.
Quite possibly hundreds of them. Your program is
far too strong for me, so its *slowness* has nothing to
do with why I choose to duck the Advance and Master
levels, nosireeBob!
> Now I see you have to struggle quite hard to beat even the beginner
> level. Play 5 games with Easy Level and lets see how many of them you
> are able to win.
Okay, I admit it: in addition to the fact that it takes
half the night to play one game at the decent levels,
there is also the fact that your fastest levels have
higher ratings by far than the much slower ones, so
there is no incentive for me to torture myself very often.
> Beginner plays a bit weaker moves so you get ahead of
> it.
Ah, so that's it. I thought maybe it was psychology. ; >D
I am playing these games on a notebook computer
using wireless internet access, and there are numerous
disconnects during play. If I happen to be playing, say,
Advance level and it takes an hour to move but then
disconnects, I have to reconnect and then wait yet
another hour for it to come up with that same move.
This is very annoying.
Let me just state that even though the program has
serious problems with tactics, as clearly shown
when it is being mated and puts up little resistance,
it does on occasion play good chess and I feel that
I must play accurately to stay on top. While others
continue with their heckling, I alone seem to have
noticed that things aren't always quite so easy as
they make it sound here. Part of the reason for that
may be that a few of them are using computers, and
thus have not "felt" the improvement as Fritz/Rybka
are at the top grandmaster levels -- out of range, so
to speak.
A few of these people want to claim that GetClub
is still at the USCF Class E level, but that is not
the reality. While it still can blunder like such a
player, it can also play well at times, and my guess
is that it has reached the "club player" status on its
higher levels, Master and Advanced. But it will never
achieve true "Expert" status until the shallow tactics
have been fully mastered. No way can a program
which consistently allows itself to be checkmated in
two moves instead of five or six be taken very seriously,
even if it pulls off a few upsets, now and then.
-- help bot
|
| |
Date: 18 Sep 2007 15:29:08
From: help bot
Subject: Re: Can you tell me about pawn structure?
|
On Sep 18, 8:18 am, Mihai Suba <[email protected] > wrote:
> No more slice quots (at least from me), so I can write: "I'm not an
> idiot", without the fear to get commented: 'The author says: "I'm" and
> later:"an idiot"!
>
> Most of the human chess player's strength does not stay in the
> conscious mind which can only deal with calculation, remembering
> things and perhaps a few more activities. All other stuff lays in his
> subconscious mind ready to be retrieved somehow. According to experts
> the subconscious takes 90-95% of total brain capacity.
> Depending how the latter is organized, can assimilate etc., the
> efficiency of what we retrieve can change a lot. We can hardly explain
> what's there, the same as we can hardly explain most of our dreams. In
> our memory, we may have bad examples, incomplete or mistaken rules,
> wrongly associated concepts, etc. which we cannot forget, meaning that
> the concient can call the subconscient but not command it. In a
> program is much simpler, apart from quick calculation, 100% focus on
> subject, 100% health and shape, 0% mania and so on, you can just
> update the information, delete the redundant, wrong or even
> contradictory information etc. We can put or leave the queen "en
> prise" and get mated in one, just to mention the extremes, but the
> program shouldn't. Bad omen if our friend's program does it, but a
> simple checking-against-blunders function will solve it. The problem
> is not tactical calculation (that's what computers and programs are
> for) but, the length of variations (horizon) and the evaluation stick
> at their end. As I said before, and gave a link to an article, the
> evaluation function is the most important, as it helps trimming the
> exponential tree of variations and choosing the main lines to go, thus
> getting a larger "horizon"; moreover, the labels stick at the end of
> those "main lines" be more trustworthy. BTW, the idea of a secondary
> function to qualify the "trustworthyness" or stability of the
> evaluation looks great to me (or was it already done?). This way the
> ridiculous jumps in evaluation (-10.10 to White mates in 8 -BTW did
> you proof it?)
>
> If Sunny lost confidence in his program and uses it as a Kempelen's
> Turk (I saw you are playing some little money out there!), it means we
> are losing our time.
I have read technical stuff, by programmers such as
Robert Hyatt, which talked about pruning techniques
and all sorts of tricks to speed up the search. But
as far as I can see, Sanny has yet to master even
the most basic concepts; actually, it is supposedly
his "team" of programmers who are lacking here.
IMO, the single most important step, apart from
screening out illegal moves and knowing something
about material valuation, is implementing tactical
search extensions so that you don't get hung up at
say, five plys and hang a piece. The program needs
these extensions so it can look to the end whenever
there are exchanges or checks, and not get chopped
off at some random point in the middle. Okay, if the
checks never end you must cut it off somewhere, but
normally there will be a series of piece and pawn
exchanges no deeper than ten or so plies, and
getting this right is crucial, just as it is for human
players. Having played his program countless
times and read him explaining what went wrong
afterward, it is obvious that in terms of tactics the
program can stoop as low as USCF 200! : >D
So when Sanny comes and asks how to implement
the subtleties of Pawn Structure Chess by Andy Soltis
or Pawn Power in Chess by Hans Kmoch, I can only
sigh, and shake my head in disbelief. It is akin to a
USCF 1000 player asking to learn the main line of the
Najdorf Sicilian twenty moves deep, so he can take on
the local Experts.
-- help bot
|
| |
Date: 18 Sep 2007 20:04:21
From: Macchess
Subject: Re: Can you tell me about pawn structure?
|
Sanny a �crit :
> What type of Pawn structures are good and what type of pawn structure
> are bad?
Sanny,
Could you avoid cross posting to this newsgroup? You are polluting.
Your questions is a good question, but it does belong more in the
rec.games.chess.analysis group than it does here.
Nevertheless, you should consult some chess theory about it. It all
depends what level you play on. If you are around the level of your
program, I would say that you should first improve on tactics by
searching deeper. As a clue, unless you beat GNUChess, tactics remain
most important.
At a level of 1700 (not your score, but FIDE elo) pawn structure starts
to be important. At levels of 2000 and above (again FIDE) in serious
games (2h/40 moves), pawn structure often decides the game.
BTW, I wonder what structure your program has. AB searching with
null-move? Stress on evaluation function or stress on depth? How deep
are you following exchanges and how wide is the tree you search?
Best regards,
Yves
http://macchess.internetcontact.be
|
| |
Date: 18 Sep 2007 09:52:14
From: Sanny
Subject: Re: Can you tell me about pawn structure?
|
> Recently, I have gotten the general impression that
> I am playing a human opponent at GetClub, not a
> computer. It's all booked up in some opening lines,
> and makes human-like moves while still overlooking
> obvious tactics on occasion -- tactics it never would
> have missed a few months ago. There are also a lot
> of disconnects during play. The one thing which
> argues against a human behind the curtain is stupid
> Rook moves; computers are notorious for these, or
> at least the very weak ones are. When Sanny comes
> here and says "try to beat the program NOW" I wonder
> if he is pulling off some trick, but every time it is still
> beatable.
Out of Last 10 games you played 9 games with beginner and only 1 with
easy level. If you had played with Easy & Normal levels there would
have been a few more defeats.
Now I see you have to struggle quite hard to beat even the beginner
level. Play 5 games with Easy Level and lets see how many of them you
are able to win. Beginner plays a bit weaker moves so you get ahead of
it.
Bye
Sanny
Play Chess at: http://www.getclub.com/Chess.html
|
| |
Date: 18 Sep 2007 13:18:43
From: Mihai Suba
Subject: Re: Can you tell me about pawn structure?
|
No more slice quots (at least from me), so I can write: "I'm not an
idiot", without the fear to get commented: 'The author says: "I'm" and
later:"an idiot"!
Most of the human chess player's strength does not stay in the
conscious mind which can only deal with calculation, remembering
things and perhaps a few more activities. All other stuff lays in his
subconscious mind ready to be retrieved somehow. According to experts
the subconscious takes 90-95% of total brain capacity.
