Main
Date: 19 Jan 2008 07:43:57
From: Sanny
Subject: Please Analyze this Game between Help Bot & Beginner.
Game Played between help bot and beginner at GetClub.com

--------------------------------------------------------------------------------
help bot: (White)
beginner: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM14332&game=Chess
--------------------------------------------------------------------------------

White -- Black
(help bot) -- (beginner)

1. e2-e3{4} e7-e5{0}
2. d2-d4{4} e5-d4{6}
3. e3-d4{2} d7-d5{10}
4. Bf1-d3{6} Nb8-c6{8}
5. c2-c3{8} Ng8-f6{10}
6. Bc1-g5{6} Qd8-d6{10}
7. Ng1-e2{12} h7-h6{6}
8. Bg5-f4{8} Qd6-e7{6}
9. Ke1-g1{6} Bc8-g4{10}
10. f2-f3{12} Bg4-e6{6}
11. Rf1-e1{8} g7-g5{8}
12. Bf4-g3{22} Ke8-c8{6}
13. Nb1-d2{16} Bf8-g7{6}
14. Nd2-b3{10} Nf6-h5{6}
15. Bg3-f2{8} Kc8-b8{22}
16. a2-a4{14} Kb8-a8{28}
17. a4-a5{6} Rd8-e8{34}
18. Nb3-c5{6} Nh5-f6{6}
19. Qd1-b3{24} Re8-b8{10}
20. Bd3-b5{24} Qe7-d6{6}
21. Bf2-g3{6} Qd6-e7{6}
22. a5-a6{52} Be6-h3{10}
23. a6-b7{16} Rb8-b7{18}
24. Bb5-c6{10} Rh8-b8{6}
25. Bc6-b7{14} Rb8-b7{6}
--------------------------------------------------------------------------------
help bot: (White)
beginner: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM14332&game=Chess

I want to know which were the wrong moves that beginner lost in just
25 Moves.

Bye
Sanny

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





 
Date: 23 Jan 2008 19:28:58
From: help bot
Subject: Re: Please Analyze this Game between Help Bot & Beginner.
On Jan 23, 6:55 pm, chipschap <[email protected] > wrote:
> > Master: 2550 (5-10 min/move)
>
> This is quite a claim. Can you in any way back it up? This makes it
> almost as strong as Hiarcs for the Palm with an ARM processor....

That number looks ridiculous to me.

The GetClub program still has major problems
in the endgame, and it still likes to sacrifice a
piece for less than full compensation, so how
is it going to reach the 2500+ level? A: It can't.

But consider for a moment the conditions
under which the human opponents play: the
games last for many hours and it is not only
possible, but likely that at some point there
will be an unforced error. GC is gradually
getting both faster and stronger, so even if
we can dismiss the above claim as a gross
exaggeration, there is a possibility that the
top levels can score *above 2000* against
humans, who may grow impatient or tired
and play below their "normal" level.

One reason that Sanny is constantly
*guessing* at the strength is that so few
players are willing to try their hand and
report an independent estimate of their
own. If a lot of rated players were to do
that, we could perhaps average them out
and see what that yields. The other issue
is that there is no real time limit for the
human players; they can't lose on time or
be forced to blunder in time pressure; but
this seems to be more than compensated
for by other factors.


-- help bot




 
Date: 23 Jan 2008 15:55:27
From: chipschap
Subject: Re: Please Analyze this Game between Help Bot & Beginner.

> Master: 2550 (5-10 min/move)

This is quite a claim. Can you in any way back it up? This makes it
almost as strong as Hiarcs for the Palm with an ARM processor....


 
Date: 23 Jan 2008 10:06:20
From: Sanny
Subject: Re: Jester Vs Beginner
On Jan 23, 10:22=A0pm, [email protected] (Andy Walker) wrote:
> In article <[email protected].=
com >,
> tin Brown =A0<


 
Date:
From: Martin Brown
Subject: Re: Jester Vs Beginner


  
Date: 23 Jan 2008 17:22:39
From: Andy Walker
Subject: Re: Jester Vs Beginner
In article <a6e2ab0f-95ce-41d5-b958-d5aa7553e36a@h11g2000prf.googlegroups.com >,
tin Brown <


 
Date: 23 Jan 2008 02:00:04
From: Sanny
Subject: Re: Please Analyze this Game between Help Bot & Beginner.
> > This is just a bit assumption.
>
> A "bit assumption" ? More like 500 ELO points too optimistic. Beginner
> mode is presently missing obvious mate in twos and regularly moving
> its queen or king into harms way. That is not characteristic of ELO
> 1600 play.
>
> Regards,
> tin Brown-

Game was a further improved today So now it will be able to see Mate
in Two and play better moves. Can you win the Beginner Level?

