| |
Main
Date: 07 Jun 2008 09:34:10
From: Sanny
Subject: What is Rating of Rybka and other Programs?
|
I read on internet Rybka is 3100 rated. And Fritz is 2900 Rated is it
true?
When they say Rybka is 3100 Rated How much time they give it for each
Move.
As I played Rybka (5 sec / move) against Normal Level (100 sec / move)
and found both play equally well.
So when Rybka is playing at 5 sec / move is it playing with 3100
Strength? What is the time per move at which Rybka Rating equals 3100?
What are the Programs which Rank 2ns and 3rd and what is their
ratings?
Crafty, ChessMaster, Fritz who ranks first and what is rating of each?
What is the Rank of GetClub Chess in all engines?
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
|
| |
Date: 19 Jun 2008 20:55:10
From: help bot
Subject: Re: Help Bot vs Advance Level
|
On Jun 18, 11:41 pm, Frisco Del Rosario <[email protected] > wrote:
> > True. But there were *many* who predicted
> > that a chess-playing machine could never beat
> > strong human players, and when proved wrong,
> > these folks repeatedly "moved the goalposts"
> > rather than admit their titanic error.
> The goalposts moved in very predictable fashion. Humans said, "Computers
> are getting better, but they'll never get above the class right beneath
> mine," until that last human at the top of the heap said it.
That accurately describes many of the Experts,
masters, IMs and grandmasters who predicted
failure, but not all the prognosticators were strong
players themselves; indeed, some of those who
sided with humans against the machines were
computer programmers-- supposed experts in
the field.
It reminds me of the way in which supposed
experts have been wrong so many times, about
so many other things. I can still recall a chess
game where comments were written by various
grandmasters criticizing a fellow GM for not
holding a draw as they said he ought to have;
around a month or so later another article
appeared, this time written by the winner of that
game, proving it was a /forced win/. In sum, it
often happens that the level of expertise of an
"expert" is strongly countered -- perhaps even
overwhelmed -- by the expert's own arrogance.
In the case of computers versus humans, it
was a matter of looking at how poor the results
of early efforts were, and extrapolating from
that, and the high cost of hardware, that the
distance was akin to man traveling to the stars.
Assumptions were made about the future cost
of hardware, the supposed inability to improve
dramatically on programming a chess engine,
etc., and these turned out to be wrong. But
it doesn't require an "expert" to make such
assumptions, does it? As I see it, the only
people who are farther off course than those
self-described experts were, are those who
try to argue that chess is way "out of range"
using funky math, and similar assumptions.
Let's say there are sixteen kazillion possible
chess positions, and fourteen kazillion of
them are legal. Assuming (here we go...)
that in the average position there are 29.73
legal moves, that means you cannot store
the resulting 12^524 positions even if you
could use all the atoms of the known (here
we go again...) universe. Silent assumptions
include, but are not limited to, optimal use of
atoms is what we know today; physical
things are the only way to store information;
everything must be stored at once; etc., etc.
Let me just be the first on record to predict
that computers will never play better than
/perfect chess/. Oh, they may find ways to
improve results against humans or inferior
computers via inferior moves or strategies,
but their play will not be better objectively
than /merely perfect/. The goalpost stops
here.
-- help bot
|
| |
Date: 18 Jun 2008 19:47:06
From: help bot
Subject: Re: Help Bot vs Advance Level
|
On Jun 16, 9:13 am, Martin Brown <
|
| | |
Date: 18 Jun 2008 20:41:42
From: Frisco Del Rosario
Subject: Re: Help Bot vs Advance Level
|
In article
<[email protected] >,
help bot <[email protected] > wrote:
> True. But there were *many* who predicted
> that a chess-playing machine could never beat
> strong human players, and when proved wrong,
> these folks repeatedly "moved the goalposts"
> rather than admit their titanic error.
The goalposts moved in very predictable fashion. Humans said, "Computers
are getting better, but they'll never get above the class right beneath
mine," until that last human at the top of the heap said it.
Well, it's been a couple hours... GetDumb must've been improved by
another factor of 2.
|
| |
Date: 18 Jun 2008 19:14:52
From: help bot
Subject: Re: Game with Rybka
|
On Jun 18, 1:48 pm, SBD <[email protected] > wrote:
> Hey, "hep bot"...
>
> You encouraging focus is a bit like Larry Parr encouraging honesty. :)
Hey, at least I *recognize* when the topic has
veered far, far away to some distant galaxy,
completely unrelated to chess! Mr. Parr wants
to discuss what the world might be like if the
Great War (that's what we called it then,
anyway) had taken some imaginary, different
course; an interesting subject perhaps, but
this is a /chess/ newsgroup (it is rumoured
that LP played the game in his younger days).
As for Mr. Parr's rejection of honesty, I note
that he claimed to have done an analysis of
the syntax of my posts, matching them up
with some old postings. When challenged to
present this alleged analysis, the man went to
radio silence-- just as he has in the past when
caught in other lies.
But back to what we were talking about
before: Focus Factor; I think this is obviously
one of the "drugs" which refutes the silly claim
by LP that "no drug has ever helped anyone to
win a game of chess", or something to that
effect. Or were we discussing the contempt
factor? The weather? My water bill is down
this month compared to the same month last
year, because I have been too busy getting
washed away by the monsoons to even
consider watering my yard. And how about
those Red Sox! Did you say I had trouble
focusing? Try putting food in front of me, and
just sit back and watch how I "home in"... .
-- (Um, does anyone remember my handle?)
|
| |
Date: 18 Jun 2008 10:48:34
From: SBD
Subject: Re: Game with Rybka
|
On Jun 18, 12:03 pm, help bot <[email protected] > wrote:
Hey, "hep bot"...
You encouraging focus is a bit like Larry Parr encouraging honesty. :)
|
| |
Date: 18 Jun 2008 10:03:58
From: help bot
Subject: Re: Game with Rybka
|
On Jun 18, 9:16 am, SBD <[email protected] > wrote:
> So you are recommending that GetClub add arrogance to its ignorance
> factor?
This guy seems unable to follow even the
simplest of threads... .
Try to focus, for just a moment: Sanny asked a
*question*, and I responded with an attempt to
explain what a "contempt factor" is and what it's
used for. I did *not* make any such
recommendation, except with regard to a purely
hypothetical odds match between two programs.
(It is and perhaps shall remain hypothetical
because Sanny's games are played against a
remote server, which probably does not allow
one to remove pieces before commencing play.)
> In humans, hardly a good combination, can't imagine that helps a
> computer anymore.... especially one that can't handle legal moves.
Stupid is as stupid does... (sigh). The contempt
factor was supposed to aid /Rybka/, by keeping it
from seeking premature draws when giving odds.
Sanny's program does not even have this function.
Even though SBD apparently cannot follow the
simplest of threads, he did stumble upon one
intelligent idea: the combination of ignorance
with arrogance, in humans, is not a good thing,
in spite of its growing popularity. One
illuminating example of what can happen was
the clash between Edward Winter and Larry
Evans over wrong dates-- a simple skirmish
which morphed into all-out war, leaving one side
deeply wounded and the other looking rather
petty for having skewered him over spelling and
dates, rather than some greater cause.
I leave it to the experts to explain how a
contempt factor may best be implemented; it's
a rather unimportant thing, relative to improving
GetClub's (mis)handling of tactics. I don't wish
to quibble over minutia-- like the fact that the
comment at top is phrased as a statement, yet
it has a question mark instead of a period at the
end... and "recommending" should have three
m's and a k... . ; >D
-- hep blot
|
| |
Date: 18 Jun 2008 06:16:01
From: SBD
Subject: Re: Game with Rybka
|
On Jun 18, 2:26 am, help bot <[email protected] > wrote:
> -- help bot
So you are recommending that GetClub add arrogance to its ignorance
factor?
In humans, hardly a good combination, can't imagine that helps a
computer anymore.... especially one that can't handle legal moves.
|
| |
Date: 18 Jun 2008 00:26:19
From: help bot
Subject: Re: Game with Rybka
|
On Jun 18, 12:01 am, Sanny <[email protected] > wrote:
> > My suggestion would be material odds, to
> > level the playing field a bit. If you were to
> > remove Rybka's QR, advance her QRp one
> > square, as is customary, and set her
> > contempt factor to, say, 3, we might get an
> > interesting match whereGetClubhas a
> > sporting chance.
>
> What is Contempt Factor?
Let's say you have complete control of two chess
programs, and you remove Rybka's QR for an odds
game. Problem: Rybka thinks she is a Rook down,
and will actively seek draws, not realizing that her
opponent is a complete patzer.
Solution: the contempt factor. This is where you
instruct the program (here, Rybka) to "think" it is
not down a Rook, even though it is. In making
position evaluations, the contempt factor is, um,
factored in so to speak. If the position becomes
level, unfortunately the program will now "think" it
is up a Rook, due to the contempt factor you
programmed in. It's a tricky business, as we saw
in some of the odds matches wherein the Rybka
program appeared to be weakened by improper
handling of both the contempt factor and the
special openings book. But without any such
modification, odds matches won't work out due
to the stronger program actively seeking draws,
before it has a decent chance to win.
In the case of an odds match between Rybka
and GetClub, Rybka needs to be told not to
seek draws, in spite of her material deficit; and
the GC program needs to be told to do the
opposite: seek draws in spite of being up a bit
of material. When I say "draws", I mean here
threefold repetitions of position, not trying to
forcibly trade off all the pieces and pawns--
which is a tall order indeed.
-- help bot
|
| |
Date: 18 Jun 2008 16:35:35
From: nobody
Subject: Re: Game with Rybka
|
Sanny wrote:
>
> > My suggestion would be material odds, to
> > level the playing field a bit. If you were to
> > remove Rybka's QR, advance her QRp one
> > square, as is customary, and set her
> > contempt factor to, say, 3, we might get an
> > interesting match whereGetClubhas a
> > sporting chance.
>
> What is Contempt Factor?
I think it might accurately describe how most people here feel about
your crappy chess program Sanny..
|
| |
Date: 17 Jun 2008 21:01:55
From: Sanny
Subject: Re: Game with Rybka
|
> =A0 My suggestion would be material odds, to
> level the playing field a bit. =A0If you were to
> remove Rybka's QR, advance her QRp one
> square, as is customary, and set her
> contempt factor to, say, 3, we might get an
> interesting match whereGetClubhas a
> sporting chance.
What is Contempt Factor?
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 17 Jun 2008 18:52:43
From: help bot
Subject: Re: Game with Rybka
|
On Jun 17, 2:21 pm, Sanny <[email protected] > wrote:
> Heres a game between Rybka & Normal Level.
>
> For the first time GetClub took White. Rybka was able to win in 41
> moves only.
>
> Can you spot mistakes in the game?
Yes.
> White -- Black
> (normal) -- (Rybka )
>
> 1. g2-g3{6} d7-d5{38}
> 2. d2-d3{82}
There's one.
> 2. ... Nb8-c6{34}
> 3. Bf1-g2{248} Ng8-f6{192}
> 4. Ng1-f3{94} g7-g6{30}
> 5. Bc1-f4{324} Bf8-g7{26}
> 6. Nb1-c3{104}
There's another.
? 6. ... d5-d4{22}
> 7. Nc3-b5{88}
And another.
> 7. ... Nf6-d5{12}
> 8. Bf4-g5{84} a7-a6{16}
> 9. Nb5-a3{122} Ke8-g8{18}
> 10. Ke1-g1{80} b7-b5{18}
> 11. Qd1-c1{84} Bc8-g4{22}
> 12. Bg5-h6{134} Qd8-d6{18}
> 13. Bh6-g7{132} Kg8-g7{26}
> 14. Qc1-g5{84}
And another.
> 14.... f7-f5{16}
> 15. h2-h3{124} Bg4-f3{22}
> 16. Bg2-f3{116} e7-e5{22}
> 17. Qg5-h4{142} Ra8-e8{18}
> 18. Ra1-c1{204}
Another one.
> 18. ... Nc6-b4{30}
> 19. Kg1-h1{118} Nb4-a2{20}
> 20. Rc1-a1{104} Na2-b4{24}
> 21. Bf3-g2{132} Qd6-c5{30}
> 22. Bg2-f3{114} Rf8-f6{28}
> 23. b2-b3{258}
Yet another one.
> 23. ... Nd5-c3{60}
> Hows the game. Kiiks like Rybka is playing very strong game.
The White Queen was lucky not to get trapped
in this game; probably, Rybka wasn't looking
very deep; I noticed that no thinking time was
specified and in the past, the times noted were
unreliable indicators.
> Was there a way for GetClub to save its pawn at a2?
