| |
Main
Date: 01 Jun 2008 06:11:03
From: Sanny
Subject: GetClub Chess now twice better.
|
A few modifications were done in GetClub Chess and now it plays Twice
better moves.
See this game with Rybka GetClub managed till 43 moves.
Game Played between Rybka and normal at GetClub.com
---------------------------------------------------------------------------=
=AD-----
Rybka : (White)
normal: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=3DDM20434&game=3D=
Chess
---------------------------------------------------------------------------=
=AD-----
White -- Black
(Rybka ) -- (normal)
1. e2-e4{8} e7-e6{0}
2. d2-d4{8} d7-d5{0}
3. Nb1-d2{10} Nb8-c6{110}
4. Ng1-f3{18} Bc8-d7{146}
5. Bf1-d3{16} g7-g6{92}
6. Ke1-g1{24} Ng8-f6{108}
7. e4-e5{12} Nf6-g4{170}
8. Nd2-b3{18} Bf8-g7{224}
9. h2-h3{14} Ng4-e5{156}
10. d4-e5{18} Nc6-e5{134}
11. Nf3-e5{16} Bg7-e5{138}
12. Rf1-e1{8} Qd8-f6{84}
13. c2-c3{10} Be5-d6{110}
14. Bc1-h6{16} Ke8-c8{106}
15. Qd1-d2{12} Qf6-e7{82}
16. Nb3-d4{68} e6-e5{102}
17. Nd4-b5{20} e5-e4{122}
18. Nb5-d6{26} Qe7-d6{154}
19. Bd3-c2{14} Kc8-b8{154}
20. Bh6-f4{14} Qd6-c5{94}
21. Bf4-g5{26} Rd8-e8{106}
22. Bc2-b3{20} Bd7-e6{118}
23. Ra1-d1{18} c7-c6{130}
24. Bg5-e3{18} Qc5-b5{172}
25. Qd2-d4{20} Qb5-a6{104}
26. Qd4-c5{16} Re8-c8{102}
27. Rd1-d4{20} b7-b6{176}
28. Rd4-a4{16} Qa6-b7{234}
29. Be3-f4{20} Kb8-a8{120}
30. Qc5-d4{16} c6-c5{136}
31. Qd4-f6{18} b6-b5{96}
32. Ra4-a5{14} c5-c4{390}
33. Bb3-c2{14} b5-b4{90}
34. Bf4-e3{16} b4-c3{108}
35. b2-c3{44} Rc8-b8{116}
36. Qf6-d4{20} a7-a6{96}
37. Re1-b1{12} Qb7-b1{162}
38. Bc2-b1{20} Rb8-b1{192}
39. Kg1-h2{8} Ka8-b7{116}
40. Qd4-h8{24} Rb1-b2{88}
41. Ra5-a6{16} Kb7-c7{116}
42. Be3-f4{14} Kc7-d7{82}
---------------------------------------------------------------------------=
=AD-----
Rybka : (White)
normal: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=3DDM20434&game=3D=
Chess
In this game GetClub was able to bring a good defence but later Rybka
broke the defence and took the Queen for Rook.
And due to a Mating Combination GetClub has to loose its Rook.
GetClub played wonderful moves and for sometime I thought the game
will be a draw. But Rybka managed to intrude the defence.
A very nice game.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
|
| |
Date: 02 Jun 2008 08:14:56
From: Sanny
Subject: Re: GetClub Chess now twice better.
|
> If, as you say, they play as well as each other. Why is it I've never seen
> anyone post a game whereGetClubbeats Rybka yet?
>
> I'm not really interested in that either really. It's a rhetorical question.
>
Rybka always finds some way to win the game. Only 1 game was played
after the improvement. And it hold for 43 moves.
I think I need to play a few more games and then we can really see how
they face each other.
Earlier when I used to play against Normal Level. Rybka got advantage
very quickly. In this game Normal Level was able to stop Rybka a lot
longer. And lost after 43 moves.