It is easy to do analysis of a game after it is played but much
difficult to win by himself. Just play a few games and see if you can
win the Beginner Level.

With The improvements made today I again adjust the Rating of GetClub
to new Ratings.

The Rating of GetClub games as per there new strength would be. +100
to each level.

Beginner: 1950 (5-20 seconds/move)

Easy: 2150 (20-60 seconds/move)
It Plays 10 times better than Beginner Level.

Normal: 2350 (1-3 min/move)
It Plays 10 times better than Easy Level

Master: 2550 (5-10 min/move)
It Plays 10 times better than Normal Level.

So if you are playing with Beginner Level you can assume you are
playing with 1950 elo player. So if you are rated lower than 1900 then
you cannot win the Beginner Level easily.

Bye
Sanny

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


 
Date: 23 Jan 2008 01:49:24
From: Sanny
Subject: Jester Vs Beginner
Here is a tough game between Jester & Beginner Please analyze if you
can find any mistakes with the game played by GetClub Beginner Level.

Game Played between sanjay11 and beginner at GetClub.com

--------------------------------------------------------------------------------
Jester: (White)
beginner: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM14593&game=Chess
--------------------------------------------------------------------------------

White -- Black
(Jester) -- (beginner)

1. e2-e4{2} e7-e5{0}
2. f2-f4{6} d7-d5{0}
3. e4-d5{8} e5-e4{0}
4. d2-d3{6} Qd8-d5{8}
5. Nb1-c3{10} Bf8-b4{6}
6. Bc1-d2{10} Bb4-c3{6}
7. Bd2-c3{8} f7-f6{10}
8. Qd1-e2{12} e4-e3{8}
9. Qe2-e3{10} Ng8-e7{6}
10. Ng1-f3{10} Nb8-c6{8}
11. Bf1-e2{12} Bc8-g4{8}
12. h2-h3{10} Bg4-f5{6}
13. Ke1-g1{10} Ke8-c8{6}
14. Nf3-h4{14} Rh8-e8{10}
15. Be2-f3{12} Qd5-d6{14}
16. Bf3-c6{14} Qd6-c6{8}
17. Qe3-a7{16} Bf5-e6{12}
18. Ra1-e1{18} b7-b6{12}
19. Nh4-f3{16} Ne7-f5{14}
20. Re1-e4{12} Nf5-g3{8}
21. Nf3-d4{16} Rd8-d4{8}
22. Qa7-a6{14} Kc8-b8{6}
23. Re4-d4{12} Ng3-e2{6}
24. Kg1-h2{12} Ne2-d4{6}
25. Bc3-d4{12} Qc6-c2{6}
26. Rf1-e1{12} Qc2-c6{6}
27. Bd4-c3{20} h7-h6{8}
28. d3-d4{24} Re8-g8{6}
29. Re1-e3{12} Be6-d5{8}
30. g2-g3{10} g7-g6{6}
31. b2-b3{14} f6-f5{6}
32. Qa6-e2{12} h6-h5{12}
33. Qe2-e1{12} Kb8-a8{6}
34. Re3-e7{14} h5-h4{6}
35. g3-h4{16} Bd5-e4{14}
36. Qe1-g3{12} Be4-d5{6}
37. Qg3-e3{20} Qc6-d6{6}
38. Qe3-e5{16} Rg8-c8{6}
39. h4-h5{12} g6-h5{6}
40. Bc3-b4{14} Qd6-e5{8}
41. Re7-e5{12} Bd5-e4{8}
42. Bb4-e7{20} Rc8-g8{8}
43. Be7-g5{12} c7-c6{6}
44. Kh2-g3{12} b6-b5{8}
45. Kg3-h4{32} Rg8-h8{6}
46. Bg5-f6{12} Rh8-h7{8}
47. Kh4-g5{10} h5-h4{8}
48. Re5-e7{14} Rh7-e7{6}
49. Bf6-e7{12} Ka8-b8{8}
50. Kg5-h4{12} Be4-g2{8}
51. a2-a4{14} b5-a4{6}
52. b3-a4{26} Kb8-b7{8}
53. Kh4-g3{10} Bg2-e4{6}
54. h3-h4{10} Kb7-a7{6}
55. h4-h5{10} Ka7-a8{6}
56. h5-h6{10} c6-c5{6}
57. Be7-c5{10} Ka8-b7{24}
58. h6-h7{10} Be4-c6{8}
59. a4-a5{12} Kb7-a8{6}
60. a5-a6{12} Ka8-b8{8}
61. Qh7-h8{Q}{12} Kb8-c7{12}
62. Qh8-f6{14} Bc6-a8{6}
63. Qf6-d6{12} Kc7-c8{10}
--------------------------------------------------------------------------------
Jester: (White)
beginner: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM14593&game=Chess