Yes; just leave the Rook at a1, where it already
was. The move p-b3 created a gaping hole on c3,
as we saw. After 23. ... Nc3, 24. d4 White's
pieces are driven back (or trapped).
I have not criticized any of Rybka's moves
because it is obvious that she was not given
very much time to think; by simply running
the game through Rybka with more time, you
could easily find her many improvements.
My suggestion would be material odds, to
level the playing field a bit. If you were to
remove Rybka's QR, advance her QRp one
square, as is customary, and set her
contempt factor to, say, 3, we might get an
interesting match where GetClub has a
sporting chance.
-- help bot
|
| |
Date: 17 Jun 2008 11:21:31
From: Sanny
Subject: Game with Rybka
|
Heres a game between Rybka & Normal Level.
For the first time GetClub took White. Rybka was able to win in 41
moves only.
Can you spot mistakes in the game?
Game Played between Rybka and normal at GetClub.com
--------------------------------------------------------------------------------
Rybka : (Black)
normal: (White)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM20839&game=Chess
--------------------------------------------------------------------------------
White -- Black
(normal) -- (Rybka )
1. g2-g3{6} d7-d5{38}
2. d2-d3{82} Nb8-c6{34}
3. Bf1-g2{248} Ng8-f6{192}
4. Ng1-f3{94} g7-g6{30}
5. Bc1-f4{324} Bf8-g7{26}
6. Nb1-c3{104} d5-d4{22}
7. Nc3-b5{88} Nf6-d5{12}
8. Bf4-g5{84} a7-a6{16}
9. Nb5-a3{122} Ke8-g8{18}
10. Ke1-g1{80} b7-b5{18}
11. Qd1-c1{84} Bc8-g4{22}
12. Bg5-h6{134} Qd8-d6{18}
13. Bh6-g7{132} Kg8-g7{26}
14. Qc1-g5{84} f7-f5{16}
15. h2-h3{124} Bg4-f3{22}
16. Bg2-f3{116} e7-e5{22}
17. Qg5-h4{142} Ra8-e8{18}
18. Ra1-c1{204} Nc6-b4{30}
19. Kg1-h1{118} Nb4-a2{20}
20. Rc1-a1{104} Na2-b4{24}
21. Bf3-g2{132} Qd6-c5{30}
22. Bg2-f3{114} Rf8-f6{28}
23. b2-b3{258} Nd5-c3{60}
24. g3-g4{214} e5-e4{136}
25. g4-g5{84} Rf6-d6{20}
26. d3-e4{154} f5-e4{28}
27. Qh4-h6{336} Kg7-g8{18}
28. Bf3-g4{130} d4-d3{16}
29. e2-d3{150} e4-d3{18}
30. Ra1-c1{222} d3-d2{20}
31. Rc1-a1{112} Nc3-e4{20}
32. Kh1-h2{110} Ne4-f2{20}
33. Bg4-f3{142} Re8-f8{30}
34. Bf3-b7{248} Qd2-d1{Q}{20}
35. Ra1-d1{92} Rd6-d1{18}
36. Rf1-d1{388} Nf2-d1{24}
37. Qh6-h4{98} Nb4-c2{32}
38. Na3-c2{114} Qc5-c2{96}
39. Bb7-g2{130} Nd1-e3{28}
40. Qh4-e4{148} Ne3-f1{34}
41. Kh2-h1{238}
--------------------------------------------------------------------------------
Rybka : (Black)
normal: (White)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM20839&game=Chess
Hows the game. Kiiks like Rybka is playing very strong game.
Was there a way for GetClub to save its pawn at a2?
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 17 Jun 2008 10:03:53
From: help bot
Subject: Re: GetClub Strength increased Twice.
|
On Jun 17, 11:41 am, Sanny <[email protected] > wrote:
> > I just started another game with the much "improved"
> > Normal level. After something like: 1. d4 Nf6, 2. c4 c5,
> > 3. d5 e5, 4. Nc3 d6, the program went bonkers,
> > insisting that its next move was 5. R-h1, and freezing
> > my computer (I intended to meet that with 5. ... Nc8!,
> > which keeps the tension while *not* losing the right to
> > castle King-side, as White just did).
>
> Looks like some Bug. Same problem happened with Zebediah Game. He
> Played Black but Computer Played R-h1 as first Move. So Zebediah
> continued the game with Black Pieces First.
>
> Zebediahs game:http://www.getclub.com/playgame.php?id=DM20812&game=Chess
>
> I think that is due to some disconnection.
>
> 2 things are improtant This happend today to two games. Must be some
> micoscopic bug eating the threads of the GetClub Program.
>
> Searching such bugs is like searching a fish in a Big River.
When I finally finished that game, I was able to play
over it without any problem, so maybe the R-h1 move
repaired itself when I rebooted and logged back on.
> But a few precautionary majors were taken So that It do not make such
> silly moves again.
>
> Next time such thing happen do remind me.
>
> This Bug has been partially removed. Since the correct reasion of
> thgis happening is not known.
I have noticed that, unlike most other programs,
with GetClub the Java applet loads very slowly,
and sometimes not at all unless I am looking at
that particular Windows tab. It seems as though
the applet is not grabbing enough system
resources to start properly, but once a game is
in progress, it hogs plenty enough to make my
computer's fan kick on and run non-stop.
This can be very complicated, I expect, since
you are dealing not only with a chess engine
but also the GUI, the user's Web browser and
the Windows OS.
As mentioned before, the simple workaround
for chess' rule that the King cannot be captured
is to assign Kings a very large value relative to
all the other men. If you still have difficulties
in defending against short-range mates, then
there must be something horribly wrong with
the way the program handles tactics (the most
critical aspect of the game). Even a twenty
year old program running on archaic hardware,
should be able to avoid unforced mates by
simply using check-and-capture extensions,
as I read about many, many years ago-- long
before Google, Yahoo!, Vista, XP, Windows95,
etc., were ever even thought of. In fact, I read
about this stuff before rgc even existed. My
guess is that Sanny has not correctly
implemented this technique; his "expert"
programmers may not know how to do it,
because it is chess-specific.
-- help bot
|
| |
Date: 17 Jun 2008 08:41:29
From: Sanny
Subject: Re: GetClub Strength increased Twice.
|
> =A0 I just started another game with the much "improved"
> Normal level. =A0After something like: 1. d4 =A0Nf6, 2. c4 =A0c5,
> 3. d5 =A0e5, =A04. Nc3 =A0d6, the program went bonkers,
> insisting that its next move was 5. R-h1, and freezing
> my computer (I intended to meet that with 5. ... Nc8!,
> which keeps the tension while *not* losing the right to
> castle King-side, as White just did).
Looks like some Bug. Same problem happened with Zebediah Game. He
Played Black but Computer Played R-h1 as first Move. So Zebediah
continued the game with Black Pieces First.
Zebediahs game: http://www.getclub.com/playgame.php?id=3DDM20812&game=3DChes=
s
I think that is due to some disconnection.
2 things are improtant This happend today to two games. Must be some
micoscopic bug eating the threads of the GetClub Program.
Searching such bugs is like searching a fish in a Big River.
But a few precautionary majors were taken So that It do not make such
silly moves again.
Next time such thing happen do remind me.
This Bug has been partially removed. Since the correct reasion of
thgis happening is not known.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 16 Jun 2008 10:38:21
From: help bot
Subject: Re: GetClub Strength increased Twice.
|
On Jun 16, 12:49 pm, Bjoern <[email protected] > wrote:
> > You do realize that in chess -- unlike checkers --
> > the King is not something which can be sacrificed?
> > Hence, a chess program should give very heavy
> > weight to the defense thereof. In my ultra-simple
> > materialistic theoretical program, the value of the
> > King could be, say, 99 (or 999), thus preventing its
> > "loss" whenever within the program's meager power.
>
> > A properly-working program will reveal that it
> > values the King properly by throwing away all the
> > other pieces first, in order to push the loss beyond
> > its horizon or at least near the edge of it.
>
> Indeed. Additionally a simple refinement is being mated now = -999,
> mated in 1 = -998, mated in 2 = -997, mated in 3 = -996, mated in 4 =
> -995 etc., so as to pick the defence that avoids mate for the longest time.
Those numbers only work for mediocre programs like
Fritz, Hiarcs, Rybka and Shredder. The GetClub
program is so "improved" that Sanny would probably
have to add (at least) another digit; thus, --9999 = mate
now; --9998 = mate tomorrow; --9997 = mate the next
day; etc. Remember: only Sanny's program has been
"improved" by a factor of sixty-three million, and it is
the only one which can see even deeper than the
Hubble telescope.
I just started another game with the much "improved"
Normal level. After something like: 1. d4 Nf6, 2. c4 c5,
3. d5 e5, 4. Nc3 d6, the program went bonkers,
insisting that its next move was 5. R-h1, and freezing
my computer (I intended to meet that with 5. ... Nc8!,
which keeps the tension while *not* losing the right to
castle King-side, as White just did).
-- help bot
|
| |
Date: 16 Jun 2008 08:48:34
From: help bot
Subject: Re: GetClub Strength increased Twice.
|
On Jun 16, 1:19 am, Sanny <[email protected] > wrote:
> > > Just Play with Normal for a good Challenge. And if you are very Strong
> > > then play with Master Level.
>
> > Here is my latest effort against the Normal level:
>
> >http://www.getclub.com/playgame.php?id=DM20722&game=Chess
>
> > I completely overlooked the move ...Nc6, and this
> > cost me two pawns. However, the program did not
>
> This Game was played before the improvement. Yesterday an improvement
> was done so that it plays Twice Strong. After that you haven't yet
> played a game.
That may well be true... but it was played after some
HUNDREDS of other supposed "improvements" were
done, so I doubt very much if the most recent one is
going to make any significant difference.
Looked at another way: if the program cannot see a
looming mate coming just two moves in advance, the
number of supposed "improvements" isn't going to
matter; that's because I can easily set up such
threats, even if I am losing badly. In my most recent
game, I could very well have eventually lost due to my
oversight (...Nc6), but this would require decent play
and not just munching on pawns, only to allow itself
to be checkmated a few moves later.
You do realize that in chess -- unlike checkers --
the King is not something which can be sacrificed?
Hence, a chess program should give very heavy
weight to the defense thereof. In my ultra-simple
materialistic theoretical program, the value of the
King could be, say, 99 (or 999), thus preventing its
"loss" whenever within the program's meager power.
A properly-working program will reveal that it
values the King properly by throwing away all the
other pieces first, in order to push the loss beyond
its horizon or at least near the edge of it.
-- help bot
|
| | |
Date: 16 Jun 2008 18:49:29
From: Bjoern
Subject: Re: GetClub Strength increased Twice.
|
help bot wrote:
> You do realize that in chess -- unlike checkers --
> the King is not something which can be sacrificed?
> Hence, a chess program should give very heavy
> weight to the defense thereof. In my ultra-simple
> materialistic theoretical program, the value of the
> King could be, say, 99 (or 999), thus preventing its
> "loss" whenever within the program's meager power.
>
> A properly-working program will reveal that it
> values the King properly by throwing away all the
> other pieces first, in order to push the loss beyond
> its horizon or at least near the edge of it.
Indeed. Additionally a simple refinement is being mated now = -999,
mated in 1 = -998, mated in 2 = -997, mated in 3 = -996, mated in 4 =
-995 etc., so as to pick the defence that avoids mate for the longest time.
|
| |
Date: 16 Jun 2008 07:24:19
From: [email protected]
Subject: Re: GetClub Strength increased Twice.
|
> I did a bit of calculating, and it turns out that after
> factoring in all the many improvements announced
> here in rgc by Sanny, the GetClub program weighs
> in with a rating of 6,219,847.21 USCF, which is
> not too bad.
I find that hard to believe as it is larger than an unsigned 32-bit
integer. I think GitClub has maxed out at 4 294 967 296 USCF and now
has nowhere to go until Sanny either moves to 64-bit architecture or
recodes with arbitrary precision arithmetic.
|
| |
Date: 15 Jun 2008 22:19:09
From: Sanny
Subject: Re: GetClub Strength increased Twice.
|
> > Just Play with Normal for a good Challenge. And if you are very Strong
> > then play with Master Level.
>
> =A0 Here is my latest effort against the Normal level:
>
> http://www.getclub.com/playgame.php?id=3DDM20722&game=3DChess
>
> =A0 I completely overlooked the move ...Nc6, and this
> cost me two pawns. =A0However, the program did not
This Game was played before the improvement. Yesterday an improvement
was done so that it plays Twice Strong. After that you haven't yet
played a game.
Play with Normal Level and see how you fair it.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 15 Jun 2008 07:23:02
From: help bot
Subject: Re: GetClub Strength increased Twice.
|
On Jun 15, 5:26 am, Sanny <[email protected] > wrote:
> Today again the Strength of GetClub Chess was increased 2 times by
> removing a bug.
