Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: GM/IM opening stats against crafty

Author: Matthew Hull

Date: 21:45:50 06/23/05

Go up one level in this thread


On June 23, 2005 at 23:08:24, Robert Hyatt wrote:

>On June 23, 2005 at 17:13:16, Robin Smith wrote:
>
>>On June 23, 2005 at 16:43:19, Matthew Hull wrote:
>>
>>>On June 23, 2005 at 15:53:46, Robin Smith wrote:
>>>
>>>>On June 23, 2005 at 14:57:35, Robert Hyatt wrote:
>>>>
>>>>>On June 23, 2005 at 11:29:32, Robin Smith wrote:
>>>>>
>>>>>>On June 23, 2005 at 09:39:48, Robert Hyatt wrote:
>>>>>>
>>>>>>>On June 23, 2005 at 03:37:51, Robin Smith wrote:
>>>>>>>
>>>>>>>>On June 22, 2005 at 16:20:32, Robert Hyatt wrote:
>>>>>>>>
>>>>>>>>>On June 22, 2005 at 13:51:40, Robin Smith wrote:
>>>>>>>>>
>>>>>>>>>>On June 22, 2005 at 03:10:00, Drexel,Michael wrote:
>>>>>>>>>>
>>>>>>>>>>>On June 21, 2005 at 23:00:37, Robin Smith wrote:
>>>>>>>>>>>
>>>>>>>>>>>>On June 21, 2005 at 18:36:34, Robert Hyatt wrote:
>>>>>>>>>>>>
>>>>>>>>>>>>>On June 21, 2005 at 16:44:21, Torstein Hall wrote:
>>>>>>>>>>>>>
>>>>>>>>>>>>>>On June 21, 2005 at 15:30:03, Robert Hyatt wrote:
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>On June 21, 2005 at 14:19:44, Robin Smith wrote:
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>On June 21, 2005 at 14:11:23, Mark Young wrote:
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>On June 21, 2005 at 14:04:37, Ted Summers wrote:
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>To sum it up " He played a drawish opening in a tactic way. " Not a good idea
>>>>>>>>>>>>>>>>>>when computers are able to hang with the best and proving themself as better
>>>>>>>>>>>>>>>>>>than humans in open tactical positions. However I still think GM Adams can pull
>>>>>>>>>>>>>>>>>>it together and Win or Draw this match.
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>[D] r2q1rk1/1pp3pp/p2b4/nP1p1p1b/2PPn3/3B1N1P/P1QN1PP1/1RB1R1K1 b - - 0 17
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>Having reached this position, we seemed to be watching the beginning of the end
>>>>>>>>>>>>>>>>>>for Adams in the first game but hopefully not the match.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>C4! was a killer positional shot.
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>c4 was a good move, but hardly a "killer".
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>It seems clear GM Adams missed this move when he played Na5.
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>Perhaps Adams miissed it, but it hardly seems "clear", since Black is still OK
>>>>>>>>>>>>>>>>afterwards. His loss happened later.
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>-Robin
>>>>>>>>>>>>>>>The problem here is that the kingside is already a bit open.  One does _not_, as
>>>>>>>>>>>>>>>a human, allow the computer to open _both_ sides of the board in the same game.
>>>>>>>>>>>>>>>It invites a debacle such as this.  Of course, he made a couple of tactical
>>>>>>>>>>>>>>>errors around the point where the rook on C8 was hanging, but he was already in
>>>>>>>>>>>>>>>the wrong kind of position...
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>All the comps were suggesting the same moves as played by Hydra, so there was no
>>>>>>>>>>>>>>>real surprises from the white side, just black making an error here, an error
>>>>>>>>>>>>>>>there, before long he fell off the rim of the canyon.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>This is in my view far to general. Black was at least = uptil move 23.Be6
>>>>>>>>>>>>>>[D]2rq1r1k/6pp/p2bB3/2p1Np1b/3Pn3/7P/P1Q2PP1/1RB1R1K1 b - - 0 23
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>Define "equal".  Here I am considering the important detail that white is a
>>>>>>>>>>>>>computer, black is a human.  In that regard, black is _not_ equal up to move 23.
>>>>>>>>>>>>
>>>>>>>>>>>>By that logic Adams was already much worse after 1.e4 no matter what he did.
>>>>>>>>>>>>Let's face it, Hydra is stronger. Adams will probably be under presure in every
>>>>>>>>>>>>game where he has the black pieces.
>>>>>>>>>>>>
>>>>>>>>>>>>> In fact, I don't believe black is anywhere near equal.
>>>>>>>>>>>>
>>>>>>>>>>>>He is equal unless you use your "considering the important detail that white is
>>>>>>>>>>>>a computer" logic.
>>>>>>>>>>>>
>>>>>>>>>>>>>He isn't lost, but he is far from equal and is at best fighting for a draw.
>>>>>>>>>>>>
>>>>>>>>>>>>>But in an open position.
>>>>>>>>>>>>>And he just has no chance in that kind of position.
>>>>>>>>>>>>
>>>>>>>>>>>>He was under presure, yes. That is a far cry from "has no chance".
>>>>>>>>>>>>
>>>>>>>>>>>>>But I would take white anywhere along the way in that game, as a human playing
>>>>>>>>>>>>>another human.  And by the way, any move after the "knight to the rim" move
>>>>>>>>>>>>>finds white better IMHO.
>>>>>>>>>>>>
>>>>>>>>>>>>Your opinion is wrong, unless perhaps you mean that white had a very slight
>>>>>>>>>>>>advantage. That is the norm in chess, by the way.
>>>>>>>>>>>>
>>>>>>>>>>>>>>Adams played 23...Rc7 while 23...cxd4 looks like it holds everything nicely
>>>>>>>>>>>>>>together.
>>>>>>>>>>>>>
>>>>>>>>>>>>>Doesn't quite hold everything nicely together.  The comps were at about +1 here
>>>>>>>>>>>>>already, went to +1.5 on the Rc7 move.