Initially Jester was 2 pawns up but GetClub played very good moves and
took back one pawn. But in end game Jester was able to kill pawns by
its King and get Queen to Check Mate in 64 Moves.

Bye
Sanny

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



 
Date:
From: Martin Brown
Subject: Re: Please Analyze this Game between Help Bot & Beginner.


 
Date: 22 Jan 2008 21:40:06
From: Sanny
Subject: Re: Please Analyze this Game between Help Bot & Beginner.
> =A0 The higher levels, which I rarely play now,
> are even better. =A0Because there is no real
> time limit on the human player, it is no
> easy matter to determine an approximate
> rating, but it would not surprise me if the
> Master level is now over 2000 USCF.
>

Here is the Rating for each level

The Rating of GetClub games as per there new strength would be.

Beginner: 1850 (5-20 seconds/move)

Easy: 2050 (20-60 seconds/move)
It Plays 10 times better than Beginner Level.

Normal: 2250 (1-3 min/move)
It Plays 10 times better than Easy Level

Master: 2450 (5-10 min/move)
It Plays 10 times better than Normal Level.

This is just a bit assumption. Have a match with Fritz and let me know
if I am correct.

Bye
Sanny

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





 
Date: 22 Jan 2008 13:02:51
From: help bot
Subject: Re: Please Analyze this Game between Help Bot & Beginner.
On Jan 22, 1:55 pm, Sanny <[email protected] > wrote:

> And I think now even Taylor Kingston can not beat the Master Level.
> May be he may win Master Level in 1/4 times. One day I remember Taylor
> Kingston said Master Level will never be able to beat him. Today I
> think the Time has Come when Master Level can beat Taylor Kingston.
> That is only because the GetClub Chess has improved a lot.

I think TK is too worried about protecting his
50-0 record to risk it. If he does play, it will be
under some other moniker, or else he will use
Fritz to ensure that no embarrassing loss
spoils his perfect-looking numbers; I say
perfect-looking because as we all know, the
program occasionally "defeated" everyone,
and the result now seen on the site is what
Sanny has decided, or negotiated with the
"losers". Take "help bot", for instance; just
recently a game was "given" to the GetClub
program in which it could not mate with a
King and Rook; that is on top of all the games
where it was decided that I "must have"
clicked the "resigns" button, plus the ones
where I actually did resign.

The main thing is to somehow get these
people to stop assuming that the program is
just as bad as it used to be; to get them to
log on and get their butts kicked, after which
of course they will go silent and stop their
moaning and groaning. ; >D


-- help bot


 
Date: 22 Jan 2008 12:53:16
From: help bot
Subject: Re: Please Analyze this Game between Help Bot & Beginner.
On Jan 22, 1:28 pm, "[email protected]" <[email protected] > wrote:

> It's so predictable. Sanny announces new and fabulous playing
> strength improvements (he even asked again if Fritz could beat
> "Master" level on GetClub). Then Sanny posts a game asking what went
> wrong, he'd like to fix it. Then (still to come in this cycle) Sanny
> will say the mistake has been corrected and now GetClub cannot be
> defeated.... on and on. This has been literally years now!

So, you can't rely on what Sanny says; you
have to do your own "research".

My early "research" had me thinking the
GetClub program had reincarnated me as
Paul Morphy! Every game allowed for a
multitude of sacrifices-- which always
worked.