I did a bit of calculating, and it turns out that after
factoring in all the many improvements announced
here in rgc by Sanny, the GetClub program weighs
in with a rating of 6,219,847.21 USCF, which is
not too bad.
> Now, I think Normal Level will play Stronger than Rybka at 5-sec/
> move.
That's probably a true statement.
> Now, GetClub is 8 times weaker than Rybka.
Possibly also true.
> So what Rybka thinks in 5
> seconds GetClub is able to think in 5*8=40 seconds.
A non sequitur (i.e. it does not follow, logically).
> I will try a game with Normal Level and see if Normal wins the game or
> not. As according to my estimates Normal will play Stronger than Rybka
> at 5-sec/ move.
Your estimates need work.
> So Help Bot will find it much difficult to win the Normal & Master
> Levels. Remember never play with Advance Level as it may think for
> Hours together on some moves.
>
> Just Play with Normal for a good Challenge. And if you are very Strong
> then play with Master Level.
Here is my latest effort against the Normal level:
http://www.getclub.com/playgame.php?id=DM20722&game=Chess
I completely overlooked the move ...Nc6, and this
cost me two pawns. However, the program did not
see that I was threatening a mating attack-- even
when it was within two moves of being checkmated.
One issue I have noticed with the program is a
strong preference for premature Queen moves
(like 2. Qd3 in this game). So, not only is there a
severe problem with its endgame, but at the very
opposite end-- the opening, it has some serious
problems. Nevertheless, tactics remain the key
issue, as far as I can tell. I am still easily besting
the program in tactical skirmishes, which is
absurd since I do not (yet) have an AMD Turion64
microprocessor, while the GetClub program does
(when run on my machine). While it is true that I
am a bazillion times smarter than this program
overall, it should at least be able to beat me at
simple math (i.e. tactics).
-- help bot
|
| |
Date: 15 Jun 2008 03:38:22
From: Sanny
Subject: Re: GetClub Strength increased Twice.
|
> I will try a game with Normal Level and see if Normal wins the game or
> not. As according to my estimates Normal will play Stronger than Rybka
> at 5-sec/ move.
>
> Bye
> Sanny
I played a game between Rybka & Normal Level.
I found that GetClub was able to save itself from a trap in which it
used to loose a knight. But later fall in another trap and lost its
knight for nothing.
Do you think Normal Level, Played any wrong move?
Even after the improvement why Normal Level lost to Rybka?
Game Played between sanjay11 and normal at GetClub.com
--------------------------------------------------------------------------------
Rybka: (White)
normal: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM20749&game=Chess
--------------------------------------------------------------------------------
White -- Black
(Rybka) -- (normal)
1. e2-e4{6} e7-e5{0}
2. Ng1-f3{12} Nb8-c6{0}
3. Bf1-b5{14} a7-a6{0}
4. Bb5-a4{12} Ng8-f6{0}
5. Ke1-g1{16} Bf8-c5{120}
6. c2-c3{22} Nf6-e4{150}
7. Qd1-e2{18} f7-f5{118}
8. d2-d3{32} Ne4-f2{162}
9. d3-d4{24} Nf2-e4{134}
10. d4-c5{18} Qd8-f6{126}
11. Qe2-c4{22} Ke8-d8{106}
12. Ba4-c2{20} h7-h6{82}
13. Bc2-e4{12} Nc6-e7{210}
14. Be4-c2{22} Rh8-f8{114}
15. Qc4-e2{18} Ne7-g6{80}
16. Nb1-d2{18} d7-d5{132}
17. c5-d6{16} c7-d6{150}
18. Nd2-c4{16} e5-e4{80}
19. Bc1-e3{20} Ng6-e7{96}
20. Nf3-d2{16} d6-d5{82}
21. Nd2-e4{14} d5-e4{84}
22. Ra1-d1{18} Bc8-d7{94}
23. Nc4-b6{14} Kd8-e8{98}
24. Nb6-d7{16} Qf6-c6{142}
25. Qe2-h5{26} g7-g6{142}
26. Qh5-h6{10} Rf8-g8{80}
27. Bc2-b3{14} Qc6-d7{86}
28. Rd1-d7{18} Ke8-d7{134}
29. Rf1-d1{4} Kd7-e8{124}
30. Qh6-h7{18} Rg8-h8{126}
--------------------------------------------------------------------------------
Rybka: (White)
normal: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM20749&game=Chess
What were the major mistakes in GetClubs game. Why it was not able to
draw/ win the game?
Was there a way out of that Knight Trap? Why do GetClub not take the
Bishop with Pawn when Bishop took the knight.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 15 Jun 2008 02:26:44
From: Sanny
Subject: GetClub Strength increased Twice.
|
> > Simple positional factors are quick to compute compared to the nasty
> > time consuming quiessence search.
>
> =A0 Really? =A0I had the impression that the selective
> search folks were beaten precisely because the
> opposite is true: tactics are not only more crucial,
> they are also less time-consuming to calculate.
Today again the Strength of GetClub Chess was increased 2 times by
removing a bug.
Now, I think Normal Level will play Stronger than Rybka at 5-sec/
move.
Now, GetClub is 8 times weaker than Rybka. So what Rybka thinks in 5
seconds GetClub is able to think in 5*8=3D40 seconds.
I will try a game with Normal Level and see if Normal wins the game or
not. As according to my estimates Normal will play Stronger than Rybka
at 5-sec/ move.
So Help Bot will find it much difficult to win the Normal & Master
Levels. Remember never play with Advance Level as it may think for
Hours together on some moves.
Just Play with Normal for a good Challenge. And if you are very Strong
then play with Master Level.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 14 Jun 2008 15:07:15
From: help bot
Subject: Re: What is Rating of Rybka and other Programs?
|
David Richerby wrote:
> > Indeed. Why not make the engine your engine play on FICS for free?
> GetClub would lose so many games on time that the resulting rating
> would tell you nothing at all about its strength.
Another issue is the fact that when playing at GetClub,
the human player has no time limit; in effect, there is no
real time pressure, unlike in competitive chess.
A lot of posters here have bragged about how easy it
allegedly was for them to beat the computer, but the
fact is they had no time constraints at all, while the
deathly-slow GetClub program did.
I imagine that FICS allows for time controls suitable
for the fastest GetClub levels, although there might be
relatively few takers. Deep down, Sanny probably
knows -- or at least suspects -- that his program
would not do well, so the suggestion to compete on
FICS will likely fall upon deaf ears-- like corncobs.
-- help bot
|
| | |
Date: 15 Jun 2008 12:11:32
From: David Richerby
Subject: Re: What is Rating of Rybka and other Programs?
|
help bot <[email protected] > wrote:
> Deep down, Sanny probably knows -- or at least suspects -- that his
> program would not do well, so the suggestion to compete on FICS will
> likely fall upon deaf ears-- like corncobs.
If there were corncobs falling on Sanny's deaf ears, well, that would
explain a lot.
Dave.
--
David Richerby Cyber-Gerbil (TM): it's like a
www.chiark.greenend.org.uk/~davidr/ children's pet that exists only in
your computer!
|
| | | |
Date: 15 Jun 2008 13:19:10
From: Mike Murray
Subject: Re: What is Rating of Rybka and other Programs?
|
On 15 Jun 2008 12:11:32 +0100 (BST), David Richerby
<[email protected] > wrote:
>help bot <[email protected]> wrote:
>> Deep down, Sanny probably knows -- or at least suspects -- that his
>> program would not do well, so the suggestion to compete on FICS will
>> likely fall upon deaf ears-- like corncobs.
>If there were corncobs falling on Sanny's deaf ears, well, that would
>explain a lot.
Midwestern bot must acknowledge that "rougher than a cob" refers not
to any disturbance and discomfort about the ears, unless, of course,
one is adept at yoga.
|
| |
Date: 14 Jun 2008 09:00:07
From: Sanny
Subject: Game with Bonsai.
|
> 1.d4 Nf6 2.Bg5 Ne4 3.Bf4 e6 4.f3 Bb4+ 5.c3 Bd6 6.Bxd6 Nxd6 7.e4 Nc6
> 8.Nd2 0-0 9.Bd3 Qg5 10.Qe2 f6 11.0-0-0 b6 12.h4 Qa5 13.Kb1 Bb7 14.g4 Qa4
> 15.g5 Rf7 16.Nh3 e5 17.d5 Ne7 18.Rdg1 Ng6 19.h5 Ne7 20.g6 hxg6 21.hxg6
> Rff8 22.Qh2 Nef5 23.Bc2 Qa5 24.exf5 Nf7 25.gxf7+ Kxf7 26.Rxg7+ Kxg7
> 27.Nf4 Kf7 28.Qh7+ Ke8 29.Ng6 Qxa2+ 30.Kc1 Bxd5 1-0
It was good you win Master Level so easily. Looks like Zebediah too
wins using his brain and not Computer. Since I find many strong
players able to kill the Master Level.
Play Chess at: http://www.GetClub.com/Chess.html
> The master level was a real pushover here. Black's (the computer)
> position was fine out of the opening (up to move six), but almost every
> single move from 7 to 15 is mediocre, pointless or just plain poor.
> 18...Ng6 by the computer is then just tactical suicide, how can any
> computer not see that if it has a vaguely reasonable evaluation
What was the correct move. The White Rooks, Bishop and Queen all are
hanging on Black King, So Master Level played this move to give a bit
support to Black King. But has to return back as the pawn advanced.
Black Position was very bad as its king was not save at move 18....
Ng6. What do you think is the correct move?
> function? I guess the answer is thatGetClubdoesn't evaluate positions
> well, at all. Time and again the program just seems to fail to see a
> need to play actively itself, it just plays its pieces about without
> impproving its position or actually threatening anything.
What were the correct moves that should be played instead of what
GetClub Played.
Here is the game Played by you as per the Recorded.
Game Played between bonsai and master at GetClub.com
--------------------------------------------------------------------------------
bonsai: (White)
master: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM20713&game=Chess
--------------------------------------------------------------------------------
White -- Black
(bonsai) -- (master)
1. d2-d4{4} Ng8-f6{0}
2. Bc1-g5{6} Nf6-e4{510}
3. Bg5-f4{14} e7-e6{386}
4. f2-f3{26} Bf8-b4{446}
5. c2-c3{756} Bb4-d6{504}
6. Bf4-d6{70} Ne4-d6{326}
7. e2-e4{4} Nb8-c6{380}
8. Nb1-d2{14} Ke8-g8{366}
9. Bf1-d3{12} Qd8-g5{332}
10. Qd1-e2{20} f7-f6{436}
11. Ke1-c1{18} b7-b6{716}
12. h2-h4{30} Qg5-a5{840}
13. Kc1-b1{34} Bc8-b7{620}
14. g2-g4{46} Qa5-a4{416}
15. g4-g5{142} Rf8-f7{410}
16. Ng1-h3{36} e6-e5{558}
17. d4-d5{48} Nc6-e7{766}
18. Rd1-g1{44} Ne7-g6{408}
19. h4-h5{18} Ng6-e7{1668}
20. g5-g6{48} h7-g6{436}
21. h5-g6{14} Rf7-f8{388}
22. Qe2-h2{12} Ne7-f5{696}
23. Bd3-c2{160} Qa4-a5{368}
24. e4-f5{8} Nd6-f7{332}
25. g6-f7{28} Kg8-f7{644}
26. Rg1-g7{4} Kf7-g7{462}
27. Nh3-f4{16} Kg7-f7{708}
28. Qh2-h7{52} Kf7-e8{414}
29. Nf4-g6{10} Qa5-a2{1862}
30. Kb1-c1{160} Bb7-d5{2}
--------------------------------------------------------------------------------
bonsai: (White)
master: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM20713&game=Chess
So what do yoiu think were the best moves?
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 13 Jun 2008 06:45:33
From: help bot
Subject: Re: Help Bot vs Advance Level
|
On Jun 13, 4:41 am, Martin Brown <
|
| | |
Date:
From: Martin Brown
Subject: Re: Help Bot vs Advance Level
|
|
| |
Date: 12 Jun 2008 10:22:32
From: help bot
Subject: Re: Help Bot vs Advance Level
|
On Jun 12, 12:27 pm, Martin Brown <
|
| | |
Date:
From: Martin Brown
Subject: Re: Help Bot vs Advance Level
|
|
| |
Date: 12 Jun 2008 08:32:23
From: help bot
Subject: Re: Help Bot vs Advance Level
|
On Jun 12, 6:49 am, help bot <[email protected] > wrote:
> > View Recorded Game:http://www.getclub.com/playgame.php?id=DM20686&game=Chess
> > 34. f2-f4{1776} e5-e4{132}
>
> Here, I overlooked: 34. ...Nxc3, 35. Rxe5+ Rxe5,
> 36. fxe5 Ne2+, winning. Apparently, so did the
> GetClub Advance level, since it thought 33. e5 was
> a good move... .