Depending how the latter is organized, can assimilate etc., the
efficiency of what we retrieve can change a lot. We can hardly explain
what's there, the same as we can hardly explain most of our dreams. In
our memory, we may have bad examples, incomplete or mistaken rules,
wrongly associated concepts, etc. which we cannot forget, meaning that
the concient can call the subconscient but not command it. In a
program is much simpler, apart from quick calculation, 100% focus on
subject, 100% health and shape, 0% mania and so on, you can just
update the information, delete the redundant, wrong or even
contradictory information etc. We can put or leave the queen "en
prise" and get mated in one, just to mention the extremes, but the
program shouldn't. Bad omen if our friend's program does it, but a
simple checking-against-blunders function will solve it. The problem
is not tactical calculation (that's what computers and programs are
for) but, the length of variations (horizon) and the evaluation stick
at their end. As I said before, and gave a link to an article, the
evaluation function is the most important, as it helps trimming the
exponential tree of variations and choosing the main lines to go, thus
getting a larger "horizon"; moreover, the labels stick at the end of
those "main lines" be more trustworthy. BTW, the idea of a secondary
function to qualify the "trustworthyness" or stability of the
evaluation looks great to me (or was it already done?). This way the
ridiculous jumps in evaluation (-10.10 to White mates in 8 -BTW did
you proof it?)
If Sunny lost confidence in his program and uses it as a Kempelen's
Turk (I saw you are playing some little money out there!), it means we
are losing our time.
|
| |
Date: 18 Sep 2007 02:40:17
From: help bot
Subject: Re: Can you tell me about pawn structure?
|
On Sep 17, 9:05 am, Mihai Suba <[email protected] > wrote:
> All this pile-up controversy is nonsense, as the subject was to help
> (or discourage) a chess programmer.
> Hitler could have learned from Napoleon, thanks for the comparison
> anyway.
> 'When the writer asks "why not move X",
> where move X is the move preferred by Fritz or Rybka,
> they get really annoyed and have no substantive
> answer; in essence, the answer amounts to "because
> I am a GM, and I say so, nyaaaaaaaaa!" LOL'
> It seems the right thing to do when the discussion turns into a veg-
> ket-like:
> "-Good morning, ma'm!
> -Plums, dear, plum, plum!"
> Ask Fritz and Rybka then, and apart from some lines, far less
> deciphrable than a GM comment,
Looking at Chess Life (maybe you get a far better
magazine) where a writer asks the winner why he
played X or does not like Fritz's moves: the replies
are gibberish; jumbled lines of raw calculation with
wrongheaded evaluations, and it's all a mess. Far
from getting the good stuff from 2600+ grandmasters
or from the anno-Fritzators, we are getting rubbish
from writers who *are* asking both Fritz AND the
GMs. As far as I can see, the sole exception was
GM Suzan Polgar with Paul Truong, and they seem
to have asked no one anything, thinking for
themselves (except regarding /why/ BF played
...Bxh2 against GM Spassky).
In one case the losing GM had a far better grasp
of what was going on, which seems rather odd; he
was more objective, but not quite as good at
calculating variations perhaps.
As for the decipher-ability of human commentary,
my latest go-'round in this arena found some odd
comments by 2300+ Taylor Kingston praising
Fritz for what he interpreted as weaving a mating
net; but my computer seemed to conclude that
Fritz could not see a mate-in-six to save its own
life; that it was floundering and the defense deserved
some credit for the catch. I know that sometimes it
will miss a better move, but never does it annouce
mate-in-6 when there is no such mate. The claim
that Fritz was weaving a net was simply debunked by
its utter inability to see a forced mate at a *far lesser*
depth than the one in the actual game.
> you will get an answer which, in
> essence, ammounts to: "because
> I am Fritz or Rybka, and I say so, bu(from 'Bucarest')uuuuuuuu!
Fritz always backs his evaluations with concrete
analysis, even if it may be totally wrong.
One way to check this is by slowly stepping forward
along its main variation to see what changes.
But the key is that if the contempt factor is zero,
Fritz is 100% objective; he never tries harder for one
side than for the other (and he never gives lines
where one side moves thrice in a row).
> Moreover I am distributed and provided 'AS IS' and with no warranties
> of any kind, whether express or implied, and I can change my
> materialistic mind from '-10.10' to 'White mates in 8' at any moment,
> without previous advice, as you can proof with the miner's position."
> Beware the laws in your country, though, before killing the
> annotators. In mine, for instance, evidence given by Fritz or Rybka is
> not valid!
From what I have seen, you seem to be bashing the
preferred tool of nearly every annotator in Chess Life.
You can tell the one or two who don't use computers,
by their occasional horrific tactical oversights, and by
their ability to explain every "why" in simple English.
> Many times, during important championships, corr. chess fans came,
> with their folders of active games, and gathered master moves from one
> or another. After IM Ghizdavu suggested a move, the other guy asked
> "why?" That was not a mate-in-2 problem, how could one explain, in a
> few words, his long years of learning and practice. No way, even to
> himself! "Dorelius" found the answer: "Do you want to win the game?
> than play this move!
> PS the "miner's position" can be found inhttp://menssana.co.uk/eng/pgns.htm
I think correspondence chess is much higher in quality
than OTB chess, and the many articles I have read which
portray openings theory as being shaped by OTB are
either misguided or else the OTB-theory itself is suspect.
Yet there are some works which have been cut down by
the book reviewers; for instance the one by Hans Berliner;
possibly he makes ludicrous, unsupportable claims.
Still, if you cannot explain in words or in variations why
you must be winning, the impression is that you *think*
you are winning, just as most people think they are
above average drivers, more intelligent than average, etc.
And when your analysis gives multiple moves in-a-row
for your side while pretending the opponent is frozen like
a block of ice, it lends an impression of a lack of true
objectivity. As you say, if you want to win, play this
move; and as I say, if you want objectivity, good luck
finding any amongst human chess players. : >D
I tried to play your system against GetClub tonight,
and things were a lot tougher than I had expected;
I had nice pawns, as you said, and he had little in
the way of counterplay, but managed to put up a tough
defense and I lost my advantage some time after
trading Queens; of course the Rook ending was easy
because of numerous mistakes by the program, so I
won anyway.
What Sanny needs is technical help in implementing
tactical search extensions, for when there are checks
and captures. His program *still* does not foresee any
coming mates on itself, so no matter how deep it may
be looking selectively, it has a severe weakness which
can easily be exploited, again and again. Just tonight,
I checkmated the program on Beginner level, which
means it did not foresee mate-in-one in time to resign.
Recently, I have gotten the general impression that
I am playing a human opponent at GetClub, not a
computer. It's all booked up in some opening lines,
and makes human-like moves while still overlooking
obvious tactics on occasion -- tactics it never would
have missed a few months ago. There are also a lot
of disconnects during play. The one thing which
argues against a human behind the curtain is stupid
Rook moves; computers are notorious for these, or
at least the very weak ones are. When Sanny comes
here and says "try to beat the program NOW" I wonder
if he is pulling off some trick, but every time it is still
beatable.
-- help bot
|
| |
Date: 17 Sep 2007 14:05:53
From: Mihai Suba
Subject: Re: Can you tell me about pawn structure?
|
All this pile-up controversy is nonsense, as the subject was to help
(or discourage) a chess programmer.
Hitler could have learned from Napoleon, thanks for the comparison
anyway.
'When the writer asks "why not move X",
where move X is the move preferred by Fritz or Rybka,
they get really annoyed and have no substantive
answer; in essence, the answer amounts to "because
I am a GM, and I say so, nyaaaaaaaaa!" LOL'
It seems the right thing to do when the discussion turns into a veg-
ket-like:
"-Good morning, ma'm!
-Plums, dear, plum, plum!"
Ask Fritz and Rybka then, and apart from some lines, far less
deciphrable than a GM comment, you will get an answer which, in
essence, ammounts to: "because
I am Fritz or Rybka, and I say so, bu(from 'Bucarest')uuuuuuuu!
Moreover I am distributed and provided 'AS IS' and with no warranties
of any kind, whether express or implied, and I can change my
materialistic mind from '-10.10' to 'White mates in 8' at any moment,
without previous advice, as you can proof with the miner's position."