But lately the program has been getting
tougher and tougher, except when Sanny
decides to do another speed-up. In spite
of all the weaknesses, and the fact that it
can't win a K & R vs. K ending, you may
find the program to be somewhere near
the Class B level these days, which is a
huge improvement from the old days.

The higher levels, which I rarely play now,
are even better. Because there is no real
time limit on the human player, it is no
easy matter to determine an approximate
rating, but it would not surprise me if the
Master level is now over 2000 USCF.


-- help bot




 
Date: 22 Jan 2008 10:55:08
From: Sanny
Subject: Re: Please Analyze this Game between Help Bot & Beginner.
On Jan 22, 11:28=A0pm, "[email protected]" <[email protected] >
wrote:
> It's so predictable. =A0Sanny announces new and fabulous playing
> strength improvements (he even asked again if Fritz could beat
> "Master" level onGetClub). =A0Then Sanny posts a game asking what went
> wrong, he'd like to fix it. =A0Then (still to come in this cycle) Sanny
> will say the mistake has been corrected and nowGetClubcannot be
> defeated.... on and on. =A0This has been literally years now!

--- >This has been literally years now!

The GetClub Chess is Just 17 months old. So How many years back did
you heard about GetClub Chess? I am only pointing those games where
GetClub Loose.

You can see that now no one dares to play with Easy & Normal Level.
Everyone is playing with Beginner Level which makes moves in 5-10
seconds. While the players are thinking for 20-30 seconds. So despite
improvements what can I do if everyone just play with Beginner Level
instead of higher levels..

Play with Easy/Normal & Master Level and they will beat you in 9/10
games.

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

Earlier Help Bot was even unbeatable by Master Level. But now even
Beginner Level has beaten Help Bot 3-4 times. And ask him if he can
win the Easy & Normal Levels anymore.

Only 1 person Mr Phil was able to beat Easy Level recently, Rest all
are afraid to play with Easy & Normal Level. They think for 20-40
seconds but play very tough moves.

Here is a brief History of GetClub Chess.

Jun 2006 : It played First Move in 1 hour.
Dec 2006: It played a bit sensible moves (10 min/ move)
Apr 2007: It played all legal moves (2/3 min per move)
Nov 2007: It started playing good games (20 sec per move )
Jan 2008: It is playing very good moves (6-10 sec per move)

So you can see how GetClub Chess has been improving. So am I wrong in
any way. And today everyone is afraid of Easy & Normal Levels.

And I think now even Taylor Kingston can not beat the Master Level.
May be he may win Master Level in 1/4 times. One day I remember Taylor
Kingston said Master Level will never be able to beat him. Today I
think the Time has Come when Master Level can beat Taylor Kingston.
That is only because the GetClub Chess has improved a lot.

Bye
Sanny

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



 
Date: 22 Jan 2008 10:28:25
From: [email protected]
Subject: Re: Please Analyze this Game between Help Bot & Beginner.
It's so predictable. Sanny announces new and fabulous playing
strength improvements (he even asked again if Fritz could beat
"Master" level on GetClub). Then Sanny posts a game asking what went
wrong, he'd like to fix it. Then (still to come in this cycle) Sanny
will say the mistake has been corrected and now GetClub cannot be
defeated.... on and on. This has been literally years now!


 
Date: 22 Jan 2008 03:21:45
From: help bot
Subject: Re: Please Analyze this Game between Help Bot & Beginner.
On Jan 22, 3:09 am, Bjoern <[email protected] > wrote:

> > King Rook - King is always a Win So I gave Beginner a Win.
>
> Giving it the win is ridiculous, exactly because it does not know how to
> win.

I'm convinced it can't win, which is why I played
on.

Another matter is, why does the program stop
making moves? It is eerily akin to there being
a Wizard-of-Oz behind the curtain, pulling levers.
When the wiz detects a situation in which his
creation cannot make headway, he freezes up,
and refuses to move. What's up with that? I
mean, am I playing a Java-machine, or a man-
machine combination, an android?