I actually meant that the GetClub program thought
34. f4 was a good move, which of course it wasn't
because of the potential Knight fork at e2 that I
missed when I chose 34. ... e4.
-- help bot
|
| | |
Date:
From: Martin Brown
Subject: Re: Help Bot vs Advance Level
|
|
| |
Date: 12 Jun 2008 08:14:46
From: Sanny
Subject: Re: Help Bot vs Advance Level
|
It was a great struggling game for you in the middle game. But in end
game GetClub could not do anything good despite having +2 Rooks for
your Rook+Knight pair.
Advance Level atleast get the priviledge of getting ahead in the
middle game. But End game it could not do anything with 1 point plus.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 12 Jun 2008 03:49:30
From: help bot
Subject: Re: Help Bot vs Advance Level
|
On Jun 11, 8:40 am, Sanny <[email protected] > wrote:
> View Recorded Game:http://www.getclub.com/playgame.php?id=DM20686&game=Chess
> --------------------------------------------------------------------------------
>
> White -- Black
> (master) -- (help bot)
>
> 1. d2-d4{1320} Ng8-f6{14}
> 2. Nb1-c3{4382} d7-d5{16}
> 3. Ng1-f3{2716} e7-e6{38}
> 4. Nf3-e5{2414} Nb8-d7{78}
> 5. a2-a3{1808} Nd7-e5{62}
If I had it to do over, I think I would avoid the complications
which follow and instead of taking on e5, just develop with
5. ...Bd6 instead.
> 6. d4-e5{1752} Nf6-d7{24}
> 7. e2-e4{2850} d5-e4{112}
> 8. Bc1-f4{1520} Bf8-e7{184}
> 9. Nc3-e4{2134} Ke8-g8{34}
> 10. Bf1-d3{1974} Qd8-e8{260}
This move (...Qe8) is directed at the possibility of
11. Qh5, with deadly threats against h7. Trouble is,
White can attack successfully in other ways.
> 11. Ke1-g1{2426} Nd7-b6{182}
After this move White should win. I did not see
the move 13. Nf6 coming, although I did feel very
uncomfortable about my King's safety.
> 12. Qd1-g4{1738} Kg8-h8{30}
> 13. Ne4-f6{1762} Be7-f6{500}
> 14. e5-f6{1448} g7-g6{44}
> 15. Bf4-h6{1974} Rf8-g8{448}
> 16. Ra1-d1{5230} Bc8-d7{584}
> 17. Rf1-e1{2790} Bd7-c6{422}
> 18. Bh6-g7{2068} Rg8-g7{44}
White should have tried for more than just
"the exchange for a pawn". Black is in no
position to defend his King, while White can
advance the h-pawn and/or do a Rook lift.
> 19. f6-g7{2460} Kh8-g7{10}
> 20. Qg4-f4{1436} Qe8-e7{120}
> 21. Bd3-e4{1350} Bc6-e4{70}
> 22. Qf4-e5{2276} Qe7-f6{124}
> 23. Qe5-f6{1314} Kg7-f6{8}
Strange as it may seem, I was relieved to
have the Queens off the board, despite the
bad ending.
> 24. Re1-e4{2328} Kf6-e7{50}
> 25. Rd1-d4{1304} h7-h5{308}
> 26. b2-b4{3836} a7-a5{1178}
> 27. b4-b5{1934} Ra8-c8{58}
> 28. c2-c3{1996} c7-c5{486}
The move p-c3 makes no sense to me. White
needs to open another file somewhere, and take
control of it; the Rooks are clumsy in closed
positions.
> 29. Rd4-d2{2286} c5-c4{154}
> 30. Rd2-d4{2012} Rc8-c5{130}
> 31. a3-a4{2130} f7-f5{94}
> 32. Re4-e3{3476} Nb6-a4{134}
> 33. h2-h4{1632} e6-e5{94}
> 34. f2-f4{1776} e5-e4{132}
Here, I overlooked: 34. ...Nxc3, 35. Rxe5+ Rxe5,
36. fxe5 Ne2+, winning. Apparently, so did the
GetClub Advance level, since it thought 33. e5 was
a good move... .
> 35. Re3-g3{1772} Rc5-b5{408}
> 36. Rd4-c4{3608} Na4-b2{200}
> 37. Rc4-d4{1310} Rb5-b6{120}
> 38. c3-c4{1548} Nb2-d3{46}
> 39. c4-c5{1914} Nd3-c5{42}
> 40. Rg3-e3{2512} Rb6-b1{172}
> 41. Kg1-h2{2136} b7-b5{60}
> 42. Rd4-d5{1344} Nc5-e6{72}
> 43. g2-g3{2708} Rb1-b2{162}
> 44. Kh2-h3{1586} Ne6-c7{154}
Moving to h3 was obviously wrong; there is
no potential stalemate trick, yet the King is
trapped out of play here.
> 45. Rd5-d3{2440} b5-b4{1066}
> 46. Rd3-d1{4880} a5-a4{60}
> 47. Re3-e1{6504} a4-a3{88}
> 48. Re1-f1{5490} a3-a2{62}
> 49. Rd1-c1{7012} Nc7-d5{70}
> 50. Rf1-e1{2008} Nd5-c3{40}
> 51. Rc1-c3{4706} b4-c3{36}
> 52. Re1-a1{330} c3-c2{48}
> 53. g3-g4{3172} h5-g4{44}
> 54. Kh3-g3{470} Rb2-b3{78}
> 55. Kg3-g2{512} Rb3-a3{42}
> 56. Kg2-f1{350} Rc2-c1{R}{28}
> 57. Ra1-c1{356} Qa2-a1{Q}{50}
> 58. Rc1-a1{358} Ra3-a1{6}
> 59. Kf1-f2{502} Ke7-e6{40}
> 60. Kf2-e2{808} g4-g3{32}
> 61. Ke2-d2{600} Ra1-a3{44}
> 62. Kd2-c2{1180} Ra3-d3{22}
> 63. h4-h5{612} g6-h5{14}
> 64. Kc2-b2{
> --------------------------------------------------------------------------------
> help bot: (Black)
> master: (White)
Hey, I sat there for six hundred hours (it felt like)
and now the game is being described as against
the lowly Master level? I think I already had it won
by the time Sanny arbitrarily intervened to change
levels near the finish. Look at the thinking time for
move 51: 4706 seconds to sac' it's Rook for my
Knight. After that, it made no difference, for even
a child could win it-- apart from the stalemate trap
that never appeared, but which I was looking out
for. : >D
-- help bot
|
| |
Date: 11 Jun 2008 05:40:43
From: Sanny
Subject: Help Bot vs Advance Level
|
GetClub Lost because of weak end game.
I think the End game was quite weak while middle game was well played
by GetClub.
Game Played between help bot and master at GetClub.com
--------------------------------------------------------------------------------
help bot: (Black)
master: (White)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM20686&game=Chess
--------------------------------------------------------------------------------
White -- Black
(master) -- (help bot)
1. d2-d4{1320} Ng8-f6{14}
2. Nb1-c3{4382} d7-d5{16}
3. Ng1-f3{2716} e7-e6{38}
4. Nf3-e5{2414} Nb8-d7{78}
5. a2-a3{1808} Nd7-e5{62}
6. d4-e5{1752} Nf6-d7{24}
7. e2-e4{2850} d5-e4{112}
8. Bc1-f4{1520} Bf8-e7{184}
9. Nc3-e4{2134} Ke8-g8{34}
10. Bf1-d3{1974} Qd8-e8{260}
11. Ke1-g1{2426} Nd7-b6{182}
12. Qd1-g4{1738} Kg8-h8{30}
13. Ne4-f6{1762} Be7-f6{500}
14. e5-f6{1448} g7-g6{44}
15. Bf4-h6{1974} Rf8-g8{448}
16. Ra1-d1{5230} Bc8-d7{584}
17. Rf1-e1{2790} Bd7-c6{422}
18. Bh6-g7{2068} Rg8-g7{44}
19. f6-g7{2460} Kh8-g7{10}
20. Qg4-f4{1436} Qe8-e7{120}
21. Bd3-e4{1350} Bc6-e4{70}
22. Qf4-e5{2276} Qe7-f6{124}
23. Qe5-f6{1314} Kg7-f6{8}
24. Re1-e4{2328} Kf6-e7{50}
25. Rd1-d4{1304} h7-h5{308}
26. b2-b4{3836} a7-a5{1178}
27. b4-b5{1934} Ra8-c8{58}
28. c2-c3{1996} c7-c5{486}
29. Rd4-d2{2286} c5-c4{154}
30. Rd2-d4{2012} Rc8-c5{130}
31. a3-a4{2130} f7-f5{94}
32. Re4-e3{3476} Nb6-a4{134}
33. h2-h4{1632} e6-e5{94}
34. f2-f4{1776} e5-e4{132}
35. Re3-g3{1772} Rc5-b5{408}
36. Rd4-c4{3608} Na4-b2{200}
37. Rc4-d4{1310} Rb5-b6{120}
38. c3-c4{1548} Nb2-d3{46}
39. c4-c5{1914} Nd3-c5{42}
40. Rg3-e3{2512} Rb6-b1{172}
41. Kg1-h2{2136} b7-b5{60}
42. Rd4-d5{1344} Nc5-e6{72}
43. g2-g3{2708} Rb1-b2{162}
44. Kh2-h3{1586} Ne6-c7{154}
45. Rd5-d3{2440} b5-b4{1066}
46. Rd3-d1{4880} a5-a4{60}
47. Re3-e1{6504} a4-a3{88}
48. Re1-f1{5490} a3-a2{62}
49. Rd1-c1{7012} Nc7-d5{70}
50. Rf1-e1{2008} Nd5-c3{40}
51. Rc1-c3{4706} b4-c3{36}
52. Re1-a1{330} c3-c2{48}
53. g3-g4{3172} h5-g4{44}
54. Kh3-g3{470} Rb2-b3{78}
55. Kg3-g2{512} Rb3-a3{42}
56. Kg2-f1{350} Rc2-c1{R}{28}
57. Ra1-c1{356} Qa2-a1{Q}{50}
58. Rc1-a1{358} Ra3-a1{6}
59. Kf1-f2{502} Ke7-e6{40}
60. Kf2-e2{808} g4-g3{32}
61. Ke2-d2{600} Ra1-a3{44}
62. Kd2-c2{1180} Ra3-d3{22}
63. h4-h5{612} g6-h5{14}
64. Kc2-b2{
--------------------------------------------------------------------------------
help bot: (Black)
master: (White)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM20686&game=Chess
What were the moves that were not correct?
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 11 Jun 2008 05:15:59
From: Sanny
Subject: Re: What is Rating of Rybka and other Programs?
|
On Jun 11, 4:26=A0pm, help bot <[email protected] > wrote:
> On Jun 9, 11:59 am, Sanny <[email protected]> wrote:
>
> > Do you want to Close this game without completing it? I will remove
> > this game?
>
> =A0 Why is it that Sanny only offers to eradicate in-
> progress games when I am clearly winning? =A0Can
> this be pure coincidence? =A0Every time?
>
> > Or have the computer think for 6 hours, in meantime go out and come
> > back after 6 hours when it has played the move.
>
> =A0 Until I read here that the Advance level can
> sometimes take as many as six hours, I was
> under the impression that there was some
> sort of malfunction. =A0TheGetClubWeb site
> does not say "20 minutes to six hours"; it
> says "20 to 40 minutes"; and yet it seems
> that Sanny knew about this all along.
>
> > I always say never play with Advance Level as it takes 30 min- 1 hiur
> > per move and for a few moves can go arround 6 hours of thinking. After
> > 6 hours of thinking If it do not find any good move it will just play
> > the best move it liked on the board. So maximum time any move can take
> > is 6 hours.
>
> =A0 Cool. =A0I still think it has to do with the
> possibility of a pawn promotion (or two),
> since it took less time earlier, in far more
> complex positions. =A0As I see it, the
> proper way to handle pawn promotions is
> to just add them to the list of legal moves,
> so you get p=3DQ, p=3DR, p=3DB, p=3DN, or four
> additional legal moves, which should slow
> things down a little, but not so much in
> comparison to a complex middle game.
>
> =A0 Apparently, some programs handle them
> in some bizarre way, which slows things
> down to a crawl-- even in Rook endings
> where most of the pawns are locked up
> and the number of legal moves is not high.
>
> > Normal Level plays as good as Rybka (5 sec/ move)
>
> =A0And yet I can beat the Normal level, like
> a carrot.