Only after a few more games I can clearly say the difference. Normal
takes 16 times more time against Rybka. So If GetClub takes 20 times
more time the game will always be equal.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| | |
Date: 02 Jun 2008 18:22:46
From: David Richerby
Subject: Re: GetClub Chess now twice better.
|
Sanny <[email protected] > wrote:
> Rybka always finds some way to win the game. Only 1 game was played
> after the improvement. And it hold for 43 moves.
So, if Rybka always wins the games, that means that Rybka is better,
doesn't it?
Dave.
--
David Richerby Incredible Dangerous Apple (TM):
www.chiark.greenend.org.uk/~davidr/ it's like a tasty fruit but it could
explode at any minute and it'll blow
your mind!
|
| | | |
Date: 02 Jun 2008 13:11:12
From: Guest
Subject: Re: GetClub Chess now twice better.
|
"David Richerby" <[email protected] > wrote in message
news:UID*[email protected]...
> Sanny <[email protected]> wrote:
>> Rybka always finds some way to win the game. Only 1 game was played
>> after the improvement. And it hold for 43 moves.
>
> So, if Rybka always wins the games, that means that Rybka is better,
> doesn't it?
Only 1 game was played after the improvement.... Gee, just one game. And
other programmers do hundreds or even thousands of games to determine who's
better and by how much. I wonder what a statistician would say about the
error range for a single datum point?
And how is it lasting 43 moves actually significant... A loss is a loss.
Rybka could just be hunting for more raterial to raise its scores. At
reduced time controls, programs have been known to get greedy rather than
going for a win that is just outside of their search depth.
And if you increase the time for each side by a factor of 10, does GetClub
still match up? Hard to say with Rybka at just 5 seconds a move.
And just what is Rybka's strength at 5sec a move? That's got to be a pretty
drastic reduction. 5 seconds barely gives it enough time to do much of
anything, and certainly not enough time for it to reach its full abilities.
Programs play a lot differently with such drastically reduced search times.
Sure, in blitz games against humans a program's strength goes up. But
against a computer... No.
Would uMax or TSCP or Gnu Chess v1 beat Rybka if they were given 100 seconds
like GetClub?? (grin)
Heck, I might be willing to give odds to Sargon 1 at those time differences.
Especially if all it has to do is last to the 43rd move.
Hey, I have one of those keychain chess computers from Excalibur. Maybe I
should pit it against GetClub chess at those time differences....
No, I'm not actually going to. I don't have any desire to actually go to
GetClub and play chess. The only reason I'd go there is if they described
in detail the chess program there. Unless there is something truely unusual
about the program (selective search, pure knowledge based approach, neural
network, etc.), then I really don't see anything significant about it.
There are too many better programs that are public domain, or MIT or even
GPL'ed.
If the program did indeed use some unusual technique, then I'd be
interested. It's always interesting to hear about some unusual approach.
----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
|
| |
Date: 02 Jun 2008 07:46:53
From: Sanny
Subject: Re: GetClub Chess now twice better.
|
> > ie Rybka think in 1 second what GetClub could in 20 seconds.
>
> Just because Rybka is making it's move 20 times faster doesn't directly
> equate how much weaker GetClub is. Because, not only is Rybka making it's
> move 20 times faster... the moves it makes are also *better*.
>
> If GetClub played the same moves it does now, but made them faster than
> rybka did. It would still be weaker than Rybka because the moves it makes
> aren't as good. So no, GetClub *isn't* 20 times weaker than Rybka, it's much
> weaker than that.
Rybka at 5 seconds/ move plays as good as Normal Level at 100
seconds / move.
Both plays comparably well. So I think GetClub is 20 times slower. Its
simple maths. Have a few match with Normal Levels and you will see
both Rybka and Normal Level plays equally good.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| | |
Date: 02 Jun 2008 16:00:52
From: ave
Subject: Re: GetClub Chess now twice better.
|
> Both plays comparably well. So I think GetClub is 20 times slower. Its
> simple maths. Have a few match with Normal Levels and you will see
> both Rybka and Normal Level plays equally good.
I'm not interested in playing GetClub, nor comparing engines for you.
If, as you say, they play as well as each other. Why is it I've never seen
anyone post a game where GetClub beats Rybka yet?