Beware the laws in your country, though, before killing the
annotators. In mine, for instance, evidence given by Fritz or Rybka is
not valid!
Many times, during important championships, corr. chess fans came,
with their folders of active games, and gathered master moves from one
or another. After IM Ghizdavu suggested a move, the other guy asked
"why?" That was not a mate-in-2 problem, how could one explain, in a
few words, his long years of learning and practice. No way, even to
himself! "Dorelius" found the answer: "Do you want to win the game?
than play this move!
PS the "miner's position" can be found in http://menssana.co.uk/eng/pgns.htm
|
| |
Date: 16 Sep 2007 19:43:24
From: help bot
Subject: Re: Can you tell me about pawn structure?
|
On Sep 16, 6:13 am, Mihai Suba <[email protected] > wrote:
> help bot sais:
> " Without any opening book, chess programs have glaring
> weaknesses. One of the most obvious is a tendency to
> shut in their own Queen by blocking the QB-pawn with the
> Knight."
> Correct, and finding other tendencies like that would improve any
> program.
Apparently, most programmers work around this by
utilizing a by-rote openings book in which their program
may not like the book moves, but will play it just the
same. It's only because I do not have these books
installed that I noticed this glaring weakness.
> After quoting "1.e3! g6 2.f4 Nf6?!"
> help bot sais:
> "What is wrong with this move? Playing ...g6 in no way
> obligates one to follow up *immediately* with Bg7."
> In fact, I should place a "?" because allowing White to oppose bishops
> on long diagonal is, for practical purposes, diminishing the mobility
> of two pieces.
The Knight on f6 and the Bishop on g7 will not want to
move because of the exchange BxB; yet White is going
to a lot of trouble over this by giving up his KB and by
sliding his Queen the wrong way with Q-c1. IMO, your
exclams are unwarranted, except perhaps as pointing
out where you anticipated a potential glaring weakness
of the program; for instance, no way is 1.e3 worthy of an
exclam, objectively speaking.
> The more so, if the game were against a human, because it's idiotic to
> prevent "3.e4" after White played "1.e3"
Yet this is the correct way to program a computer;
you can't tell the computer "he won't do this, because
it might make him feel stupid for an earlier move". The
computer must remain objective, apart from any
contempt factor that is; this factor is a tool for Rybka
to avoid making too many draws, but I don't know if it
helps or harms the program's results.
> Later:
> "3.b3! Bg7 4.Bb2 O-O 5.Nf3 Nc6?! 6.Bb5!
> Why is this move so good? After ...Nb8 the Bishop
> is misplaced, IMO."
> Your O is correct, but the bishop won't go back to "f1" hence, 1 tempo
True, one tempo -- with the understanding that it must
be used for a retreat, and that Black may, or may not,
attack the Bishop at his convenience and by any means
he desires. In sum, if a tempo were worth a third-point,
you would gain maybe one sixth point here; how does
that compare to the best move other than B-b5, I wonder?
> gain.
> Later:
> "a6?
> Obviously, Rybka, just as a few of its predecessors
> -- including GM Fischer -- is programmed to prefer
> Bishops to Knights. To humans, this is almost
> incomprehensible, yet we do it with Rooks vs. Bishops
> and think nothing of it, calling it winning the exchange."
> Classics called B vs N = "small exchange", but a rook is much better
> than a bishop far more than a bishop vs. knight.
Right; yet the program doesn't care if its calculations
show a mathematical gain; it will win your Queen; it will
win the exchange; it will win a pawn; it will take the
minor exchange -- just so long as it sees inadequate
compensation;and here it seems to value the Bishop
over a Knight more so than it minds doubled pawns.
As I recall, the program "M-Chess" held the Bishop
as a half-point better than a Knight, and as a result it
would often beat programs which played very
aggressively, because they did not like having to
retreat Bishops with a loss of time. Even so, there
were a few games where this hyper-preference would
become M-chess's undoing, and it would land with a
bad Bishop in the ending.
> Generrally a B is better than a N due to endgame pairing up: B+B > B
> +N, B+B > N+N, R+B > R+N, R+B+B ~ R+R+N while only Q+N > Q+B.
> With a damaged PAWN STRUCTURE, the superiority of the bishop is very
> questionable.
In one scenario, Black immediately pins the annoying
White Knight with ...Bg4, and later chops it off. If White
is very careful to avoid this, the Bishop is still an active
piece. The trick is to read your mind and foresee the
crushing f5 move. : >D
My game had my pawn structure wrecked in the normal
Nimzowitch type manner, but here things aren't quite so
easy. I should probably mention that my opponent, some
hundreds of rating points higher than me, got it in his head
that the game was already decided, as he no doubt read
in some book on this opening; things weren't quite that
simple though, because I "knew" that I was in big trouble
while he did not realize that I am *very* experienced in
having bad positions!
> Later:
> "7.Bxc6 dxc6? between +/- and +-.
> Come on now; just because you won does not
> mean that you were winning all the way back to
> here. White has the e5 square, but Black has the
> d-file, a pair of Bishops, and no other glaring
> weakness to speak of. I think Rybka may have
> erred later on to lose this."
> Not winning (the symbols say between clearly and much better) and
> Ribka hardly errs "later on".
Why are we not allowed to see the rest of the game
then? Are we to assume that if you say no mistakes
were made, that you are infallible or something? : >D
In many annotated games in Chess Life, even the
winner did not seem to comprehend what was going
on in complex middle games, judging by their own
comments. When the writer asks "why not move X",
where move X is the move preferred by Fritz or Rybka,
they get really annoyed and have no substantive
answer; in essence, the answer amounts to "because
I am a GM, and I say so, nyaaaaaaaaa!" LOL
> Most times "she" crushes me from
> theoretical positions, irrespective of evaluation,
I replayed Rybka's games against several programs
and her play was fantastic, except against ZapChess.
In the opening, Rybka generally came out on top, and
in the middle game, she seemed to always have the
safer King -- except when she had grabbed material
which was normally converted to a win after some
impressive defensive gymnastics.
> "She" erred quite a
> lot in concept by now and was uncapable to understand the position and
> elaborate a long term plan of defence.
Generally, at normal time controls a program will
not anticipate a long-range attack but instead will
accidentally thwart it by very active piece play in
another sector. The stronger the program, the
more likely this counter play will come fast and
hard enough to thwart the human's "plan".
> Later:
> "White's plan is O-O, Qc1
> (development) d3, e4 (centre) Nbd2, Nc4 and possibly Ne3 (improving
> general energy, dynamic potential I call it) and finally f5! Even
> her diabolic calculation could not overcome the strategical defects
> of the position, that is, pawn structure and relative piece
> positioning.
> Okay, so let's see how you did all that while Rybka
> just twiddled her thumbs.
> Sorry, I neither keep evidence, nor remember all games against
> programs, do you?
No. But most of my games are recorded by the Web
site on which they are played. I have games at GetClub,
RedHotPawn (my losses here /could/ be to computers,
but certainly none of my wins were), and besides which
I am just a weak player, so I only recall bits and pieces
of my games. ; >D
I believe my Fritz program automatically records my
games against it, whether I want it to or not (I don't
like having so many quick losses recorded). I have
not played Rybka, and until I can figure out how to
remove her Queen, I don't see much point. It goes
without saying that without her Queen, the program
would not only be deprived of her openings book, but
also exhibit the same sort of problems seen in the
odds match against GM Benjamin: 1.e3 (frees the
program's Bishop) ...e6 (frees my Bishop and *also
my Queen*, ha ha!).
> I only recall that, in other games of the session, she answered 1.e3
> e5 2.d4 e4 which is also dubious, but not refutable.
Please notify John Watson; even with move in hand,
the White side of the French Defense is considered
irrefutable! Take that, stubborn French players.
> Later:
> "BTW, somewhere in here
> the move ...Bg4 can pin the e5-loving Knight on f3: a
> real downer, if you plan on attacking with minor pieces."