>
> > Master Level plays as good as Rybka (20 sec / move)
>
> =A0 And yet I can beat the Master level, like
> a beet.
>
> =A0 Let's face it: Rybka is the strongestchess
> engine in the world, and they need to give the
> human grandmasters material odds in order
> to make things interesting. =A0In one competition,
> Rybka made mincemeat of /all but one/ of its
> fellowchessprograms (Zap or Zappa), so it is
> not a big deal if theGetClubprogram does not
> play as well.
>
> =A0 Note well: the big money is made by programs
> which are mass-marketed; not by the programs
> which happen to be strong. =A0At one time, the
> strongestchessprogram in the world was
> essentially unknown, except to a relative few.
> Yet many knew of or owned such programs as
> "Chessmaster", because it was among the few
> which were mass-marketed.
>
> =A0 -- help bot
To speed up the game Level was rolled down to Master Level. Now it
will play that move in 1.5 hour only.
Now, even Zebediah do not play with Advance Level but Master & Normal
Levels.
I like that you were so ahead with this game. Looks like there is
still scope of improvement. After the game is complete we will analyze
what went wrong that Advance has to loose the game.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 11 Jun 2008 04:26:55
From: help bot
Subject: Re: What is Rating of Rybka and other Programs?
|
On Jun 9, 11:59 am, Sanny <[email protected] > wrote:
> Do you want to Close this game without completing it? I will remove
> this game?
Why is it that Sanny only offers to eradicate in-
progress games when I am clearly winning? Can
this be pure coincidence? Every time?
> Or have the computer think for 6 hours, in meantime go out and come
> back after 6 hours when it has played the move.
Until I read here that the Advance level can
sometimes take as many as six hours, I was
under the impression that there was some
sort of malfunction. The GetClub Web site
does not say "20 minutes to six hours"; it
says "20 to 40 minutes"; and yet it seems
that Sanny knew about this all along.
> I always say never play with Advance Level as it takes 30 min- 1 hiur
> per move and for a few moves can go arround 6 hours of thinking. After
> 6 hours of thinking If it do not find any good move it will just play
> the best move it liked on the board. So maximum time any move can take
> is 6 hours.
Cool. I still think it has to do with the
possibility of a pawn promotion (or two),
since it took less time earlier, in far more
complex positions. As I see it, the
proper way to handle pawn promotions is
to just add them to the list of legal moves,
so you get p=Q, p=R, p=B, p=N, or four
additional legal moves, which should slow
things down a little, but not so much in
comparison to a complex middle game.
Apparently, some programs handle them
in some bizarre way, which slows things
down to a crawl-- even in Rook endings
where most of the pawns are locked up
and the number of legal moves is not high.
> Normal Level plays as good as Rybka (5 sec/ move)
And yet I can beat the Normal level, like
a carrot.
> Master Level plays as good as Rybka (20 sec / move)
And yet I can beat the Master level, like
a beet.
Let's face it: Rybka is the strongest chess
engine in the world, and they need to give the
human grandmasters material odds in order
to make things interesting. In one competition,
Rybka made mincemeat of /all but one/ of its
fellow chess programs (Zap or Zappa), so it is
not a big deal if the GetClub program does not
play as well.
Note well: the big money is made by programs
which are mass-marketed; not by the programs
which happen to be strong. At one time, the
strongest chess program in the world was
essentially unknown, except to a relative few.
Yet many knew of or owned such programs as
"Chessmaster", because it was among the few
which were mass-marketed.
-- help bot
|
| |
Date: 11 Jun 2008 04:06:26
From: help bot
Subject: Re: What is Rating of Rybka and other Programs?
|
On Jun 9, 11:35 am, "Guest" <[email protected] > wrote:
> >> > As I recall, some "expert" explained that in
> >> > order to utilize multi-core processors properly,
> >> > the user needs to have Vista Professional or
> >> > Windows XP Pro-- fairly rare OSs in home
> >> > computers.
> >> Not true. (Well, you do need an OS that supports multi-core or
> >> multi-processor, yes. But they aren't rare.)
>
> > "...in home computers".
>
> Yes, home computers.
>
> As I said, as long as you didn't buy the low end, 'average consumer' grade
> cheap stuff from Dell or eMachines etc. XP Pro was not rare.
I wrote "...fairly rare".
> If you bought any sort of high end system (like for gaming
Professional is not the preferred choice for gaming.
> or like you'd do for serious chess
People like, say, Robert Hyatt would indeed go for
heavy-hitter hardware. That's not exactly a huge
market though. Wasn't Sanny asking about this
as a way to /make money/?
> or video editing
Now you're talking "Pro".
> etc.) the odds were good that XP Pro
> would either be installed or at least an option.
Moving the goalposts, again. I did not write that
such OSs were "not an option".
> Whether it was from Dell
> or some other company.
Changing the subject (to brand names). It really
makes no difference, as the main brands compete
head-to-head on price. I've owned just about all of
them at one time or another, except Dell and
Compaq.
> Higher end home systems had XP Pro available or it
> came with it by default.
Originally, the two most expensive choices were
XP Pro and I believe it was Media Center edition, so
these were "fairly rare" in comparison to the other
choices, at first, but once Vista was threatened, the
prices came down. I am talking about what you
might have seen in a major electronics retailer, such
as Best Buy, for instance. The big sales tended to
focus on the lower end, and these hot items tended
to have low-to-middle end OSs.
> > At work, I saw outdated versions of what Microsoft
>
> At work places, if you look hard enough, you can probably find Win95.
No need to "look"; computers are ubiquitous in
most businesses, these days. The key seems to
be *stability*, and salesmen appear to have pushed
specific "business" versions over what we already
know are unstable ones-- the ones "we" have been
stuck with for decades.
> Heck, I've heard about places still using DOS
[Gasp!]
> & Win3.11 for some stuff. The
> system works, the hardware is still working, so there's been no reason to
> change it.
[Software compatibility.]
> Wait until it breaks and then replace it.
Windows requires no waiting; it is broken, right
out of the box! (I wonder if that could be touted
as a selling point?)
> > liked to describe as "more stable" OSs (which tells
> > us very little since even the leaning Tower of Pizza
> > qualifies).
> >> First, XP Pro is actually quite common.
>
> > Hmm. I don't frequent Best Buy stores much
> > anymore, but back when I bought a couple of
> > computers there, the XP Pro was rather rare
>
> Places like BestBuy and WalMart tend to focus on the low end,
Time-warp. Back then, Wal-mart was not a major
retailer of computers; thinking back, I never bought
even one there. Now, things have changed. The
trouble is, they stick a few (if you're lucky) behind
some plexiglass, and even if they're in-stock you
can't see them powered-on. Even with the addition
of the Dell brand, the only reason they will sell very
many is that they are open 24 hours a day-- when
the competition is closed.
> slower
> systems. The kind of stuff the average consumer would buy so they could get
> onto that new fangled thing called the internet.
>
> Higher end systems (like for gamers or whoever) weren't as common there
> because they were more expensive and therefor didn't sell as well. But they
> did have them.
>
> If you are going to be buying a faster system (like what a dual core or dual
> processor system would have been years ago), then XP Pro would have been
> there.
Yet many of these systems were sold hyping
the "dual core" chip, without being properly set
up for whiz-bang chess programs. Relatively
few people will be eager to spring for an OS
upgrade, just to run a faster version of a chess
engine.
> If you were even buying a single core high end system, the odds were good XP
> Pro would have been available.
Key word: available; not standard.
> But not on the low end stuff. At that end, price was too important to add
> an extra us$50 for something a buyer wouldn't understand.
Out-of-touch with reality. Walk into Microsoft
headquarters and plop down a paltry $50; pound
the table and say you want a fancy OS-- you
may get blank stares, or they might just laugh.
> I didn't say XP Home wasn't more common. I said XP Pro wasn't rare.
And this was supposed to address my comment
"...fairly rare"? Sanny wants to make money, and
you want to pretend that multi-core systems and
the required OSs for running such programs are
ubiquitous. I think he should just forget the whole
idea of creating such a version-- even if he cannot
be stopped from his other boondoggles.
> > in comparison to cheaper OSs. Now, I expect
> > that Vista follows the same pattern: the top
> > version is the rarest, the cheapest one the
> > most common at first, then the middle version
> > catches up and finally surpasses sales of the
> > "basic" Vista.
>
> Vista is a little different.
How so?
It seems to me that the very same pattern is
repeating; they seem to think this model is not
broken.
> I think most systems come with Home Premium because Home Basic is too
> limited. Only the ultra-cheap systems bother with that. Usually where it
> comes with only 512m or they've skimped on the graphics hardware and Vista
> Areo can't run on it.
>
> And few systems come with Vista Ultimate because there is very little extra
> to be worth the substantial extra price.
Deja vu.
> So by trying to have three version, Microsoft kind of screwed themselves and
> made the Premium version suitable for nearly everybody. Few even want
> ultimate and few will encounter Basic (except on ultra low end laptops Like
> $400 or below kind of stuff.)
True. But the low-end systems are hyped in big
sales ads, and they probably sell very well due to
their low price.
> >> If you bought some low end Dell system, for example
...which I never have; hey, why is this always
about you straining to push a cheap Dell
computer off on me? You can keep all your
cheap junk. : >D
The computer I'm typing on right now does not
have the supposed "Premium" version of Vista.
It has Media Center edition, and that gradually
became more and more common. I remember
that at first, it was only on the higher-end
systems. Maybe my next computer will have
an OS which is designed for whiz-bang chess
engines... .
-- help bot
|
| |
Date: 10 Jun 2008 23:03:47
From: Sanny
Subject: Re: What is Rating of Rybka and other Programs?
|
> =A0 Seriously, I am through with waiting for well over
> an hour per move. =A0This is the last time... . =A0I mean
> it this time!
Game level reduced from Advance to Master Level, So now it will play
withing 1-2 hours Maximum.
Now, This game has been converted to Master Level. So never play with
Advance Level. Master level is a good player, Play with it.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 09 Jun 2008 17:45:33
From: SBD
Subject: Re: Rating of GetClub???
|
On Jun 9, 5:08 pm, "[email protected]" <[email protected] > wrote:
> > Beginner: 2000+
> > Easy: 2100+
> > Normal: 2200+
> > Master: 2300+
>
> Sanny, you can't just pull a bunch of ratings out of your butt and
> then claim they are correct.
Sure he can and does, regularly.
> No one believes you.
That's another issue entirely.
|
| |
> Beginner: 2000+
> Easy: 2100+
> Normal: 2200+
> Master: 2300+
Sanny, you can't just pull a bunch of ratings out of your butt and
then claim they are correct. You need some kind of organized
process. No one believes you. I see you left out Advanced, what is
that, 2400+ according to your butt?
|
| |
Date: 09 Jun 2008 08:59:27
From: Sanny
Subject: Re: What is Rating of Rybka and other Programs?
|
On Jun 9, 1:04=A0pm, help bot <[email protected] > wrote:
> Sanny wrote:
> > Yes, Advance level thinks for 1/2 an hour atleast. And most of the
> > moves takes less than 2 hours. But for some typical moves it can think
> > for arround 6 hours also. So try till 6 hours as Advance thinks a lot.
>
> > After 6 hours it will force Advance level to play the move which it
> > thinks best.
>
> > Thats why I always suggest not to play with Advance Level. Its better
> > to play with Master Level which makes moves in 10-30 min. And maximum
> > it will take 1.5 hours.
>
> =A0 I only hope that the next time you feel like trying
> the cheap ploy of suggesting that, although I beat
> level X, if only I were to play the *Advance level*, I
> would lose... the next time you feel one of those
> ploys coming on, try to /restrain/ yourself.
Do you want to Close this game without completing it? I will remove
this game?
Or have the computer think for 6 hours, in meantime go out and come
back after 6 hours when it has played the move.
I always say never play with Advance Level as it takes 30 min- 1 hiur
per move and for a few moves can go arround 6 hours of thinking. After
6 hours of thinking If it do not find any good move it will just play
the best move it liked on the board. So maximum time any move can take
is 6 hours.
> =A0 Seriously, I am through with waiting for well over
> an hour per move. =A0This is the last time... . =A0I mean
> it this time!
Yes, I too think you should now play with Master instead as now Master
will play as good as Advance Level due to improvement. Leave Advance
Level for Zebediah who can still beat the Advance Level thinking upto
6 hours / move
I suggest you should play Master & Normal Level only as they are Quick
& Quite strong.
Normal Level plays as good as Rybka (5 sec/ move)
Master Level plays as good as Rybka (20 sec / move)
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 09 Jun 2008 01:04:18
From: help bot
Subject: Re: What is Rating of Rybka and other Programs?
|
Sanny wrote:
> Yes, Advance level thinks for 1/2 an hour atleast. And most of the
> moves takes less than 2 hours. But for some typical moves it can think
> for arround 6 hours also. So try till 6 hours as Advance thinks a lot.