> > The game understand 3-Fold repetition. If the same Moves are made by
> > Black & White for 3 times it results in a draw. I think the Computer
> > had changed the Move as it was in Advantage.
>
> I suspect you have implemented it incorrectly, it is not really about
> the same moves being made. It is about the same position occuring 3
> times at any point in the game (or being about to occur, but let's say
> it would be enough to make sure the game recognises that the same
> position has occured 3 times (and executing such a move on the Java
> board would be a way of saying "okay, I would make this move and claim
> the draw). The thing to make sure in the programming is that the
> position can occur a lot of moves apart (say first time move 100, then
> on move 120, then on move 148) and that it is not necessarily created by
> the same move, e.g. if white has his rook on c2 and a king on a1, black
> has his king on e5, then the following is a threefold repetiotion:
> 1.Rb2 Kf5 (first occurence of the position)
> 2.Kb1 Ke6
> 3.Kc2 Ke5
> 4.Kb3 Kf5
> 5.Ka2 Kf6
> 6.Ka1 Kf5 (second occurence of the position)
> 7.Ka2 Ke5
> 8.Kb3 Kf5
> 9.Kc3 Ke5
> 10.Kc2 Ke6
> 11.Kb1 Kf5
> 12.Ka1 (not an occurence, because the move rights are the wrong way around)
> 12...Kf6
> 13.Kb1 Ke6
> 14.Ka1
> and now black can claim a draw if he declares that he intends to play
> 14...Kf5.

At my local chess club, there was a game in
which one player got irate when, he said, the
position had been repeated three times, but he
was no longer keeping score because of time
pressure and thus, his claim could not be
verified. The TD remained silent, yet I wondered
if the fact that there *was* an eye-witness could
be relevant; a third party was sitting at the table
in question, watching the game the whole time.
On the other hand, he may have been watching
the game because he was friends with the guy
who went on to win, so maybe it would have
been naive to think he would "testify" to any
threefold repetition. (He, too, remained silent.)


-- help bot



 
Date: 21 Jan 2008 11:45:37
From: help bot
Subject: Re: Please Analyze this Game between Help Bot & Beginner.
On Jan 21, 7:32 am, tin Brown <

 
Date:
From: Martin Brown
Subject: Re: Please Analyze this Game between Help Bot & Beginner.


 
Date: 20 Jan 2008 09:39:29
From: help bot
Subject: Re: Please Analyze this Game between Help Bot & Beginner.
On Jan 20, 1:40 am, Sanny <[email protected] > wrote:

> > In order to determine the last "playable position",
> > I could run this through Fritz. But here's a clue:
> > count up the number of times the Beginner level
> > moved its Queen (over and over and over). Apart
> > from saving the Queen from being taken, those
> > were all wasted tempos which the opponent can
> > use to build up an attack.
>
> Beginner moved its queen only 4 times in this game.

Getting checkmated around move 25, that looks
like a lot of Queen moves to me.


> Out of Which First It Played Q-d6 on 6th move.and when attacked by
> Bishop it played Queen to Q-e7 on 8th Move.
>
> After that it played Queen on 20th move to Q-d6 to save Knight and
> then took Back Queen to Q-e7 as it was again threathened by Bishop.
>
> So Out of 4 times Queen was Moved 2 times it played for Saving its
> Queen, And one time it played to save Rook. So It was only in 6th Move
> Q-d6 was played without any attack.

But all of these moves were unnecessary, since
for example, the program ought to have been
developing its minor pieces and castling, not
playing around with the Queen.


> Do you think 6. .... Qd6 was wrong?

Yes. The Queen belongs on d8-- right
where she stands on move one.


> Why what was then the correct move?

After my B-g5, a logical response would be
...B-e7, intending to castle king-side next.


> > In this kind of position, no unguarded entry square
> > exists on the open e-file, so the heavy pieces
> > (Rooks and Queens) will not get traded. Instead,
> > each side can attack the enemy King, willy-nillly.
>
> Should Black Go for an Attack on White King instead of putting Rooks &
> Queen on e file?

Yes. In this kind of position, all the entry
points on the e-file are guarded by minor
pieces (Knights and Bishops), so it is not
possible to penetrate with a Rook or
Queen (unless they are tripled on the file,
and an exchange sacrifice is made).

> Yes in King-Rook ending it is incapable of giving Check Mate. The
> reasion is it thinks till 10 depth. Incase it do not find a Mate in 10
> then all moves are given equal benefit. Since there is nothing the
> King- Rook is gaining by killing any piece/pawn in 10 moves. Since all
> moves has equal point it is unable to find Mate and plays King & Rook
> randomly.

Perhaps, but I note a distinct tendency
for the program to play its King toward the
corners, as if worried about King-safety.
This is a H-U-G-E weakness in the
endgame.