> The idea to get rid of a piece which can easily become a "big pawn" or
> a "stone" is not bad but requires understanding and admitting
> inferiority (contradicts the 2B stuff mentioned before)
To the contrary, rather than admitting that *correct*
play inevitably leads to a bad QB, I am simply taking
your assertion that you have a crushing attack and
showing one small step by which it might be taken
back a peg, slowed down. Of course, you can still
claim to have an unstoppable heavy piece assault in
the works, but if you allow the pin ...Bg4, your minors
are cut in number by half.
> After, say: 8.O-O Bg4 9.Qc1 Bxf3 10.Rxf3 White still has the same plan
> in centre and kingside, "f5" is in the air and the rook on f3,g3,h3 is
> helpful. I don't see Black's counterplay
This is precisely what Hitler said when he invaded
Russia! (Who suspected that mud could attack his
tanks? That winter could cripple his troops?)
In my experience (so far), only ZapChess has been
able to consistently "see" the dangers presented by
Rybka. And the odd thing is that Rybka has more
than one possible setting, so maybe one of the more
tactical ones would do better against Zap.
> and the plan with "...f5"
> described below is more questionable (now White has superiority on
> both black and white squares)
Yeah, but you give White several free moves and
assume that Black will do nothing. Maybe the time
control was quick and Rybka did nothing; that just
shows you nixed tricky tactics and got a strategical
win through the eye of the needle; it does not prove
that you are objectively winning after Bxc6. In fact,
it appears to indicate that this is one of those types
of positions where computers are relatively weak.
> Than:
> "A good human player would have find the toughest defence
> and possibly counterplay: "Ne8, f5 (freezing White's centre, improving
> king's safety) c5, b5, c4 (white squares).
>
> That is an excellent description of how I handled
> this as Black against a superior player many years
> ago He lost due to overconfidence; recklessness,
> even though my position was a bit worse than this
> one."
> I begin to understand you!
> Than:
> "I would reformulate help
> bot's, 'The reason is simply that even if you got your program
> to play like a grandmaster in pawn-handling, its results
> would barely budge since tactics are 90+% of the game.' changing
> 'tactics' by 'tactics + evaluation+long term planning' Fischer beat
> Spassky because he was stronger in most technical aspects (opening
> preparation, calculation speed, endgame, shape and psichology -
> including motivation, will power, boldness, etc.) but Spassky had more
> IDEAS.
>
> That was just one match. Before the 1972 match, GM
> Spassky had never lost even a single game to GM Fischer!
> Psychology seems to have been a key player."
> Right, except from a few games e.g, one 'poisoned pawn' (Nb1!) and two
> Sicilians (all in the match), Spassky always got worse after the
> opening, as Fisher's preparation was outstanding (although he was the
> last winner of "World's INDIVIDUAL Chess Championship").
GM Karpov (among others) has attributed this to
GM Spassky's extreme laziness. In fact, he also
claims to have beaten reigning champion Spassky
in some training games just before the 1972 match.
I never saw any of those games, either; how did
GM Spassky lose? Did they work on "un-busting"
the King's Gambit, I wonder?
> Before the
> match, Spassky scored 2,5-0,5 from 3 lost positions. About 6 weeks ago
> I was lucky enough to be in his company for a few days (he and
> Vaganian were invited in Romania, as honour directors of Romgaz Eforie
> Open). His favourite motto is Tartakover's: "Combination is a victory
> of spirit against matter".
At my level, combination is what inevitably happens
when I was winning, but got into time pressure. "I had
him", "he got lucky", and my favorite, "I was easily
winning before this blunder, where I gave you the
game". Of course, next time we come back and do
the same sort of thing again.
> Than:
> "There are several books which deal with how to handle
> the pawns in a specific opening, and some which deal
> with pawn handling in general, but at the level I have
> seen at GetClub, this is largely irrelevant. Example: the
> program cannot see a mate-in-two coming, and instead
> of averting it walks right into it. I am talking about a game
> against the Advance level -- the slowest, and strongest
> level there is. This clearly indicates that any work on
> pawn handling will for the most part be wasted in so far
> as results go.
> -- help bot- "
> That's right, the greater the victory if we can help the author to
> improve these results.
> We must encourage him to improve his own chess erudition and wisdom,
> before trying to translate them to the machine.
> Good luck again in your games and/or programming!
According to his own account, Sanny has a "team"
of programmers and does not himself write the code.
Most likely, these folks also write the HTML for his
Web site and it, too, is buggy. But the program has
improved in some areas, and at times can play good
moves for quite a stretch. It is because of these
good moves in complex positions that I find it so
ludicrous that his program cannot avert a mate-in-two.
-- help bot
|
| |
Date: 16 Sep 2007 18:02:17
From: help bot
Subject: Re: Can you tell me about pawn structure?
|
On Sep 15, 2:26 am, Sanny <[email protected] > wrote:
> I saw you played 4 games yesterday all with beginner Level, You win 3
> Games and one was Drawn.
>
> I Challenge you, that you will not be able to beat Easy and Normal
> Levels as they will play very strong moves. Why play with weak Level.
> If you want tough competition play with Easy & Normal Levels.
>
> They will beat you 2/4 times. (As per my Calculations)
That is an interesting formula. By my calculations, a
mouse error whereby I accidentally promote a pawn to
Knight and thus draw a won game simply because the
button for "Knight" is right below the one for "Queen"
and I am in blitz mode, is very, very rare; in fact, I
estimate that it may not happen again until after the
reappearance of Halley's comet! LOL
-- help bot
|
| |
Date: 16 Sep 2007 11:13:02
From: Mihai Suba
Subject: Re: Can you tell me about pawn structure?
|
help bot sais:
" Without any opening book, chess programs have glaring
weaknesses. One of the most obvious is a tendency to
shut in their own Queen by blocking the QB-pawn with the
Knight."
Correct, and finding other tendencies like that would improve any
program.
After quoting "1.e3! g6 2.f4 Nf6?!"
help bot sais:
"What is wrong with this move? Playing ...g6 in no way
obligates one to follow up *immediately* with Bg7."
In fact, I should place a "?" because allowing White to oppose bishops
on long diagonal is, for practical purposes, diminishing the mobility
of two pieces.
The more so, if the game were against a human, because it's idiotic to
prevent "3.e4" after White played "1.e3"
Later:
"3.b3! Bg7 4.Bb2 O-O 5.Nf3 Nc6?! 6.Bb5!
Why is this move so good? After ...Nb8 the Bishop
is misplaced, IMO."
Your O is correct, but the bishop won't go back to "f1" hence, 1 tempo
gain.
Later:
"a6?
Obviously, Rybka, just as a few of its predecessors
-- including GM Fischer -- is programmed to prefer
Bishops to Knights. To humans, this is almost
incomprehensible, yet we do it with Rooks vs. Bishops
and think nothing of it, calling it winning the exchange."
Classics called B vs N = "small exchange", but a rook is much better
than a bishop far more than a bishop vs. knight.
Generrally a B is better than a N due to endgame pairing up: B+B > B
+N, B+B > N+N, R+B > R+N, R+B+B ~ R+R+N while only Q+N > Q+B.
With a damaged PAWN STRUCTURE, the superiority of the bishop is very
questionable.
Later:
"7.Bxc6 dxc6? between +/- and +-.
Come on now; just because you won does not
mean that you were winning all the way back to
here. White has the e5 square, but Black has the
d-file, a pair of Bishops, and no other glaring
weakness to speak of. I think Rybka may have
erred later on to lose this."
Not winning (the symbols say between clearly and much better) and
Ribka hardly errs "later on". Most times "she" crushes me from
theoretical positions, irrespective of evaluation, "She" erred quite a
lot in concept by now and was uncapable to understand the position and
elaborate a long term plan of defence.
Later:
"White's plan is O-O, Qc1
(development) d3, e4 (centre) Nbd2, Nc4 and possibly Ne3 (improving
general energy, dynamic potential I call it) and finally f5! Even
her diabolic calculation could not overcome the strategical defects
of the position, that is, pawn structure and relative piece
positioning.