>
> After 6 hours it will force Advance level to play the move which it
> thinks best.
>
> Thats why I always suggest not to play with Advance Level. Its better
> to play with Master Level which makes moves in 10-30 min. And maximum
> it will take 1.5 hours.
I only hope that the next time you feel like trying
the cheap ploy of suggesting that, although I beat
level X, if only I were to play the *Advance level*, I
would lose... the next time you feel one of those
ploys coming on, try to /restrain/ yourself.
Seriously, I am through with waiting for well over
an hour per move. This is the last time... . I mean
it this time!
-- help bot
|
| |
Date: 08 Jun 2008 22:40:26
From: Sanny
Subject: Re: What is Rating of Rybka and other Programs?
|
> > Waiting for this game to end and see how such interesting moves were
> > played.
>
> =A0 Me too. =A0The program thought for nearly an
> hour and a half before I decided to shut it
> down and try again. =A0Now it's back up well
> over an hour, and still cranking away. =A0That's
> a lot of thinking for just one move. =A0I think it
Yes, Advance level thinks for 1/2 an hour atleast. And most of the
moves takes less than 2 hours. But for some typical moves it can think
for arround 6 hours also. So try till 6 hours as Advance thinks a lot.
After 6 hours it will force Advance level to play the move which it
thinks best.
Thats why I always suggest not to play with Advance Level. Its better
to play with Master Level which makes moves in 10-30 min. And maximum
it will take 1.5 hours.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 08 Jun 2008 21:42:19
From: help bot
Subject: Re: What is Rating of Rybka and other Programs?
|
On Jun 8, 10:45 am, "Guest" <[email protected] > wrote:
> > As I recall, some "expert" explained that in
> > order to utilize multi-core processors properly,
> > the user needs to have Vista Professional or
> > Windows XP Pro-- fairly rare OSs in home
> > computers.
>
> Not true. (Well, you do need an OS that supports multi-core or
> multi-processor, yes. But they aren't rare.)
"...in home computers".
At work, I saw outdated versions of what Microsoft
liked to describe as "more stable" OSs (which tells
us very little since even the leaning Tower of Pizza
qualifies).
> First, XP Pro is actually quite common.
Hmm. I don't frequent Best Buy stores much
anymore, but back when I bought a couple of
computers there, the XP Pro was rather rare
in comparison to cheaper OSs. Now, I expect
that Vista follows the same pattern: the top
version is the rarest, the cheapest one the
most common at first, then the middle version
catches up and finally surpasses sales of the
"basic" Vista.
> If you bought some low end Dell system, for example,
Egads man, what do you take me for? ; >D
> you would have gotten XP Home, but any higher end
> system from back then would have had XP Pro available, if not standard.
Ah, so then you are discussing "high end"
computers, while I am/was discussing home
computers (see above).
> Second, if you were buying a dual core system (then or now), a suitable OS
> would have been installed on it so you could take advatange of the dual
> cores and other advanced features. (With a few exceptions where some
> laptops or low end systems were shipped with XP Home or configured so the
> second core wasn't used.)
In reality, many systems were touted as
containing the faster chips, but in efforts to
keep the retail price competitive, they were
not equipped with "professional strength"
OSs.
> As for Vista... There's actually some confusion about whether Vista Home
> Basic can handle multiple cores or not. Microsoft's own site says "While
> all editions of Windows Vista can support multiple core CPUs, only Windows
> Vista Business, Ultimate, and Enterprise can support dual processors"
>
> Which gives the impression that even Vista Basic can handle multiple cores
> provided it's not just two...
This is what I was talking about before; the
fact that a computer "works" with a cheap OS
and a fancy chip in no way guarantees that it
will properly "support" multi-core chess engines
such that they run at lightning speed.
Back when I bought the computer I am typing
on now, Vista Ultimate was relatively rare... in
home computers. That's probably because they
have to compete on price, and it's easy to tout
a dual-core processor (while downplaying the
fact that in order to run supeRybkaX, you would
have to upgrade the OS).
These are all moot points if you are a nutter
who selects his computer specifically to run
GrandeFritz12.0, 16 processor version.
> I've never checked into whether it can or not because I don't care.
>
> The point is, if you have a multi-core or multi-processor system, you'll
> have an OS that can take advantage of it. In the XP days, not uncommon.
> These days, very common.
Ah, but you seem to think you know more
than Microsoft... which is impossible. Take
co-founder Bill Gates, for instance; even as
far back as 1962 he forecast the coming
boom in personal computers:
"I see no advantage to a "graphical" operating
system".
Oops-- wrong quote. Um, well.. he was the
richest crook in America, which should count
for something.
How about some instructions on how to
"enable" fast processors under manufacturer
equipped OSs? I have an AMD Turion64,
which was touted as being very fast (for a
*home* computer) at the time, but it came
with Windows XP Media Center Edition. I
once installed a free chess engine which was
designed for 64-bit OSs, and it didn't work.
So I installed the 32-bit version, and it did
work. Anyway, I literally melted one of the
pivots which hold the screen on, so this one
is done for... I won't be putting money into
upgrading it.
-- help bot
|
| | |
Date: 09 Jun 2008 10:35:48
From: Guest
Subject: Re: What is Rating of Rybka and other Programs?
|
"help bot" <[email protected] > wrote in message
news:98798e2c-4489-4c58-847c-ef0f29ad0577@z72g2000hsb.googlegroups.com...
> On Jun 8, 10:45 am, "Guest" <[email protected]> wrote:
>
>> > As I recall, some "expert" explained that in
>> > order to utilize multi-core processors properly,
>> > the user needs to have Vista Professional or
>> > Windows XP Pro-- fairly rare OSs in home
>> > computers.
>>
>> Not true. (Well, you do need an OS that supports multi-core or
>> multi-processor, yes. But they aren't rare.)
>
> "...in home computers".
Yes, home computers.
As I said, as long as you didn't buy the low end, 'average consumer' grade
cheap stuff from Dell or eMachines etc. XP Pro was not rare.
If you bought any sort of high end system (like for gaming, or like you'd do
for serious chess, or video editing, etc.) the odds were good that XP Pro
would either be installed or at least an option. Whether it was from Dell
or some other company. Higher end home systems had XP Pro available or it
came with it by default.
> At work, I saw outdated versions of what Microsoft
At work places, if you look hard enough, you can probably find Win95.
(shrug)
Heck, I've heard about places still using DOS & Win3.11 for some stuff. The
system works, the hardware is still working, so there's been no reason to
change it. Wait until it breaks and then replace it.
> liked to describe as "more stable" OSs (which tells
> us very little since even the leaning Tower of Pizza
> qualifies).
>
>
>> First, XP Pro is actually quite common.
>
> Hmm. I don't frequent Best Buy stores much
> anymore, but back when I bought a couple of
> computers there, the XP Pro was rather rare
Places like BestBuy and WalMart tend to focus on the low end, slower
systems. The kind of stuff the average consumer would buy so they could get
onto that new fangled thing called the internet.
Higher end systems (like for gamers or whoever) weren't as common there
because they were more expensive and therefor didn't sell as well. But they
did have them.
If you are going to be buying a faster system (like what a dual core or dual
processor system would have been years ago), then XP Pro would have been
there.
If you were even buying a single core high end system, the odds were good XP
Pro would have been available.
But not on the low end stuff. At that end, price was too important to add
an extra us$50 for something a buyer wouldn't understand.
I didn't say XP Home wasn't more common. I said XP Pro wasn't rare.
> in comparison to cheaper OSs. Now, I expect
> that Vista follows the same pattern: the top
> version is the rarest, the cheapest one the
> most common at first, then the middle version
> catches up and finally surpasses sales of the
> "basic" Vista.
Vista is a little different.
I think most systems come with Home Premium because Home Basic is too
limited. Only the ultra-cheap systems bother with that. Usually where it
comes with only 512m or they've skimped on the graphics hardware and Vista
Areo can't run on it.
And few systems come with Vista Ultimate because there is very little extra
to be worth the substantial extra price.
So by trying to have three version, Microsoft kind of screwed themselves and
made the Premium version suitable for nearly everybody. Few even want
ultimate and few will encounter Basic (except on ultra low end laptops Like
$400 or below kind of stuff.)
>> If you bought some low end Dell system, for example,
>
> Egads man, what do you take me for? ;>D
>
>
>> you would have gotten XP Home, but any higher end
>> system from back then would have had XP Pro available, if not standard.
>
> Ah, so then you are discussing "high end"
> computers, while I am/was discussing home
> computers (see above).
I am talking *home* computers.
The kind of systems you'd find in a persons home. Not an office or a home
office or such.
But actually in a person's den or a kid's bedroom.
That last is particularly notable, because kids will tend to have the higher
end systems so they can play the newer games.
Even many dens would have them so they could do video editing of their home
movies, or because the sales person said it was 'better' than this other
cheaper system. (And so the father could do the occasional game when nobody
was looking.)
When I said "home", I meant home. For actual personal daily use.
Not in the homes of people who prefer to buy the cheapest one they can, no.
But definetly in a lot of homes.
It wasn't rare. It wasn't even 10%, but that still isn't rare.
>
>
>> Second, if you were buying a dual core system (then or now), a suitable
>> OS
>> would have been installed on it so you could take advatange of the dual
>> cores and other advanced features. (With a few exceptions where some
>> laptops or low end systems were shipped with XP Home or configured so the
>> second core wasn't used.)
>
> In reality, many systems were touted as
> containing the faster chips, but in efforts to
> keep the retail price competitive, they were
> not equipped with "professional strength"
> OSs.
I can't vouch for 'many'.
I do know that in the early days of dual-core systems, some low end systems
were done like that. Especially with laptops.
None of the mid or higher end systems were, though.
And none of the Vista systems, since even Basic can handle multi-core.
>> As for Vista... There's actually some confusion about whether Vista Home
>> Basic can handle multiple cores or not. Microsoft's own site says "While
>> all editions of Windows Vista can support multiple core CPUs, only
>> Windows
>> Vista Business, Ultimate, and Enterprise can support dual processors"
>>
>> Which gives the impression that even Vista Basic can handle multiple
>> cores
>> provided it's not just two...
>
> This is what I was talking about before; the
> fact that a computer "works" with a cheap OS
> and a fancy chip in no way guarantees that it
> will properly "support" multi-core chess engines
> such that they run at lightning speed.
I *think* what Microsoft is meaning is that Vista Basic *will* work with
systems that have multiple cores within one physical chip, but not systems
that have two physically seperate chips.
Microsoft is making an arbitrary distinction between multi-core &
multi-processors even though to the programmer they look the same.
But I haven't checked into it because I don't care.
I do know that nearly all systems sold today have Vista Home Premium, which
can definetly handle multi-core systems that are common today. Both dual
core and quad core.
So you can be pretty sure that every dual core or quad core system sold
today comes with an OS that allows those cores to be used.
>
> Back when I bought the computer I am typing
> on now, Vista Ultimate was relatively rare... in
> home computers. That's probably because they
> have to compete on price, and it's easy to tout
Partially price and partially because Vista Ultimate offers so little extra
that even most hard core enthusiasts don't want it.
Things are either a pain to use because the are poorly designed or other
products do a better job. (For example, with bitlocker you can't use a
password. You have to use the function keys to enter a 48 digit number.)
So except for exotic systems (where you might have multiple processors,
etc.), even most high end system owners don't want it.
> a dual-core processor (while downplaying the
> fact that in order to run supeRybkaX, you would
> have to upgrade the OS).
>
> These are all moot points if you are a nutter
> who selects his computer specifically to run
> GrandeFritz12.0, 16 processor version.
>
>
>> I've never checked into whether it can or not because I don't care.
>>
>> The point is, if you have a multi-core or multi-processor system, you'll
>> have an OS that can take advantage of it. In the XP days, not uncommon.
>> These days, very common.
>
> Ah, but you seem to think you know more
> than Microsoft... which is impossible. Take
> co-founder Bill Gates, for instance; even as
> far back as 1962 he forecast the coming
> boom in personal computers:
>
> "I see no advantage to a "graphical" operating
> system".
>
> Oops-- wrong quote. Um, well.. he was the
> richest crook in America, which should count
> for something.
>
> How about some instructions on how to
> "enable" fast processors under manufacturer
> equipped OSs? I have an AMD Turion64,
> which was touted as being very fast (for a
> *home* computer) at the time, but it came
It probably was. Even in single core usage, it was / is as fast or faster
than many of the Pentium-4 systems that were considered high end just a year
before.