>>>>>>>>>>>>
>>>>>>>>>>>>Maybe Craqfty sees +1, but the top programs don't see anything near +1 until
>>>>>>>>>>>>_after_ Rc7. Before Rc7 black was fine.
>>>>>>>>>>>
>>>>>>>>>>>You don´t have a clue.
>>>>>>>>>>
>>>>>>>>>>And you do?  :-)
>>>>>>>>>>
>>>>>>>>>>>It´s always easy to sacrifice the exchange of others. In order to play this
>>>>>>>>>>>sacrifice you have to calculate correctly some very concrete lines.
>>>>>>>>>>
>>>>>>>>>>Of course. That is obvious and I never said otherwise. All I said was that black
>>>>>>>>>>is OK if he plays cxd4 instead of Rc7.
>>>>>>>>>>
>>>>>>>>>>>For example 24.Bxc8 Bxe5 ( The ending after 24...Qxc8 is very difficult to play
>>>>>>>>>>>for black) 25.Bxf5 d3 24.Qc6 d2 27.Bxd2 Nxd2 28.Rxe5 Nxb1 29.Bxb1 Qd1+ 30.Kh2
>>>>>>>>>>>Qxb1 31.Qd6 Kg8 32.Rxh5 Qxa2 = and the position after 25.Qxc8 Qf6 26.Qc4 Qxe5
>>>>>>>>>>>27.g3 is very difficult to play for black.
>>>>>>>>>>>
>>>>>>>>>>>Definitely not the typ of position you want to play against a computer.
>>>>>>>>>>
>>>>>>>>>>I agreed this is not the type of position a human wants to be in in another
>>>>>>>>>>post. Did you read it before you shot off your mouth?
>>>>>>>>>>
>>>>>>>>>>>Therefore Adams Rc7 is a completely understandable decision.
>>>>>>>>>>
>>>>>>>>>>I agree that Adams decision was understandable. I never said otherwise. It was
>>>>>>>>>>also a mistake, that's all; an understandable mistake. I have always agreed that
>>>>>>>>>>by this point Adams was in the type of position that is hard for a human to
>>>>>>>>>>play. That does not mean he made mistakes earlier. It is easier for white to
>>>>>>>>>>create open, messy positions that are hard for a human to play than it is for
>>>>>>>>>>black to prevent it, so just because it happens does not mean Adams made
>>>>>>>>>>mistakes prior to getting into such a position. Hyrda won because Hydra played
>>>>>>>>>>well, not because Adams "blundered" or made "outright stupid" choices or "GM
>>>>>>>>>>Adams missed this move". I think it is disrespectful to GM Adams when people say
>>>>>>>>>>such things, especially since Adams _didn't_ blunder.
>>>>>>>>>>
>>>>>>>>>>-Robin
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>I'll say it again.  You can throw high, fast and outside to a big hitter, and
>>>>>>>>>when he slaps it over the fence, you can say "good shot".  Or you can say "lousy
>>>>>>>>>pitch."  In this game, it was a lousy pitch by Adams.  If he chooses to avoid
>>>>>>>>>anti-computer type chess, that's fine, and no it isn't a blunder.
>>>>>>>>
>>>>>>>>Then why in http://www.talkchess.com/forums/1/message.html?432636 did you say
>>>>>>>>"He was guilty of a different type of blunder. Namely of playing 1. ... e5
>>>>>>>>against the computer." Was it a blunder or not? Have you changed your position
>>>>>>>>so that now we agree?
>>>>>>>>
>>>>>>>>-Robin
>>>>>>>
>>>>>>>No.  It was a mistake, or a blunder, or a foolhardy opening choice.  You pick
>>>>>>>the description.  But it was clearly the wrong approach to playing a computer.
>>>>>>>Anyone that has played them often will say the same thing...
>>>>>>>
>>>>>>>I don't see why this turns into an argument when the basic premise is so well
>>>>>>>understood by so many...
>>>>>>
>>>>>>Bob, believe it or not I understand the desirability of keeping the position
>>>>>>closed. Over and over again I have agreed with that. But there is a second basic
>>>>>>premise, also understood by so many ... play openings you know. You keep not
>>>>>>addressing that.
>>>>>>
>>>>>>-Robin
>>>>>>
>>>>>
>>>>>There is nothing to address then.  Let 'im keep playing 1. e4 e5.  He has lost
>>>>>both of those as black, in sterling fashion.  He can continue to do so, or he
>>>>>can decide to vary as Kasparov and others have when playing computers.  We know
>>>>>what sticking with e5 is going to produce.  We know the computer is not going to
>>>>>change.  So either he does, or he goes down in flames.  Which would you
>>>>>suggest???
>>>>
>>>>If I were to suggest something to Adams, it would be for him to play the Caro.
>>>>At least he has played it on occaision, so it is not completely unfamiliar. But
>>>>if he does opt for 1...e5 I won't hurl insults at him, implying he is stupid, as
>>>>you have done. If he plays 1...e5 I would assume that he is more comfortable
>>>>playing such openings against computers than he is playing other openings. All
>>>>the GMs practice with computers these days. Do you honestly belive you know
>>>>better than Adams what openings he does best against computers?
>>>
>>>
>>>If you can show double king pawn games where GMs on the black side do well
>>>against _strong_ computers, you will have made your point.
>>
>>Fair enough. How about:
>>
>>Deep Blue - Kasparov, 1996. Kasparov won.
>>Deep Fritz - Kramnik 2002, game 1. An easy draw for Kramnik
>>Deep Fritz - Kramnik 2002, game 3. Kramnik won.
>>Kasparov - X3D Fritz, 2003, game 2. Even though Kasparov lost this game he was
>>in a _superior_ and _relatively closed_ position when he blundered away at least
>>an easy draw (with some _winning_ chances) on move 32.
>>
>>Is that enough?
>>
>>-Robin
>>
>
>
>
>
>How about listing _all_ such games?  Might there be a _different_ overall
>conclusion based on all e4 e5 games, rather than cherry-picking one here and
>there?