Okay, so let's see how you did all that while Rybka
just twiddled her thumbs.
Sorry, I neither keep evidence, nor remember all games against
programs, do you?
I only recall that, in other games of the session, she answered 1.e3
e5 2.d4 e4 which is also dubious, but not refutable.
Later:
"BTW, somewhere in here
the move ...Bg4 can pin the e5-loving Knight on f3: a
real downer, if you plan on attacking with minor pieces."
The idea to get rid of a piece which can easily become a "big pawn" or
a "stone" is not bad but requires understanding and admitting
inferiority (contradicts the 2B stuff mentioned before)
After, say: 8.O-O Bg4 9.Qc1 Bxf3 10.Rxf3 White still has the same plan
in centre and kingside, "f5" is in the air and the rook on f3,g3,h3 is
helpful. I don't see Black's counterplay and the plan with "...f5"
described below is more questionable (now White has superiority on
both black and white squares)
Than:
"A good human player would have find the toughest defence
and possibly counterplay: "Ne8, f5 (freezing White's centre, improving
king's safety) c5, b5, c4 (white squares).
That is an excellent description of how I handled
this as Black against a superior player many years
ago He lost due to overconfidence; recklessness,
even though my position was a bit worse than this
one."
I begin to understand you!
Than:
"I would reformulate help
bot's, 'The reason is simply that even if you got your program
to play like a grandmaster in pawn-handling, its results
would barely budge since tactics are 90+% of the game.' changing
'tactics' by 'tactics + evaluation+long term planning' Fischer beat
Spassky because he was stronger in most technical aspects (opening
preparation, calculation speed, endgame, shape and psichology -
including motivation, will power, boldness, etc.) but Spassky had more
IDEAS.
That was just one match. Before the 1972 match, GM
Spassky had never lost even a single game to GM Fischer!
Psychology seems to have been a key player."
Right, except from a few games e.g, one 'poisoned pawn' (Nb1!) and two
Sicilians (all in the match), Spassky always got worse after the
opening, as Fisher's preparation was outstanding (although he was the
last winner of "World's INDIVIDUAL Chess Championship"). Before the
match, Spassky scored 2,5-0,5 from 3 lost positions. About 6 weeks ago
I was lucky enough to be in his company for a few days (he and
Vaganian were invited in Romania, as honour directors of Romgaz Eforie
Open). His favourite motto is Tartakover's: "Combination is a victory
of spirit against matter".
Than:
"There are several books which deal with how to handle
the pawns in a specific opening, and some which deal
with pawn handling in general, but at the level I have
seen at GetClub, this is largely irrelevant. Example: the
program cannot see a mate-in-two coming, and instead
of averting it walks right into it. I am talking about a game
against the Advance level -- the slowest, and strongest
level there is. This clearly indicates that any work on
pawn handling will for the most part be wasted in so far
as results go.
-- help bot- "
That's right, the greater the victory if we can help the author to
improve these results.
We must encourage him to improve his own chess erudition and wisdom,
before trying to translate them to the machine.
Good luck again in your games and/or programming!
|
| |
Date: 15 Sep 2007 18:27:59
From: help bot
Subject: Re: Can you tell me about pawn structure?
|
On Sep 15, 8:15 am, Mihai Suba <[email protected] > wrote:
> Dear friends,
> Allow me to re-quote from another chess forum:
> "Life is painting a picture, not doing a sum"
> Same happens to chess.
> I remember "Deep Thought" in "World Open", Philadelphia 1987
> When it announced "mate in 19" it came as a thunder and all
> professional players felt the threat!
> A couple of years later I had a discussion with Dr. Nunn about the
> matter. I hold that the program was very strong tactically and very
> weak strategically. (based on its win with Black against Larsen and
> its loss with White against Browne). He hold that, at some point, the
> tactical strength will be so high that strategy would not matter or be
> confounded with it. I could not believe that, during my life, a
> program would beat the best human player. Our attitude was somehow
> corresponding to our "playing styles", mine more logical (in my view),
> based on "strategic intuition" and his (also in my view) more
> gambling, based on "tactical intuition".
> Who was right?
> No one!
> Chess has 4 very important elements: material, space, time and
> threats. (I excluded psichology and shape factors on purpose).
> Programs are biassed more on 1 and 4, while a "positional" human GM on
> 2 and 3.
> The chess strength is to weight these elemens and produce an
> "evaluation". It is not true that programs don't know strategy. Best
> programs have hundred of mini-rules to correct the brute "material
> count" (e.g, "more exposed king") They found many of these rules
> statistically, by playing millions of games against each other. They
> "trim" the tree of variations with some evaluation.
> Programs have huge memory (which is illegal) and adapted opening books
> (also illegal) and yet, big strategy lapses. Here is the opening of
> one in a few games I won against Rybka (the latter without opening
> book)
Without any opening book, chess programs have glaring
weaknesses. One of the most obvious is a tendency to
shut in their own Queen by blocking the QB-pawn with the
Knight.
> : 1.e3! g6 2.f4 Nf6?!
What is wrong with this move? Playing ...g6 in no way
obligates one to follow up *immediately* with Bg7.
> 3.b3! Bg7 4.Bb2 O-O 5.Nf3 Nc6?! 6.Bb5!
Why is this move so good? After ...Nb8 the Bishop
is misplaced, IMO.
> a6?
Obviously, Rybka, just as a few of its predecessors
-- including GM Fischer -- is programmed to prefer
Bishops to Knights. To humans, this is almost
incomprehensible, yet we do it with Rooks vs. Bishops
and think nothing of it, calling it winning the exchange.
> 7.Bxc6 dxc6? between +/- and +-.
Come on now; just because you won does not
mean that you were winning all the way back to
here. White has the e5 square, but Black has the
d-file, a pair of Bishops, and no other glaring
weakness to speak of. I think Rybka may have
erred later on to lose this.
> White's plan is "O-O, Qc1
> (development) d3, e4 (centre) Nbd2, Nc4 and possibly Ne3 (improving
> general energy, dynamic potential I call it) and finally f5! Even
> "her" diabolic calculation could not overcome the strategical defects
> of the position, that is, pawn structure and relative piece
> positioning.
Okay, so let's see how you did all that while Rybka
just twiddled her thumbs. BTW, somewhere in here
the move ...Bg4 can pin the e5-loving Knight on f3: a
real downer, if you plan on attacking with minor pieces.
> A good human player would have find the toughest defence
> and possibly counterplay: "Ne8, f5 (freezing White's centre, improving
> king's safety) c5, b5, c4 (white squares).
That is an excellent description of how I handled
this as Black against a superior player many years
ago He lost due to overconfidence; recklessness,
even though my position was a bit worse than this
one.
> I would reformulate help
> bot's, 'The reason is simply that even if you got your program
> to "play like a grandmaster" in pawn-handling, its results
> would barely budge since tactics are 90+% of the game.' changing
> 'tactics' by 'tactics + evaluation+long term planning' Fischer beat
> Spassky because he was stronger in most technical aspects (opening
> preparation, calculation speed, endgame, shape and psichology -
> including motivation, will power, boldness, etc.) but Spassky had more
> IDEAS.
That was just one match. Before the 1972 match, GM
Spassky had never lost even a single game to GM Fischer!
Psychology seems to have been a key player.
> Rules are good, provided you look at them critically (with
> exceptions feedback). Best (including for our theme - pawn structure)
> is to teach your program patterns (don't ask me how).
> That's all by now, I need some coffee!
> Good luck in your games and/or programming!
There are several books which deal with how to handle
the pawns in a specific opening, and some which deal
with pawn handling in general, but at the level I have
seen at GetClub, this is largely irrelevant. Example: the
program cannot see a mate-in-two coming, and instead
of averting it walks right into it. I am talking about a game
against the Advance level -- the slowest, and strongest
level there is. This clearly indicates that any work on
pawn handling will for the most part be wasted in so far
as results go.