I know a guy who went from a 2.4ghz P4 to a 1.9ghz AMD X2, and even in
single core (ie: running a single program that couldn't take advantage of
the extra core), it was about 40% faster. So it was giving performace
comparable to a 3+ Ghz P4. And it was doing it with less power, less
cooling (and hence less noise), and still had one core left over for you to
do something else at the same time.
> with Windows XP Media Center Edition. I
> once installed a free chess engine which was
> designed for 64-bit OSs, and it didn't work.
A couple of questions...
1) Why would you have bought an OS designed for a "media center" and expect
it to work well for general purpose stuff?
2) Why would you try to run a 64 bit program on a 32 bit system?
Don't blame the OS on your poor choice of OS or on trying to run a 64 bit
program on it.
The switch from 32 to 64 bits is going slowly, unfortunately. Microsoft
could have speeded it up quite a bit but hasn't wanted to. They haven't
even wanted to do an 'extender' kind of thing, like they did with Win3,
where you could run some 32 bit programs under Win32s.
> So I installed the 32-bit version, and it did
> work. Anyway, I literally melted one of the
> pivots which hold the screen on, so this one
> is done for... I won't be putting money into
> upgrading it.
Sounds like the system was designed with insufficient cooling.
That's not the fault of the OS or the chess program.
The same thing could have happened if you had done non-stop video conversion
or number crunching or anything that would have generated more heat than
'typical' *media center* usage.
That's just simply lack of decent cooling.
(And since you are talking about a Turion, I assume this was a laptop style
system. In which case they are well known to run right at the edge of
cooling limits. Some systems are designed better than others, of course,
but it's just common sense that if the exhaust becomes super hot and stays
that way, then you should probably back off for a while and let it cool
down.)
----== 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: 08 Jun 2008 06:55:21
From: help bot
Subject: Re: What is Rating of Rybka and other Programs?
|
On Jun 8, 6:02 am, Sanny <[email protected] > wrote:
> Waiting for this game to end and see how such interesting moves were
> played.
Me too. The program thought for nearly an
hour and a half before I decided to shut it
down and try again. Now it's back up well
over an hour, and still cranking away. That's
a lot of thinking for just one move. I think it
may be that the possibility of an eventual
pawn promotion is dramatically increasing
the thinking times. That was a frequent
problem with old chess programs from a
decade or so ago, as I recall.
-- help bot
|
| |
Date: 08 Jun 2008 03:02:59
From: Sanny
Subject: Re: What is Rating of Rybka and other Programs?
|
> > You are a pawn down So what about GetClubs passed pawns. Are they
> > blocked by your pieces?
>
> =A0 I never allowed the program to obtain a
> passed pawn; those things can be quite
> dangerous. =A0By comparison, I now have:
>
> one protected, passed pawn;
>
> and two connected passed pawns.
Waiting for this game to end and see how such interesting moves were
played.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 08 Jun 2008 01:26:53
From: help bot
Subject: Re: What is Rating of Rybka and other Programs?
|
On Jun 8, 3:21 am, Sanny <[email protected] > wrote:
> > Right now, I am playing the Advance level at
> > GetClub, and I was losing badly in the early
> > middle game. There were simply too many
> > ways to get at my King and I failed to cover
> > them all, and I ended up losing the exchange
> > for one pawn. But the endgame is another
>
> So GetClub is one pawn ahead. May be Advance get a draw by exchanging
> all pieces.
Note: in order to attain such a draw via simplification,
it is the pawns which must (all) be exchanged, not the
pieces.
For instance, with no pawns whatever, all these
positions are very often drawn, barring use of the
endgame table-bases and provided the inferior
side is not also the inferior chess player:
KR vs. KN
KRR vs. KRN
KBB vs. KR
KNN vs. KR
KBN vs. KR
By contrast, it requires just one pawn remaining
on the board for the possibility of a win to emerge,
except:
in opposite-colored Bishop endings;
in Rook endings where the inferior side's
King can get near the pawn in time;
where the other side can sacrifice a piece for
the passed pawn, and yet still secure a draw.
> Game was improved 4 days back
Me too. I watched a video of two women
playing blitz, and although I couldn't see what
happened on the board, I learned yet another
useful technique for distracting the opponent!
Just wear long, dangling sleeves and hover
them over the board on the opponent's time.
; >D
> So you will really face hard in winning
> the games. Next time play with Master Level which plays 4 times faster
> than Advance Level.
I think I am through with playing the Advance
level, as it is too slow in the middle game. I
notice that if I switch back to the chess window,
it seems to improve the chances of the program
finally making a move, if only a little.
> But as the game was improved you will see good
> challenge even with the Master Level.
No doubt. I have noticed a lot of improvement in
strength of play, but if I take the White pieces and
hunker down, the program tends to self-destruct,
so it is far from invincible. I still feel that I am
vastly superior in the *endgame*.
> How many pieces are left in the end game?
Lots of pawns remain, and the program has
two Rooks against my Rook and Knight. I am
now ahead by three pawns, but down the
exchange; if I Queen a pawn, it's all over.
> You are a pawn down So what about GetClubs passed pawns. Are they
> blocked by your pieces?
I never allowed the program to obtain a
passed pawn; those things can be quite
dangerous. By comparison, I now have:
one protected, passed pawn;
and two connected passed pawns.
-- help bot
|
| |
Date: 08 Jun 2008 00:21:07
From: Sanny
Subject: Re: What is Rating of Rybka and other Programs?
|
> =A0 Right now, I am playing the Advance level at
> GetClub, and I was losing badly in the early
> middle game. =A0There were simply too many
> ways to get at my King and I failed to cover
> them all, and I ended up losing the exchange
> for one pawn. =A0But the endgame is another
So GetClub is one pawn ahead. May be Advance get a draw by exchanging
all pieces.
Game was improved 4 days back So you will really face hard in winning
the games. Next time play with Master Level which plays 4 times faster
than Advance Level. But as the game was improved you will see good
challenge even with the Master Level.
Play Chess at: http://www.GetClub.com/Chess.html
Earlier Zebediah used to win in 30-35 moves against Master Level. Now
he takes 40-50 moves in winning Master level. He is having lot
difficulty in winning the Advance Level. He win Advance Level in 50-60
Moves earlier he used to win in just 30-40 Moves. So in all games
GetClub takes +20 moves before loosing to Zebediah.
Still Zebediah with his brute force analysis keeps ahead of GetClub
Program.
I am looking for further improvement in program but now there is
nothing that can be improved further.
> matter; here, even the Advance level does not
> seem able to make any headway, and I am
> plodding along, despite my material deficit.
> As usual, the program's King plays no role
> whatever except as potential target, and this
> in effect places me a piece ahead-- albeit a
How many pieces are left in the end game?
> very vulnerable one. =A0The game's result is still
> in question, but I now have hopes of winning
> as I have obtained some passed pawns, and
> this, in the past, has resulted in great tragedy
> for the GC program, because it can't defend
> well.
You are a pawn down So what about GetClubs passed pawns. Are they
blocked by your pieces?
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| |
Date: 07 Jun 2008 23:53:07
From: help bot
Subject: Re: What is Rating of Rybka and other Programs?
|
On Jun 8, 1:46 am, Sanny <[email protected] > wrote:
> Does anyone have Rybka Program and tell me how it face the Normal
> Level So that I may compute the correct Rating of GetClub on same
> hardware.
I tested this in my imagination, and here are the
(imaginary) results:
Rybka 1.0 free beta version: 999,999,999,999.5
GetClub Normal level: 000,000,000,000.5
This would seem to place the Rybka program at
around 3000+, and the GetClub Normal level at
somewhere below 2200, in imaginary numbers.
: >D
> 1. Do not use dual / Quad core as GetClub do not take advantage of
> multicores.
As I recall, some "expert" explained that in
order to utilize multi-core processors properly,
the user needs to have Vista Professional or
Windows XP Pro-- fairly rare OSs in home
computers.
> 2. Do not allow Rybka to think in opponents time as it will slow down
> GetClub.
More to the point, it would negate the time
restriction on Rybka to a large degree. You
see, allowing a dullard to "think" longer has
little effect, but allowing a genius to do so is
quite a different matter.
> Once that is done I will get correct Rating of GetClub Levels.
>
> Normal Level plays in 80-120 sec / move.
The tournaments from which computer ratings
are derived tend to stretch the differences a bit;
this is because stronger engines tend to
whammy weaker ones nearly 100% of the time,
unlike in human vs. human competitions. As I
recall, some computers which were at one time
"certified" by the USCF to be of, say, Expert
strength, now appear -- if at all -- as lowly class
players, having been battered down by their
many successors. Does this make them any
weaker than before? Certainly not! They had
*fixed* programs, on *fixed* hardware, so their
level of skill is obviously fixed, and the entire
pool of human players would have to improve
markedly for the relative ratings to change as it
has on these computer-only lists.
Instead of testing the GetClub program against
Rybka, where it is clearly outclassed, why not
test it against some program with a known
strength of around 2000 or so? You want to
have fairly equal wins and losses, so you can
calculate a rating estimate *reliably*. You can't
get a reliable, meaningful number from a small
number of games against an opponent which is
several classes stronger; especially when the
GC program takes a widely-variable amount of
time, while the other program's time per move
will be strictly limited, fixed.
Right now, I am playing the Advance level at
GetClub, and I was losing badly in the early
middle game. There were simply too many
ways to get at my King and I failed to cover
them all, and I ended up losing the exchange
for one pawn. But the endgame is another
matter; here, even the Advance level does not
seem able to make any headway, and I am
plodding along, despite my material deficit.
As usual, the program's King plays no role
whatever except as potential target, and this
in effect places me a piece ahead-- albeit a
very vulnerable one. The game's result is still
in question, but I now have hopes of winning
as I have obtained some passed pawns, and
this, in the past, has resulted in great tragedy
for the GC program, because it can't defend
well.
-- help bot
|
| | |
Date: 14 Jun 2008 17:29:19
From: Bjoern
Subject: Re: What is Rating of Rybka and other Programs?
|
help bot wrote:
> Instead of testing the GetClub program against
> Rybka, where it is clearly outclassed, why not
> test it against some program with a known
> strength of around 2000 or so? You want to
> have fairly equal wins and losses, so you can
> calculate a rating estimate *reliably*.
I really wonder why this great suggestion was ignored. And to get lots
of results quickly, you should just take up the suggestion below by Guest.
Guest wrote:
> I mean really! If you want a reliable rating for your program, then the
> best way is to let other people and programs play against it.
>
> If you aren't willing to do that, then you probably already know your
> program is under 1800, which even 8 bit micro's surpassed decades ago.
Indeed. Why not make the engine your engine play on FICS for free?
http://www.freechess.org/
(see http://www.freechess.org/Help/HelpFiles/computers.html and
http://www.freechess.org/Help/HelpFiles/computer_app.html for details)
If you can't be bothered to do that, why waste anyone's time with
guesses about your program's strength?
Addtionally it would be easy to set up your computer so that it can play
automatically without your supervision, so you can let it go online,
play a few hundred games and you'll have a suggestion for how it's rated
on FICS.
Looking at the win I just easily scored against the master level while
watching the EURO, shows that you really need to work on the evaluation
function, the program just fails to see tactical threats that are coming
up on it and then committs tactical suicide with his king:
1.d4 Nf6 2.Bg5 Ne4 3.Bf4 e6 4.f3 Bb4+ 5.c3 Bd6 6.Bxd6 Nxd6 7.e4 Nc6
8.Nd2 0-0 9.Bd3 Qg5 10.Qe2 f6 11.0-0-0 b6 12.h4 Qa5 13.Kb1 Bb7 14.g4 Qa4
15.g5 Rf7 16.Nh3 e5 17.d5 Ne7 18.Rdg1 Ng6 19.h5 Ne7 20.g6 hxg6 21.hxg6
Rff8 22.Qh2 Nef5 23.Bc2 Qa5 24.exf5 Nf7 25.gxf7+ Kxf7 26.Rxg7+ Kxg7
27.Nf4 Kf7 28.Qh7+ Ke8 29.Ng6 Qxa2+ 30.Kc1 Bxd5 1-0
The master level was a real pushover here. Black's (the computer)
position was fine out of the opening (up to move six), but almost every
single move from 7 to 15 is mediocre, pointless or just plain poor.
18...Ng6 by the computer is then just tactical suicide, how can any
computer not see that if it has a vaguely reasonable evaluation
function? I guess the answer is that GetClub doesn't evaluate positions
well, at all. Time and again the program just seems to fail to see a
need to play actively itself, it just plays its pieces about without
impproving its position or actually threatening anything.