I'm not really interested in that either really. It's a rhetorical question.
ave
|
| |
Date: 01 Jun 2008 07:38:03
From: Sanny
Subject: Re: GetClub Chess now twice better.
|
> > A few modifications were done in GetClub Chess and now it plays Twice
> > better moves.
>
> It would be interesting to search Google Groups and find out exactly
> how many times Sanny has advertised improvements, often huge
> improvements, to GitClub. =A0After so many giant leaps, one would think
> that GitClub would truly beat Rybka (and everything and everyone else)
> like carrots. =A0Somehow, though, that hasn't happened. =A0Could (heaven
> forbid) Sanny be ever so slightly full of baloney?
What other Programs Fritz/ Chess Master/ Rybka etc Achieved in 20
years GetClub did that much in 2 years So natuarally you will see more
improvements about GetClub Chess.
8 months back GetClub was 400 times weaker than Rybka.
ie Rybka think in 1 second what GetClub could in 400 seconds.
5 months Back GetClub was 200 times weaker than Rybka.
ie Rybka think in 1 second what GetClub could in 200 seconds.
2 months Back GetClub was 60 times weaker than Rybka
ie Rybka think in 1 second what GetClub could in 60 seconds.
1 month back GetClub was 40 times weaker than Rybka
ie Rybka think in 1 second what GetClub could in 40 seconds.
Now GetClub is 20 times weaker than Rybka.
ie Rybka think in 1 second what GetClub could in 20 seconds.
So now GetClub takes 20 seconds to think the same move what Rybka
thinks in 1 seconds.
Now, It is very difficult to improve further. Only if good players
suggest their best advices then GetClub can be improved. Else It will
remain what it is now.
But now it is strong enough to challenge any human player.
Today a game with Par Excellence was Drawn. Earlier he always used to
win the game. So now you will really face good Challenge at GetClub
Chess.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| | |
Date: 02 Jun 2008 15:32:12
From: ave
Subject: Re: GetClub Chess now twice better.
|
> Now GetClub is 20 times weaker than Rybka.
No it isn't.
Look...
> ie Rybka think in 1 second what GetClub could in 20 seconds.
Just because Rybka is making it's move 20 times faster doesn't directly
equate how much weaker GetClub is. Because, not only is Rybka making it's
move 20 times faster... the moves it makes are also *better*.
If GetClub played the same moves it does now, but made them faster than
rybka did. It would still be weaker than Rybka because the moves it makes
aren't as good. So no, GetClub *isn't* 20 times weaker than Rybka, it's much
weaker than that.
> So now GetClub takes 20 seconds to think the same move what Rybka
> thinks in 1 seconds.
But.. it doesn't make the same moves Rybka makes, does it?
ave
|
| |
Date: 01 Jun 2008 07:17:13
From: [email protected]
Subject: Re: GetClub Chess now twice better.
|
On Jun 1, 7:11 am, Sanny <[email protected] > wrote:
> A few modifications were done in GetClub Chess and now it plays Twice
> better moves.
It would be interesting to search Google Groups and find out exactly
how many times Sanny has advertised improvements, often huge
improvements, to GitClub. After so many giant leaps, one would think
that GitClub would truly beat Rybka (and everything and everyone else)
like carrots. Somehow, though, that hasn't happened. Could (heaven
forbid) Sanny be ever so slightly full of baloney?
|
| | |
Date: 01 Jun 2008 18:01:08
From: David Richerby
Subject: Re: GetClub Chess now twice better.
|
[email protected] <[email protected] > wrote:
> On Jun 1, 7:11 am, Sanny <[email protected]> wrote:
>> A few modifications were done in GetClub Chess and now it plays
>> Twice better moves.
>
> It would be interesting to search Google Groups and find out exactly
> how many times Sanny has advertised improvements, often huge
> improvements, to GitClub.
No it wouldn't.
Dave.
--
David Richerby Expensive Poisonous Goldfish (TM):
www.chiark.greenend.org.uk/~davidr/ it's like a fish but it'll kill you
in seconds and break the bank!
|
|
