When GetClub becomes worlds best [Top 100]?
Sanny
Engines
I am thinking in 2-3 years GetClub will come in [Top 10]
Please Test it with other Engines:[Fritz/ Rybka/ ChessMaster]
http://www.GetClub.com/Chess.html
Let me know whats the performance of GetClub against other engines. I
tested against Rybka but my connection breaks in half of the game.
I want to know if GetClub Chess plays as good as [Top 10] programs.
Since there are 5 levels
Baby [2 sec / move] 2200+ As good as Help Bot/ Taylor Kingston
Beginner [15 sec / move] 2400+ As good as Phil Inns/ SamSloan [May be]
Easy [1 min/ move] 2600+ As good as International Master
Normal [4 min/ move] 2800+ As good as a Grand Master
What will be the rating of GetClub. Since taking longer time will
improve the rating. As Fritz has a rating of 3000+ How many seconds
Fritz thinks to play at 3000+ rating?
What should I do when my program starts beating Rybka/ Fritz/
ChessMaster etc?
I want some Chess Magazines to write about GetClub Chess. Saying that
"GetClub is in Top 10 Chess Programs."
Bye
Sanny
Wlodzimierz Holsztynski (Wlod)
>
> I am thinking in 2-3 years GetClub will come in [Top 10]
And a pig will fly on a broom to another galactic.
Sanny, please, do me a favor - stop using the phrase
"am thinking". Avoid all forms of the word "think".
Thank you for your ... understanding :-)
Wlod
Wlodzimierz Holsztynski (Wlod)
<sennaj...@gmail.com> wrote:
> On Nov 3, 11:52 pm, Sanny <softtank...@hotmail.com> wrote:
>
>
>
> > I am thinking in 2-3 years GetClub will come in [Top 10]
>
> And a pig will fly on a broom to another galactic.
Make it "galactica" or "galaxy".
Wlod
Sanny
>
> > And a pig will fly on a broom to another galactic.
>
> Make it "galactica" or "galaxy".
Just like Harry Potter.
You can see a small documentary on Harry Potter Movie here
http://www.getclub.com/Show/view.php?best=Movies&itemid=12
I love the way Harry Potter is tortured by his uncle & aunt.
Do you still have agony with GetClub?
What are the problems you face while playing Chess at GetClub?
I want to see your games at GetClub Chess.
Bye
Sanny
Thomas T. Veldhouse
> Now GetClub is much improved. Now, Its playing like other strong Chess
> Engines
>
Wow .. I want some of what ever the fuck that you are smoking!
Sorry for the language, but it is required to push the point home here.
--
Thomas T. Veldhouse
Religion is a crutch, but that's okay... humanity is a cripple.
Offramp
> I love the way Harry Potter is tortured by his uncle & aunt.
How awful.
Simon Krahnke
> I love the way Harry Potter is tortured by his uncle & aunt.
You love torturing? Get treatment.
mfg, simon .... l
Wlodzimierz Holsztynski (Wlod)
> * Sanny <softtank...@hotmail.com> (16:06) schrieb:
>
> > I love the way Harry Potter is tortured by his uncle & aunt.
>
> You love torturing? Get treatment.
>
> mfg, simon .... l
Ok, so Sanny is a masochist.
Does it mean that he needs a "treatment"?
I assume, simon - based on Sanny's general
comprehension level, that you mean "psychiatric treatment",
and not the (unknown to me) treatment received
by Harry Potter, whatever it might be.
Wlod
madams
>
> On Nov 4, 7:07 pm, Simon Krahnke <overl...@gmx.li> wrote:
>
> > * Sanny <softtank...@hotmail.com> (16:06) schrieb:
> >
> > > I love the way Harry Potter is tortured by his uncle & aunt.
> >
> > You love torturing? Get treatment.
> >
> > mfg, simon .... l
>
> Ok, so Sanny is a masochist.
> Does it mean that he needs a "treatment"?
--------
Yeah ! it really does mean "that" Wlodo - it means Sanny ia a half-arsed
vegetarian sub-continental "grott" who endjoys waving his smelly
"backend" to Euro financiers...
He hopes; "they" will give 'im... 'carte-blanche' to euro/monies but
he's a non-player, a pathetic waNKER - a masturbator, a sickening little
dark-skinned creep !
Sanny is the most vile & un'tooned vulgar crepsolum to be featured here
in the longest of yawwnnn... times
m.....
___________
Sanny
See the film and you will enjoy that too.
http://www.getclub.com/Show/view.php?best=Movies&itemid=12
Its just comedy.
Bye
Sanny
Sanny
> * Sanny <softtank...@hotmail.com> (16:06) schrieb:
>
> > I love the way Harry Potter is tortured by his uncle & aunt.
>
> You love torturing? Get treatment.
No it was a comedy See the film and you will enjoy that too.
Markus Islinger
It would never lose to anyone because even trees would not live long
enough to make it past the 10th move.
I wonder how anyone can have the patience to play the master level. I
rarely try normal level because I feel I am wasting my time when it is
thinking for more than 30 minutes.
Markus
tjb
> Since there are 5 levels
>
> Baby [2 sec / move] 2200+ As good as Help Bot/ Taylor Kingston
>
> Beginner [15 sec / move] 2400+ As good as Phil Inns/ SamSloan [May be]
>
> Easy [1 min/ move] 2600+ As good as International Master
>
> Normal [4 min/ move] 2800+ As good as a Grand Master
2400+ is beginner level?
Simon Krahnke
> On Nov 5, 8:07�am, Simon Krahnke <overl...@gmx.li> wrote:
>> * Sanny <softtank...@hotmail.com> (16:06) schrieb:
>>
>> > I love the way Harry Potter is tortured by his uncle & aunt.