I would bet that most reasonably strong human players would immediately
know that white is completely winning after 20.g6 with the idea of e.g.
as in the game 22.Qh2 (if black opens the h-file) or gxh7+ and attacking
down the g-file.
|
| | | |
Date: 14 Jun 2008 18:34:55
From: David Richerby
Subject: Re: What is Rating of Rybka and other Programs?
|
Bjoern <[email protected] > wrote:
> Indeed. Why not make the engine your engine play on FICS for free?
GetClub would lose so many games on time that the resulting rating
would tell you nothing at all about its strength.
Dave.
--
David Richerby Incredible Hungry Umbrella (TM):
www.chiark.greenend.org.uk/~davidr/ it's like an umbrella but it'll eat
you and blow your mind!
|
| | |
Date: 08 Jun 2008 09:45:39
From: Guest
Subject: Re: What is Rating of Rybka and other Programs?
|
"help bot" <[email protected] > wrote in message
news:c2ce52c5-f7ff-4b4c-a00f-21b0016ead7e@m36g2000hse.googlegroups.com...
> On Jun 8, 1:46 am, Sanny <[email protected]> wrote:
>
>> Does anyone have Rybka Program and tell me how it face the Normal
>> Level So that I may compute the correct Rating of GetClub on same
>> hardware.
>> 1. Do not use dual / Quad core as GetClub do not take advantage of
>> multicores.
>
> As I recall, some "expert" explained that in
> order to utilize multi-core processors properly,
> the user needs to have Vista Professional or
> Windows XP Pro-- fairly rare OSs in home
> computers.
Not true. (Well, you do need an OS that supports multi-core or
multi-processor, yes. But they aren't rare.)
First, XP Pro is actually quite common. If you bought some low end Dell
system, for example, you would have gotten XP Home, but any higher end
system from back then would have had XP Pro available, if not standard.
Second, if you were buying a dual core system (then or now), a suitable OS
would have been installed on it so you could take advatange of the dual
cores and other advanced features. (With a few exceptions where some
laptops or low end systems were shipped with XP Home or configured so the
second core wasn't used.)
As for Vista... There's actually some confusion about whether Vista Home
Basic can handle multiple cores or not. Microsoft's own site says "While
all editions of Windows Vista can support multiple core CPUs, only Windows
Vista Business, Ultimate, and Enterprise can support dual processors"
Which gives the impression that even Vista Basic can handle multiple cores
provided it's not just two...
I've never checked into whether it can or not because I don't care.
The point is, if you have a multi-core or multi-processor system, you'll
have an OS that can take advantage of it. In the XP days, not uncommon.
These days, very common.
----== 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: 07 Jun 2008 22:46:25
From: Sanny
Subject: Re: What is Rating of Rybka and other Programs?
|
> > Strength? What is the time per move at which Rybka Rating equals 3100?
>
> You'd have to check into who was saying Rybka was 3100. =A0It could be any=
> speed setting they were wanting to talk about.
Yes, I found that Rybka was playing Equivalent to 40 moves in 40
minutes on a Athlon 64 X2 4600+ (2.4 GHz)
http://computerchess.org.uk/ccrl/4040/index.html
It gave Rybka rating of 3129.
Does anyone have Rybka Program and tell me how it face the Normal
Level So that I may compute the correct Rating of GetClub on same
hardware.
1. Do not use dual / Quad core as GetClub do not take advantage of
multicores.
2. Do not allow Rybka to think in opponents time as it will slow down
GetClub.
Once that is done I will get correct Rating of GetClub Levels.
Normal Level plays in 80-120 sec / move.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| | |
Date: 08 Jun 2008 09:43:46
From: Guest
Subject: Re: What is Rating of Rybka and other Programs?
|
"Sanny" <[email protected] > wrote in message
news:4e4aa021-8f0c-46b2-ba03-8372f769ed98@j33g2000pri.googlegroups.com...
> > Strength? What is the time per move at which Rybka Rating equals 3100?
>
>> You'd have to check into who was saying Rybka was 3100. It could be any
>> speed setting they were wanting to talk about.
>Yes, I found that Rybka was playing Equivalent to 40 moves in 40
>minutes on a Athlon 64 X2 4600+ (2.4 GHz)
>
>http://computerchess.org.uk/ccrl/4040/index.html
>
>It gave Rybka rating of 3129.
That's definetly computer vs. computer, so you need to take that rating with
a proverbial grain of salt.
In other words, it may be valid for that particular tournament, but it's not
representative of its rating against humans.
>Does anyone have Rybka Program and tell me how it face the Normal
>Level So that I may compute the correct Rating of GetClub on same
>hardware.
Wont work.
To get a rating you have to play a lot more than just one program. Or even
few programs.
And you'll have to play lots of games to get a reasonable accuracy for your
program.
That's why I keep telling you that the fastest way for you to get a real
rating for your program is to modify your program so it can play on the
chess servers.
Either that or you will have to get a dozen or so programs or (better yet)
dedicated chess computers with fixed hardware and program versions that have
real ratings.
And then play hundreds of games.
As I said... it's easier if you do your program on the chess servers. And
it'll be more accurate and realistic.
----== 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: 07 Jun 2008 22:30:54
From: Sanny
Subject: Rating of GetClub???
|
> Crafty, ChessMaster, Fritz who ranks first and what is rating of each?
>
> What is the Rank of GetClub Chess in all engines?
>
> Bye
> Sanny
A few months back a few people used to testr GetClub against other
Chess Engines and used to say see how GetClub lost the game. But after
the GetClub Chess was improved they have just disappeared from the
group.
Any one having Fritz/ Chess Master/ Crafty etc can you play a few
games and let me know what is the rating of Beginner & Easy Levels at
GetClub Chess?
As per my estimates on my old computer
Beginner: 2000+
Easy: 2100+
Normal: 2200+
Master: 2300+
Do you all agree with above rating for GetClub Levels after the
improvements?
How many players here are 2300+ rated and can win the Master Level?
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| | |
Date:
From: Martin Brown
Subject: Re: Rating of GetClub???
|
|
| |
Date: 07 Jun 2008 14:43:24
From: Guest
Subject: Re: What is Rating of Rybka and other Programs?
|
"Sanny" <[email protected] > wrote in message
news:[email protected]...
>I read on internet Rybka is 3100 rated. And Fritz is 2900 Rated is it
> true?
Possibly if you limit the results to solely computer vs. computer.
When you test only computer vs. computer, you can get some really screwed up
ratings that don't match well to what you'd get against humans.
> When they say Rybka is 3100 Rated How much time they give it for each
> Move.
I suppose that would depend on where you read people were saying it was at
that rating.
I wouldn't be surprised if that was some sort of blitz game, though.
> As I played Rybka (5 sec / move) against Normal Level (100 sec / move)
> and found both play equally well.
>
> So when Rybka is playing at 5 sec / move is it playing with 3100
No.
> Strength? What is the time per move at which Rybka Rating equals 3100?
You'd have to check into who was saying Rybka was 3100. It could be any
speed setting they were wanting to talk about.
> What are the Programs which Rank 2ns and 3rd and what is their
> ratings?
>
> Crafty, ChessMaster, Fritz who ranks first and what is rating of each?
Current versions of Crafty are probably 200-250 points lower than Rybka.
And there are quite a few programs in between them.
Based on what I've seen on the web, Fritz might be less than a hundred
points lower than Rybka.
Not sure about ChessMaster.
> What is the Rank of GetClub Chess in all engines?
Probably pretty darn low, since you seem determined not to let it play on
the free internet chess servers where it can play against people and
programs and actually get a reliable rating.
I mean really! If you want a reliable rating for your program, then the
best way is to let other people and programs play against it.
If you aren't willing to do that, then you probably already know your
program is under 1800, which even 8 bit micro's surpassed decades ago.
>
> Bye
> Sanny
>
> Play Chess at: http://www.GetClub.com/Chess.html
----== 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: 07 Jun 2008 09:36:31
From: Sanny
Subject: Re: What is Rating of Rybka and other Programs?
|
> I read on internet Rybka is 3100 rated. And Fritz is 2900 Rated is it
> true?
>
> When they say Rybka is 3100 Rated How much time they give it for each
> Move.
And also Which Computer they Play to test the Rating?
Is it Celeron? Dual Core? Quad Core? or a Super Computer.
Since rating can be increased a lot by giving a good hardware.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
|
| | |
Date: 07 Jun 2008 14:49:46
From: Guest
Subject: Re: What is Rating of Rybka and other Programs?
|
"Sanny" <[email protected] > wrote in message
news:4290e7d0-abb0-4fd4-b676-dbcd17a7cb76@y22g2000prd.googlegroups.com...
>> I read on internet Rybka is 3100 rated. And Fritz is 2900 Rated is it
>> true?
>>
>> When they say Rybka is 3100 Rated How much time they give it for each
>> Move.
>
> And also Which Computer they Play to test the Rating?
>
> Is it Celeron? Dual Core? Quad Core? or a Super Computer.
How should we know?
You are the one who read on the web that these programs had those ratings.
So, look on the web. Look for actual test results rather than somebody just
saying it or just playing a few games.
It could be any type of system. If it was done on an even tournament, then
possibly all of them were running on the same hardware.
If it was a normal tournmanet style test, then they could have been running
on any hardware they watned. You could have one running an old 32 bit
celeron laptop, while some other is running on a massive 256 core cluster
supercomputer.
So unless you know what hardware it was done on, and the time control
settings, those ratings could go up or down by 300 points or more.
> Since rating can be increased a lot by giving a good hardware.
Sure. A fast 64 bit system is going to help a bitboard program a lot. It
wont help a mailbox program much, though.
If you force a bitboard program to run on a 32 bit system, then you've just
slowed the program down by half or so. (32 vs. 64 bits, plus the loss of
the extra registers)
So if those tests were all done on the same hardware, and it wasn't 64 bit,
then some programs could be at a serious disadvantage.
> Bye
> Sanny
>
> Play Chess at: http://www.GetClub.com/Chess.html
>
>
>
>
----== 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: 07 Jun 2008 20:17:10
From: Morten Skarstad
Subject: Re: What is Rating of Rybka and other Programs?
|
Sanny skrev:
>> I read on internet Rybka is 3100 rated. And Fritz is 2900 Rated is it
>> true?
>>
>> When they say Rybka is 3100 Rated How much time they give it for each
>> Move.
>
> And also Which Computer they Play to test the Rating?
>
> Is it Celeron? Dual Core? Quad Core? or a Super Computer.
>
> Since rating can be increased a lot by giving a good hardware.
Uhm, no. Rating of pure software chess engines is based on how they
perform on _equal_ grounds. That is they have the same hardware, same
time controls, same number of pieces on the board etc.
If I made a chess engine and said "hey, wow, my engine on a quad core
computer can beat Rybka on a Pentium classic most of the time if I give
my engine 2 hours per move versus Rybka 10 seconds per move, my engine
plays white all the time, and my engine starts with a queen up!" it
would mean exactly squat.
|
| | | |
Date: 07 Jun 2008 15:01:40
From: Guest
Subject: Re: What is Rating of Rybka and other Programs?
|
"Morten Skarstad" <`whoami`@localhost > wrote in message
news:[email protected]...
> Sanny skrev:
>>> I read on internet Rybka is 3100 rated. And Fritz is 2900 Rated is it
>>> true?
>>>
>>> When they say Rybka is 3100 Rated How much time they give it for each
>>> Move.
>>
>> And also Which Computer they Play to test the Rating?
>>
>> Is it Celeron? Dual Core? Quad Core? or a Super Computer.
>>
>> Since rating can be increased a lot by giving a good hardware.
>
> Uhm, no. Rating of pure software chess engines is based on how they
> perform on _equal_ grounds. That is they have the same hardware, same time
> controls, same number of pieces on the board etc.
Depends on who's doing the tests.
If somebody is actually running those programs in their own tournament, then
they probably are going to do it on similar hardware. It could be whatever
they had. An antique Athlon, an even older 486, or a modern Core 2 Quad.
32 or 64 bit OS.
On the other hand, if this was done based on a chess server, then everybody
will be running their own programs on their own hardware. Which may or not
be their full strength tournament hardware. And you'll have all sorts of
hardware. A celeron laptop, an old spare P4, a dual core, a quad core, a
few spare cores on some cluster, etc.
So unless you know who did the tests and where the rating comes from, you
just don't know.
> If I made a chess engine and said "hey, wow, my engine on a quad core
> computer can beat Rybka on a Pentium classic most of the time if I give
If the Rybka author chose to use a pentium class, then that would be his
choice. And that rating would be valid.
It wouldn't be up to you to run it on that differeing hardware though.
> my engine 2 hours per move versus Rybka 10 seconds per move, my engine
> plays white all the time, and my engine starts with a queen up!" it would
> mean exactly squat.
----== 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 =---
|
|