Here are the opening stats for scrappy (Crafty versus humans only) losses in the
last year or so (opening code and # of games):

Scrappy losses playing the white pieces
A02 1
A40 1
B07 1
C02 2
A03 5
Scrappy losses playing the black pieces
A00 1
A05 1
A25 1
A46 1
C41 1
C45 1
D45 1
D85 1
E15 1
E70 1
E92 1
A79 2
B22 2
C42 2
D02 2
D10 2






>
>
>
>
>
>
>
>
>
>
>
>>>Until then, Bob has made his point, IMHO, and the recent games are eloquent
>>>testimony to that fact.
>>>
>>>>
>>>>-Robin
>>>>
>>>>>>>>
>>>>>>>>>But it _is_ a
>>>>>>>>>mistake.  You play to your opponent's weaknesses, not to his strength, for max
>>>>>>>>>advantage.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>>Michael
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>>But then the next few moves were mostly
>>>>>>>>>>>>>bad by black, turning this into a debacle.  But if there were not so many open
>>>>>>>>>>>>>files, open diagonals, etc, black wouldn't have had to be worrying about tactics
>>>>>>>>>>>>>all over the board.
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>> One line could be 23...cxd4 24.Qxc8 Qf6 25.Qc4 Qxe5 26.Qa5 and black
>>>>>>>>>>>>>>looks OK to me.
>>>>>>>>>>>>>
>>>>>>>>>>>>>But white looks better to me there.  Maybe not "winning better" but
>>>>>>>>>>>>>"significantly better".
>>>>>>>>>>>>
>>>>>>>>>>>>Try "very slightly better". Adams played well until Rc7. Hydra is very strong
>>>>>>>>>>>>and kept putting the presure on and finally Adams made a mistake.
>>>>>>>>>>>>
>>>>>>>>>>>>-Robin



This page took 0.01 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.