> For that End Game Tables are there they are 10MB-50MB Big. If I use
> End Game Tables It will take an hour to download the Chess game. So
> Thats the reasion It plays badly in King-Rook Ending.

There is no need for endgame table-bases
here; all that is needed is for the program to
be taught to maximize its mobility, or control
of squares, while reducing the opponent's
as much as possible (barring stalemate).

Old table-top computers from the 1980s, like
the Fidelity Chess Challenger for instance, had
this down pat in spite of very limited hardware
resources and speed. After checkmate and
stalemate, next in importance comes material,
and then *mobility*. Generally, you give a tiny
bonus for each square attacked by your own
men, and a tiny penalty for each square the
opponent's men control. Add it all up and the
side with greater mobility nets the higher
position score, plus the possible moves can
be ranked when otherwise everything is a
dead tie.


> > As I've noticed has happened before, somebody
> > "intervened" on the program's behalf, ending the
> > game and handing the program a win it apparently
> > could not earn on its own.
>
> Yes there was an error which was removed. Since the game played
> arround 200 Moves The Largest number of Moves Set at GetClub. That
> error was Removed by Changing highest setting to 1000 Moves. So now
> you can play a game for upto 500 Moves.

Sounds like fun. (Groan)

> > These "freebies", along with the many cases
> > in which it is claimed that the human "must have
> > clicked the resigns button", result in an inflated
> > rating for the program, and the reverse for its
> > hapless victims.
>
> King Rook - King is always a Win So I gave Beginner a Win.

My point, exactly! You "give" it a win. How
often have you "given" the human players
such wins, for free and which they could
never have earned legitimately? In my
experience, unfinished games are normally
given the ax-- erased, deleted, and so it
goes.


> > Another issue is the fact that there is *still* no
> > way for the human player to claim a draw, no
> > "draw" button or detection of threefold repetition
> > of position, nothing. In the game I mentioned
> > above, I had a threefold repetition, but noticed

> The game understand 3-Fold repetition. If the same Moves are made by
> Black & White for 3 times it results in a draw. I think the Computer
> had changed the Move as it was in Advantage.

Perhaps. Maybe it was a two-and-a-half fold
repetition, and I miscounted. The actual rule
states that the draw claim is made *before*
executing the move which *would result* in the
position being repeated for the third time, the
same side on-move, and all possible moves
being identical (en passant, castling). So the
"draw" button would technically need to
appear after two repetitions, giving the human
the *option*, but not the obligation, to make a
draw claim. If you automatically enforce a
draw, you may ensnare an unwitting human
player who had no intention of making the
claim (though I feel no sympathy for losers
who cannot make progress and who instead
butt heads with a mindless computer, ad
infinitum).


-- help bot




 
Date: 19 Jan 2008 22:40:00
From: Sanny
Subject: Re: Please Analyze this Game between Help Bot & Beginner.
> =A0 In order to determine the last "playable position",
> I could run this through Fritz. =A0But here's a clue:
> count up the number of times the Beginner level
> moved its Queen (over and over and over). =A0Apart
> from saving the Queen from being taken, those
> were all wasted tempos which the opponent can
> use to build up an attack.

Beginner moved its queen only 4 times in this game.

Out of Which First It Played Q-d6 on 6th move.and when attacked by
Bishop it played Queen to Q-e7 on 8th Move.

After that it played Queen on 20th move to Q-d6 to save Knight and
then took Back Queen to Q-e7 as it was again threathened by Bishop.

So Out of 4 times Queen was Moved 2 times it played for Saving its
Queen, And one time it played to save Rook. So It was only in 6th Move
Q-d6 was played without any attack.

Do you think 6. .... Qd6 was wrong? Why what was then the correct
move?

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

> =A0 In this kind of position, no unguarded entry square
> exists on the open e-file, so the heavy pieces
> (Rooks and Queens) will not get traded. =A0Instead,
> each side can attack the enemy King, willy-nillly.
>

Should Black Go for an Attack on White King instead of putting Rooks &
Queen on e file?

> -----------------------------------------------------------------
> -----------------------------------------------------------------
>
> =A0 In another recent game, I was outplayed by the
> Beginner level and the best I could do was get to
> King versus King and Rook. =A0Though I tried
> repeatedly to get the program to move, even
> logging off and then back on at various times, it
> just sat there, with its King in the corner, and the
> Rook nearby. =A0It seems to me that the program
> is incapable of winning this elementary ending,
> because of a flaw in the program whereby it
> keeps its own King in or near the corner.