-- help bot
|
| |
Date: 15 Sep 2007 09:46:07
From: samsloan
Subject: Re: Can you tell me about pawn structure?
|
On Sep 14, 2:34 am, Sanny <[email protected] > wrote:
> Till now GetClub Chess only avoid double pawn formation and isolated
> pawns.
> Play Chess at:http://www.GetClub.com/Chess.html
>
> What type of Pawn structures are good and what type of pawn structure
> are bad?
>
If you want to learn the answer to this question, I recommend that you
buy a book called "Pawn Power in Chess" by Hans Kmoch.
http://www.amazon.com/Pawn-Power-Chess-Hans-Kmoch/dp/0486264866/
You can hardly expect members of this group to give you beginner's
lessons.
Sam Sloan
|
| |
Date: 15 Sep 2007 06:48:35
From: help bot
Subject: Re: Can you tell me about pawn structure?
|
On Sep 15, 12:41 am, Sanny <[email protected] > wrote:
> Thats good. I find whenever a pawn is moved a hole will always
> generated on its side. Then how to tell wich hole is bigger danerous
> and which hole can be ignored?
I don't know of any Web sites that cover this, but there
may well be many.
In general, you have to use judgment to determine if a
feature such as a weak square or doubled pawns can be
exploited by the opponent, or if it is irrelevant or more
than compensated for by what is gained in exchange.
The same thing applies in determining whether or not to
exchange pieces, castle on opposite sides, and so forth.
Just as an example, my game where I sac'ed a Rook
for a far-advanced passed pawn and a potential attack
on Black's King (this game was discussed earlier) had
me giving up a lot of material, and the program should
have gone into defense mode, not grabbing the h-pawn
as if it were a "normal" position where material was key
to the outcome. All that mattered was, yes/no: will I get
to its King? So the correct approach was to defend its
King, and *quickly*.
-- help bot
|
| |
Date: 15 Sep 2007 13:15:18
From: Mihai Suba
Subject: Re: Can you tell me about pawn structure?
|
On Sep 14, 8:34 am, Sanny <[email protected] > wrote:
> Till now GetClub Chess only avoid double pawn formation and isolated
> pawns.
> Play Chess at:http://www.GetClub.com/Chess.html
>
> What type of Pawn structures are good and what type of pawn structure
> are bad?
>
> 1. If Pawns are on same rows not protecting each other {in front of
> King}. (Is it good)
> King has no place to move???
>
> 2. All pawns protecting other making a long diagonal Chain. (Is it
> good)
> A pawn Exchange can bring it down to Isolated Pawns???
>
> 3. Two pawns on side protecting middle pawn. (Is it good)
> There is a hole in Center???
>
> How to find weak spots? Whenever a Pawn is played we have toi leave a
> diagonal hole for bishop or a place for opponent Knight.
>
> Can you tell me how to tell which pawn structure is better.
>
> I went through lots of sites to know about Pawn Formations but could
> not find any concrete example.
>
> How to tell whether a Pawn formation is weak or strong?
>
> Bye
> Sanny
>
> Play Chess at:http://www.GetClub.com/Chess.html
Good question, and obviously asked to improve a chess playing
program.
By now, the most relevant answer came from "help bot":
"If I were you, I would ignore finesse work regarding pawn
structures and focus priily on improving tactics.
The reason is simply that even if you got your program
to "play like a grandmaster" in pawn-handling, its results
would barely budge since tactics are 90+% of the game.
Many times I have replayed some game in Fritz, and
no matter what the pawn structure was, no matter who
had the superior minor piece, no matter who controlled
the center, no matter which side was positionally
better, someone blundered and Fritz instantly spotted
the tactical refutation -- game over.
---------------------------------------------------------------------------=
=AD----
But if you must know, here are a few ideas...
Good:
a) pawns connected to one another (a2, b2, c2)
b) pawns which control the center (d4, e4, c4, f4)
c) pawns which are passed
d) pawns as a shield for the King (f2, g2, h2, O-O)
Bad:
a) isolated pawns (a2, no same-color pawn on b-file)
b) doubled pawns (a2, a3)
c) pawn distortion (a2, b5, c2)
d) holes in pawn formation (c5, d4, e5, empty hole at d5)
e) home pawns (f3, g4, h4, O-O) which have advanced
Obviously, many of these weaknesses are irrelevant if
there is no way for the opponent to exploit them; for
instance, if you have a K & Q and tripled, isolated rook
pawns and you opponent has just a King, you may well
promote each of your "bad" pawns to Queens, proving
that even a bad pawn makes for a decent Queen (while
no Queen could ever hope to be even an isolated pawn).
-- help bot "
Dear friends,
Allow me to re-quote from another chess forum:
"Life is painting a picture, not doing a sum"
Same happens to chess.
I remember "Deep Thought" in "World Open", Philadelphia 1987
When it announced "mate in 19" it came as a thunder and all
professional players felt the threat!
A couple of years later I had a discussion with Dr. Nunn about the
matter. I hold that the program was very strong tactically and very
weak strategically. (based on its win with Black against Larsen and
its loss with White against Browne). He hold that, at some point, the
tactical strength will be so high that strategy would not matter or be
confounded with it. I could not believe that, during my life, a
program would beat the best human player. Our attitude was somehow
corresponding to our "playing styles", mine more logical (in my view),
based on "strategic intuition" and his (also in my view) more
gambling, based on "tactical intuition".
Who was right?
No one!
Chess has 4 very important elements: material, space, time and
threats. (I excluded psichology and shape factors on purpose).
Programs are biassed more on 1 and 4, while a "positional" human GM on
2 and 3.
The chess strength is to weight these elemens and produce an
"evaluation". It is not true that programs don't know strategy. Best
programs have hundred of mini-rules to correct the brute "material
count" (e.g, "more exposed king") They found many of these rules
statistically, by playing millions of games against each other. They
"trim" the tree of variations with some evaluation.
Programs have huge memory (which is illegal) and adapted opening books
(also illegal) and yet, big strategy lapses. Here is the opening of
one in a few games I won against Rybka (the latter without opening
book) : 1.e3! g6 2.f4 Nf6?! 3.b3! Bg7 4.Bb2 O-O 5.Nf3 Nc6?! 6.Bb5! a6?
7=2EBxc6 dxc6? between +/- and +-. White's plan is "O-O, Qc1
(development) d3, e4 (centre) Nbd2, Nc4 and possibly Ne3 (improving
general energy, dynamic potential I call it) and finally f5! Even
"her" diabolic calculation could not overcome the strategical defects
of the position, that is, pawn structure and relative piece
positioning. A good human player would have find the toughest defence
and possibly counterplay: "Ne8, f5 (freezing White's centre, improving
king's safety) c5, b5, c4 (white squares). I would reformulate help
bot's, 'The reason is simply that even if you got your program
to "play like a grandmaster" in pawn-handling, its results
would barely budge since tactics are 90+% of the game.' changing
'tactics' by 'tactics + evaluation+long term planning' Fischer beat
Spassky because he was stronger in most technical aspects (opening
preparation, calculation speed, endgame, shape and psichology -
including motivation, will power, boldness, etc.) but Spassky had more
IDEAS. Rules are good, provided you look at them critically (with
exceptions feedback). Best (including for our theme - pawn structure)
is to teach your program patterns (don't ask me how).
That's all by now, I need some coffee!
Good luck in your games and/or programming!
|
| |
Date: 15 Sep 2007 00:26:16
From: Sanny
Subject: Re: Can you tell me about pawn structure?
|
I saw you played 4 games yesterday all with beginner Level, You win 3
Games and one was Drawn.
I Challenge you, that you will not be able to beat Easy and Normal
Levels as they will play very strong moves. Why play with weak Level.
If you want tough competition play with Easy & Normal Levels.
They will beat you 2/4 times. (As per my Calculations)
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 14 Sep 2007 22:41:58
From: Sanny
Subject: Re: Can you tell me about pawn structure?
|
On Sep 15, 2:56 am, help bot <[email protected] > wrote:
> On Sep 14, 1:34 am, Sanny <[email protected]> wrote:
>
>
>
>
>
> > Till now GetClub Chess only avoid double pawn formation and isolated
> > pawns.
> > Play Chess at:http://www.GetClub.com/Chess.html
>
> > What type of Pawn structures are good and what type of pawn structure
> > are bad?
>
> > 1. If Pawns are on same rows not protecting each other {in front of
> > King}. (Is it good)
> > King has no place to move???
>
> > 2. All pawns protecting other making a long diagonal Chain. (Is it
> > good)
> > A pawn Exchange can bring it down to Isolated Pawns???
>
> > 3. Two pawns on side protecting middle pawn. (Is it good)
> > There is a hole in Center???
>
> > How to find weak spots? Whenever a Pawn is played we have toi leave a
> > diagonal hole for bishop or a place for opponent Knight.
>
> > Can you tell me how to tell which pawn structure is better.
>
> > I went through lots of sites to know about Pawn Formations but could
> > not find any concrete example.
>
> > How to tell whether a Pawn formation is weak or strong?
>
> If I were you, I would ignore finesse work regarding pawn
> structures and focus priily on improving tactics.
>
> The reason is simply that even if you got your program
> to "play like a grandmaster" in pawn-handling, its results
> would barely budge since tactics are 90+% of the game.
>
> Many times I have replayed some game in Fritz, and
> no matter what the pawn structure was, no matter who
> had the superior minor piece, no matter who controlled
> the center, no matter which side was positionally
> better, someone blundered and Fritz instantly spotted
> the tactical refutation -- game over.
>
> -------------------------------------------------------------------------=
--=AD----
>
> But if you must know, here are a few ideas...
>
> Good:
>
> a) pawns connected to one another (a2, b2, c2)
>
> b) pawns which control the center (d4, e4, c4, f4)
>
> c) pawns which are passed
>
> d) pawns as a shield for the King (f2, g2, h2, O-O)
>
> Bad:
>
> a) isolated pawns (a2, no same-color pawn on b-file)
>
> b) doubled pawns (a2, a3)
>
> c) pawn distortion (a2, b5, c2)
>
> d) holes in pawn formation (c5, d4, e5, empty hole at d5)
>
> e) home pawns (f3, g4, h4, O-O) which have advanced
>
> Obviously, many of these weaknesses are irrelevant if
> there is no way for the opponent to exploit them; for
> instance, if you have a K & Q and tripled, isolated rook
> pawns and you opponent has just a King, you may well
> promote each of your "bad" pawns to Queens, proving
> that even a bad pawn makes for a decent Queen (while
> no Queen could ever hope to be even an isolated pawn).
>
> -- help bot- Hide quoted text -
>
> - Show quoted text -
Thats good. I find whenever a pawn is moved a hole will always
generated on its side. Then how to tell wich hole is bigger danerous
and which hole can be ignored?
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 14 Sep 2007 14:56:37
From: help bot
Subject: Re: Can you tell me about pawn structure?
|
On Sep 14, 1:34 am, Sanny <[email protected] > wrote:
> Till now GetClub Chess only avoid double pawn formation and isolated
> pawns.
> Play Chess at:http://www.GetClub.com/Chess.html
>
> What type of Pawn structures are good and what type of pawn structure
> are bad?
>
> 1. If Pawns are on same rows not protecting each other {in front of
> King}. (Is it good)
> King has no place to move???
>
> 2. All pawns protecting other making a long diagonal Chain. (Is it
> good)
> A pawn Exchange can bring it down to Isolated Pawns???
>
> 3. Two pawns on side protecting middle pawn. (Is it good)
> There is a hole in Center???
>
> How to find weak spots? Whenever a Pawn is played we have toi leave a
> diagonal hole for bishop or a place for opponent Knight.
>
> Can you tell me how to tell which pawn structure is better.
>
> I went through lots of sites to know about Pawn Formations but could
> not find any concrete example.
>
> How to tell whether a Pawn formation is weak or strong?
If I were you, I would ignore finesse work regarding pawn
structures and focus priily on improving tactics.
The reason is simply that even if you got your program
to "play like a grandmaster" in pawn-handling, its results
would barely budge since tactics are 90+% of the game.
Many times I have replayed some game in Fritz, and
no matter what the pawn structure was, no matter who
had the superior minor piece, no matter who controlled
the center, no matter which side was positionally
better, someone blundered and Fritz instantly spotted
the tactical refutation -- game over.
-------------------------------------------------------------------------------
But if you must know, here are a few ideas...
Good:
a) pawns connected to one another (a2, b2, c2)
b) pawns which control the center (d4, e4, c4, f4)
c) pawns which are passed
d) pawns as a shield for the King (f2, g2, h2, O-O)
Bad:
a) isolated pawns (a2, no same-color pawn on b-file)
b) doubled pawns (a2, a3)
c) pawn distortion (a2, b5, c2)
d) holes in pawn formation (c5, d4, e5, empty hole at d5)
e) home pawns (f3, g4, h4, O-O) which have advanced
Obviously, many of these weaknesses are irrelevant if
there is no way for the opponent to exploit them; for
instance, if you have a K & Q and tripled, isolated rook
pawns and you opponent has just a King, you may well
promote each of your "bad" pawns to Queens, proving
that even a bad pawn makes for a decent Queen (while
no Queen could ever hope to be even an isolated pawn).
-- help bot
|
| |
Date: 14 Sep 2007 19:08:03
From: [email protected]
Subject: Re: Can you tell me about pawn structure?
|
On Sep 14, 8:33 am, "Thomas T. Veldhouse" <[email protected] > wrote:
> I for one haven't written a chess engine [yet], so I can't say how
> much work it would be to even compete with crafty [which has full source
> available ... thanks professor!], which is an excellent engine but nowhere
> near the top.
I haven't written a chess engine yet either, but I've worked with the
code for one, and to write an engine that plays really well is not so
easy, as Sanny and company are finding out. (Writing one that plays
so-so is much easier, of course, though still by no means trivial.)
Crafty is rated as one of the best open-source engines. It is not as
good as commercial engines such as Rybka or Fritz and some others, but
it is very strong and very good. I doubt Sanny will ever come close
to Crafty's high standard.
But to make my real point: I have nothing against an engine that plays
at a certain level and is forthright about it. There are engines that
play in the 2000 ELO range and they self-profess to be "reasonably
strong" or words to that effect (certainly well above my level!). But
they don't claim to be championship material. Sanny's self-made ELO
estimates seem vastly inflated. He needs to run a real test and
publish the correct figures, whatever they might turn out to be.
|
| | |
Date: 14 Sep 2007 21:53:19
From: Thomas T. Veldhouse
Subject: Re: Can you tell me about pawn structure?
|
In rec.games.chess.computer [email protected] <[email protected] > wrote:
>
> Crafty is rated as one of the best open-source engines. It is not as
> good as commercial engines such as Rybka or Fritz and some others, but
> it is very strong and very good. I doubt Sanny will ever come close
> to Crafty's high standard.
>
Trust me, I know. I would never try to claim to the contrary and have a lot
of respect for Dr. Hyatt (sp?).
> But to make my real point: I have nothing against an engine that plays
> at a certain level and is forthright about it. There are engines that
> play in the 2000 ELO range and they self-profess to be "reasonably
> strong" or words to that effect (certainly well above my level!). But
> they don't claim to be championship material. Sanny's self-made ELO
> estimates seem vastly inflated. He needs to run a real test and
> publish the correct figures, whatever they might turn out to be.
>
He is arbitrary for sure, and he is inconsistant. But, he does offer his site
for free, so I guess that is his right, as is giving him flak the right of
members of this group.
--
Thomas T. Veldhouse
We have more to fear from the bungling of the incompetent than from the
machinations of the wicked.
|
| |
Date: 14 Sep 2007 08:25:58
From: Taylor Kingston
Subject: Re: Can you tell me about pawn structure?
|
On Sep 14, 10:33 am, "Thomas T. Veldhouse" <[email protected] > wrote:
> In rec.games.chess.computer [email protected] <[email protected]> wrote:
>
> You have to give him credit if he isn't charging you. It might not be a great
> site, and it might not be the best, but if it is free, it is tough to
> complain.