Yes in King-Rook ending it is incapable of giving Check Mate. The
reasion is it thinks till 10 depth. Incase it do not find a Mate in 10
then all moves are given equal benefit. Since there is nothing the
King- Rook is gaining by killing any piece/pawn in 10 moves. Since all
moves has equal point it is unable to find Mate and plays King & Rook
randomly.

For that End Game Tables are there they are 10MB-50MB Big. If I use
End Game Tables It will take an hour to download the Chess game. So
Thats the reasion It plays badly in King-Rook Ending.

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

> =A0 As I've noticed has happened before, somebody
> "intervened" on the program's behalf, ending the
> game and handing the program a win it apparently
> could not earn on its own.

Yes there was an error which was removed. Since the game played
arround 200 Moves The Largest number of Moves Set at GetClub. That
error was Removed by Changing highest setting to 1000 Moves. So now
you can play a game for upto 500 Moves.

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

>
> =A0 These "freebies", along with the many cases
> in which it is claimed that the human "must have
> clicked the resigns button", result in an inflated
> rating for the program, and the reverse for its
> hapless victims.

King Rook - King is always a Win So I gave Beginner a Win.

> =A0 Another issue is the fact that there is *still* no
> way for the human player to claim a draw, no
> "draw" button or detection of threefold repetition
> of position, nothing. =A0In the game I mentioned
> above, I had a threefold repetition, but noticed


The game understand 3-Fold repetition. If the same Moves are made by
Black & White for 3 times it results in a draw. I think the Computer
had changed the Move as it was in Advantage.

Bye
Sanny

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



  
Date: 22 Jan 2008 09:09:02
From: Bjoern
Subject: Re: Please Analyze this Game between Help Bot & Beginner.
Sanny wrote:
> Yes in King-Rook ending it is incapable of giving Check Mate. The
> reasion is it thinks till 10 depth. Incase it do not find a Mate in 10
> then all moves are given equal benefit. Since there is nothing the
> King- Rook is gaining by killing any piece/pawn in 10 moves. Since all
> moves has equal point it is unable to find Mate and plays King & Rook
> randomly.
>
> For that End Game Tables are there they are 10MB-50MB Big. If I use
> End Game Tables It will take an hour to download the Chess game. So
> Thats the reasion It plays badly in King-Rook Ending.
...
> King Rook - King is always a Win So I gave Beginner a Win.

Giving it the win is ridiculous, exactly because it does not know how to
win.

There's the 50 move rule (after 50 moves in which no pawn was moved or
piece captured, one or the other player can claim a draw), then there's
the rule about not trying to win by normal means and then there's the
3-fold repetition rule.

As pointed out by others there are many ways of creating rules that the
program follows in KR-K to make sure it plays that and other easy
endgames correctly.

>> Another issue is the fact that there is *still* no
>>way for the human player to claim a draw, no
>>"draw" button or detection of threefold repetition
>>of position, nothing. In the game I mentioned
>>above, I had a threefold repetition, but noticed
>
>
>
> The game understand 3-Fold repetition. If the same Moves are made by
> Black & White for 3 times it results in a draw. I think the Computer
> had changed the Move as it was in Advantage.

I suspect you have implemented it incorrectly, it is not really about
the same moves being made. It is about the same position occuring 3
times at any point in the game (or being about to occur, but let's say
it would be enough to make sure the game recognises that the same
position has occured 3 times (and executing such a move on the Java
board would be a way of saying "okay, I would make this move and claim
the draw). The thing to make sure in the programming is that the
position can occur a lot of moves apart (say first time move 100, then
on move 120, then on move 148) and that it is not necessarily created by
the same move, e.g. if white has his rook on c2 and a king on a1, black
has his king on e5, then the following is a threefold repetiotion:
1.Rb2 Kf5 (first occurence of the position)
2.Kb1 Ke6
3.Kc2 Ke5
4.Kb3 Kf5
5.Ka2 Kf6
6.Ka1 Kf5 (second occurence of the position)
7.Ka2 Ke5
8.Kb3 Kf5
9.Kc3 Ke5
10.Kc2 Ke6
11.Kb1 Kf5
12.Ka1 (not an occurence, because the move rights are the wrong way around)
12...Kf6
13.Kb1 Ke6
14.Ka1
and now black can claim a draw if he declares that he intends to play
14...Kf5.