Your own motto seems to belie that sentiment, Tom:
> We have more to fear from the bungling of the incompetent than from the
> machinations of the wicked.
If the bungling of the incompetent is to be feared, then the
appropriate reaction to Sanny might be abject terror! ;-)
|
| | |
Date: 14 Sep 2007 17:28:48
From: Thomas T. Veldhouse
Subject: Re: Can you tell me about pawn structure?
|
In rec.games.chess.computer Taylor Kingston <[email protected] > wrote:
>
> Your own motto seems to belie that sentiment, Tom:
>
>> We have more to fear from the bungling of the incompetent than from the
>> machinations of the wicked.
>
> If the bungling of the incompetent is to be feared, then the
> appropriate reaction to Sanny might be abject terror! ;-)
>
Hehe ... I forgot about that. Good point :-)
--
Thomas T. Veldhouse
We have more to fear from the bungling of the incompetent than from the
machinations of the wicked.
|
| |
Date: 14 Sep 2007 14:07:29
From: [email protected]
Subject: Re: Can you tell me about pawn structure?
|
On Sep 14, 1:50 am, Sanny <[email protected] > wrote:
> > How to tell whether a Pawn formation is weak or strong?
>
> > Bye
> > Sanny
>
> Today GetClub Chess played very good moves and Draw game With one of
> the Worlds best Chess Program Jester.
Where did you get this incredible idea that Jester is up there with
Rybka et al?
>
> Below is the Game Played at:http://www.GetClub.com/Chess.html
>
> Jester: White
> GetClub Black (Beginner Level)
>
(game deleted)
Even a lowly class player such as I can see that this game is pure
baloney with poor play to say the least.
>
> Game Draw.!!! with Black Pieces.
Draw by repetition after 19 moves in a bad game? And you are proud of
this?
> So It gives me relief the Program is now playing very good moves.
You should not be relieved too early.
> Yesterday I said of shutting my Program
So far you haven't made a lot of progress, and there are many, many
superior options. Surely you should spend your time and money on more
useful things.
> Its after long hard work the sucess came.
Everyone can be proud of success after hard work ... but you're
unfortunately not there yet.
|
| | |
Date: 14 Sep 2007 14:33:30
From: Thomas T. Veldhouse
Subject: Re: Can you tell me about pawn structure?
|
In rec.games.chess.computer [email protected] <[email protected] > wrote:
>
> So far you haven't made a lot of progress, and there are many, many
> superior options. Surely you should spend your time and money on more
> useful things.
>
>> Its after long hard work the sucess came.
>
> Everyone can be proud of success after hard work ... but you're
> unfortunately not there yet.
>
You have to give him credit if he isn't charging you. It might not be a great
site, and it might not be the best, but if it is free, it is tough to
complain. I for one haven't written a chess engine [yet], so I can't say how
much work it would be to even compete with crafty [which has full source
available ... thanks professor!], which is an excellent engine but nowhere
near the top.
I would hope that he can take suggestions on how to improve his process [I am
not to sure there is a team working behind him on this] and put out a better
product, as it seems like he would like to someday charge for it.
--
Thomas T. Veldhouse
We have more to fear from the bungling of the incompetent than from the
machinations of the wicked.
|
| |
Date: 14 Sep 2007 01:24:29
From: Sanny
Subject: Re: Can you tell me about pawn structure?
|
> Its after long hard work the sucess came. And I am very happy about
> it.
>
> Bye
> Sanny
>
> Play Chess at: http://www.GetClub.com/Chess.html
Here is the new ratings I give to GetClub Chess. +50 to each.
Beginner: 1750 =========Nomorechess in Blitz Mode
Easy: 1850 ============Nomorechess in Serious Mode
Nornal: 1950============Zebediah in Blitz Mode
Master: 2050============Zebediah in Serious Mode
So Strong Players what do you say?
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 14 Sep 2007 00:50:20
From: Sanny
Subject: Re: Can you tell me about pawn structure?
|
> How to tell whether a Pawn formation is weak or strong?
>
> Bye
> Sanny
>
> Play Chess at:http://www.GetClub.com/Chess.html
Today GetClub Chess played very good moves and Draw game With one of
the Worlds best Chess Program Jester.
Below is the Game Played at: http://www.GetClub.com/Chess.html
Jester: White
GetClub Black (Beginner Level)
Recorded Game: http://www.getclub.com/playgame.php?id=DM9355&game=Chess
1.e2-e4 e7-e5
2.Nb1-c3 Ng8-f6
3.g2-g3 Nb8-c6
4.Ng1-f3 Bf8-b4
5.Bf1-b5 Bb4xc3
6.d2xc3 Ke8-g8
7.Bc1-g5 h7-h6
8.Bg5xf6 Qd8xf6
9.Ke1-g1 Nc6-d8
10.h2-h4 c7-c5
11.Bb5-e2 Nd8-c6
12.Qd1-d5 d7-d6
13.Rf1-d1 Rf8-d8
14.Qd5-d3 Bc8-h3
15.Qd3-e3 Ra8-b8
16.a2-a3 Bh3-e6
17.Rd1-d2 Be6-h3
18.Rd2-d1 Bh3-e6
19.Rd1-d2 -------------
Game Draw.!!! with Black Pieces.
Recorded Game: http://www.getclub.com/playgame.php?id=DM9355&game=Chess
So It gives me relief the Program is now playing very good moves.
Yesterday I said of shutting my Program if it do not improve in 4
days. And today it satisfied me so now I can continue with GetClub
Chess Game without shutting it forever.
Its after long hard work the sucess came. And I am very happy about
it.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| | |
Date: 14 Sep 2007 11:03:47
From: Kenneth Sloan
Subject: Re: Can you tell me about pawn structure?
|
Sanny wrote:
>> How to tell whether a Pawn formation is weak or strong?
>>
>> Bye
>> Sanny
>>
>> Play Chess at:http://www.GetClub.com/Chess.html
>
> Today GetClub Chess played very good moves and Draw game With one of
> the Worlds best Chess Program Jester.
>
> Below is the Game Played at: http://www.GetClub.com/Chess.html
>
> Jester: White
> GetClub Black (Beginner Level)
>
> Recorded Game: http://www.getclub.com/playgame.php?id=DM9355&game=Chess
>
> 1.e2-e4 e7-e5
> 2.Nb1-c3 Ng8-f6
> 3.g2-g3 Nb8-c6
> 4.Ng1-f3 Bf8-b4
> 5.Bf1-b5 Bb4xc3
> 6.d2xc3 Ke8-g8
> 7.Bc1-g5 h7-h6
> 8.Bg5xf6 Qd8xf6
> 9.Ke1-g1 Nc6-d8
> 10.h2-h4 c7-c5
> 11.Bb5-e2 Nd8-c6
> 12.Qd1-d5 d7-d6
> 13.Rf1-d1 Rf8-d8
> 14.Qd5-d3 Bc8-h3
> 15.Qd3-e3 Ra8-b8
> 16.a2-a3 Bh3-e6
> 17.Rd1-d2 Be6-h3
> 18.Rd2-d1 Bh3-e6
> 19.Rd1-d2 -------------
>
> Game Draw.!!! with Black Pieces.
>
> Recorded Game: http://www.getclub.com/playgame.php?id=DM9355&game=Chess
>
> So It gives me relief the Program is now playing very good moves.
Here's a hint: the above are *not* "very good moves".
>
> Yesterday I said of shutting my Program if it do not improve in 4
> days. And today it satisfied me so now I can continue with GetClub
> Chess Game without shutting it forever.
>
> Its after long hard work the sucess came. And I am very happy about
> it.
>
> Bye
> Sanny
>
> Play Chess at: http://www.GetClub.com/Chess.html
>
>
--
Kenneth Sloan [email protected]
Computer and Information Sciences +1-205-932-2213
University of Alabama at Birmingham FAX +1-205-934-5473
Birmingham, AL 35294-1170 http://www.cis.uab.edu/sloan/
|
|