 
Date: 19 Jan 2008 12:53:00
From: help bot
Subject: Re: Please Analyze this Game between Help Bot & Beginner.
On Jan 19, 10:43 am, Sanny <[email protected] > wrote:
> Game Played between help bot and beginner at GetClub.com
>
> --------------------------------------------------------------------------------
> help bot: (White)
> beginner: (Black)
> Game Played at:http://www.GetClub.com/Chess.html
> View Recorded Game:http://www.getclub.com/playgame.php?id=DM14332&game=Chess
> --------------------------------------------------------------------------------
>
> White -- Black
> (help bot) -- (beginner)
>
> 1. e2-e3{4} e7-e5{0}
> 2. d2-d4{4} e5-d4{6}
> 3. e3-d4{2} d7-d5{10}
> 4. Bf1-d3{6} Nb8-c6{8}
> 5. c2-c3{8} Ng8-f6{10}
> 6. Bc1-g5{6} Qd8-d6{10}
> 7. Ng1-e2{12} h7-h6{6}
> 8. Bg5-f4{8} Qd6-e7{6}
> 9. Ke1-g1{6} Bc8-g4{10}
> 10. f2-f3{12} Bg4-e6{6}
> 11. Rf1-e1{8} g7-g5{8}
> 12. Bf4-g3{22} Ke8-c8{6}
> 13. Nb1-d2{16} Bf8-g7{6}
> 14. Nd2-b3{10} Nf6-h5{6}
> 15. Bg3-f2{8} Kc8-b8{22}
> 16. a2-a4{14} Kb8-a8{28}
> 17. a4-a5{6} Rd8-e8{34}
> 18. Nb3-c5{6} Nh5-f6{6}
> 19. Qd1-b3{24} Re8-b8{10}
> 20. Bd3-b5{24} Qe7-d6{6}
> 21. Bf2-g3{6} Qd6-e7{6}
> 22. a5-a6{52} Be6-h3{10}
> 23. a6-b7{16} Rb8-b7{18}
> 24. Bb5-c6{10} Rh8-b8{6}
> 25. Bc6-b7{14} Rb8-b7{6}
> --------------------------------------------------------------------------------
> help bot: (White)
> beginner: (Black)
> Game Played at:http://www.GetClub.com/Chess.html
> View Recorded Game:http://www.getclub.com/playgame.php?id=DM14332&game=Chess
>
> I want to know which were the wrong moves that beginner lost in just
> 25 Moves.

In order to determine the last "playable position",
I could run this through Fritz. But here's a clue:
count up the number of times the Beginner level
moved its Queen (over and over and over). Apart
from saving the Queen from being taken, those
were all wasted tempos which the opponent can
use to build up an attack.

In this kind of position, no unguarded entry square
exists on the open e-file, so the heavy pieces
(Rooks and Queens) will not get traded. Instead,
each side can attack the enemy King, willy-nillly.

-----------------------------------------------------------------
-----------------------------------------------------------------

In another recent game, I was outplayed by the
Beginner level and the best I could do was get to
King versus King and Rook. Though I tried
repeatedly to get the program to move, even
logging off and then back on at various times, it
just sat there, with its King in the corner, and the
Rook nearby. It seems to me that the program
is incapable of winning this elementary ending,
because of a flaw in the program whereby it
keeps its own King in or near the corner.

As I've noticed has happened before, somebody
"intervened" on the program's behalf, ending the
game and handing the program a win it apparently
could not earn on its own.

These "freebies", along with the many cases
in which it is claimed that the human "must have
clicked the resigns button", result in an inflated
rating for the program, and the reverse for its
hapless victims.

Another issue is the fact that there is *still* no
way for the human player to claim a draw, no
"draw" button or detection of threefold repetition
of position, nothing. In the game I mentioned
above, I had a threefold repetition, but noticed
that you *still* refuse to accept the existence of
draws! How long has it been now... perhaps a
year or more, since their existence was first
pointed out? Win, loss, draw... count 'em: 3!
Three possible outcomes in a game of chess.
(I am ignoring double forfeits, and games in
which both sides checkmate one another at
precisely the same time... ?!!?!!)


-- hep blot