>>
>> You love torturing? Get treatment.
>
> No it was a comedy See the film and you will enjoy that too.
So it's a movie where young people learn how funny torturing
small boys is. Weren't there some cases of this in Great Britain
following this film?
mfg, simon .... l
Thomas T. Veldhouse
Yeah ... and 15 seconds a move too! Isn't it great?
Martin Brown
Subtract 1000 to get normal ELO ratings from Sanny guesses.
Regards,
Martin Brown
help bot
wrote:
> >> Baby [2 sec / move] 2200+ As good as Help Bot/ Taylor Kingston
>
> >> Beginner [15 sec / move] 2400+ As good as Phil Inns/ SamSloan [May be]
>
> >> Easy [1 min/ move] 2600+ As good as International Master
>
> >> Normal [4 min/ move] 2800+ As good as a Grand Master
>
> > 2400+ is beginner level?
>
> Subtract 1000 to get normal ELO ratings from Sanny guesses.
In my latest adventure at GetClub, I tried a Budapest
Gambit as Black against the "Master" level; a lot of
trades occurred early on, and we emerged into an /al-
most level/ endgame. (Don't blame me-- it's this awful
opening!)
Apparently, the GetClub program decided to chuck
a pawn in return for a spite-check and temporary con-
trol of a few squares next to my king, wrongly thinking
this is what constitutes an "attack". Now, with pawn
in hand, I quite easily outplayed Sanny's monstrosity,
forcing mate in short order.
But let's go back to the point at which I noted that
the Master level emerged from the opening unscathed;
that is quite an improvement! In GetClub's early days,
it simply *never* managed such a feat; no, not ev-ahh.
Thus, the nay-sayers who insist the program still re-
mains at or near the same level of strength as they
themselves claimed before (i.e. at or near 1000 Elo)
are deceiving themselves almost as much as Sanny.
If indeed Sanny belongs in an asylum for the insane,
then he should at least have some rgc company, who
are themselves chessically delusional in that none
seem able to differentiate between the real and the
purely imaginary improvement... .
-- help bot
JVMe...@aol.com
No, it won't.
> Please Test it with other Engines:[Fritz/ Rybka/ ChessMaster]
There's no point in testing your engine against Rybka / Fritz /
Chessmaster, as it will lose every game.
> Let me know whats the performance of GetClub against other engines. I
> tested against Rybka but my connection breaks in half of the game.
And that's part of the reason.
> I want to know if GetClub Chess plays as good as [Top 10] programs.
Not even remotely close. Not even top 100. Possibly not even top 200.
> What should I do when my program starts beating Rybka/ Fritz/
> ChessMaster etc?
Thank the person who rewrote every line of code in it.
My guess is that your engine is not even in the top 150 of KNOWN
engines in the world, let alone all of the unpublished ones.
Here's a thought. I've gone to a relatively well-known rating list
where over 200 engines are rated using tens of thousands of games.
http://wbec-ridderkerk.nl/html/BayesianElo_ed16.htm
Try a 20-game match against the #148 engine, TJChess, using a
consistent time control throughout the match. Pick any time control
that you like. I choose TJChess because it is a recent one that can
now be downloaded at http://tonyjh.com/chess/tjchess. TJ Chess is
relatively consistently rated around 2000-2100.
Of course, having a match of this type is almost impossible because
your engine doesn't know anything at all about clock management, so it
will almost certainly lose several games on time no matter what you
do. But that's part of what REAL computer chess is all about -- as
opposed to your crappy engine and even crappier website. As long as
your engine doesn't know about a clock, anything you say about its
strength is a lie.
If you are able to actually have this match, and you count losses on
time, you'll see how bad your engine is. Remember, GetClub will have
to win the match just to be rated around 2050.
But it won't, because it isn't.
jm
Martin Brown
>> I want to know if GetClub Chess plays as good as [Top 10] programs.
>
> Not even remotely close. Not even top 100. Possibly not even top 200.
How well does it do these days on the WAC300 test set of tactical
motifs? My engine currently gets 233/300 at nominal 10s / position.
That isn't particularly good.
>> What should I do when my program starts beating Rybka/ Fritz/
>> ChessMaster etc?
>
> Thank the person who rewrote every line of code in it.
>
> My guess is that your engine is not even in the top 150 of KNOWN
> engines in the world, let alone all of the unpublished ones.
>
> Here's a thought. I've gone to a relatively well-known rating list
> where over 200 engines are rated using tens of thousands of games.
>
> http://wbec-ridderkerk.nl/html/BayesianElo_ed16.htm
>
> Try a 20-game match against the #148 engine, TJChess, using a
> consistent time control throughout the match. Pick any time control
> that you like. I choose TJChess because it is a recent one that can
> now be downloaded at http://tonyjh.com/chess/tjchess. TJ Chess is
> relatively consistently rated around 2000-2100.
That is unkind.
He would be better off testing against something around the 1800 mark
judging from the games. That way it might stand a (slight) chance of
winning in an endgame. Engines don't easily succumb to human blunders in
complex tactical positions so in an engine-engine match it would have to
see beyond the others tactical horizon. #177 Micro-max 4.8m would be a
better choice of opponent if he is to stand any chance of winning...
http://home.hccnet.nl/h.g.muller/umax4_8w.c
Full sourcecode - all 2kbytes of it - an impressive feat.
> If you are able to actually have this match, and you count losses on
> time, you'll see how bad your engine is. Remember, GetClub will have
> to win the match just to be rated around 2050.
> But it won't, because it isn't.
It would be amusing to see. Does someone with the required patience
fancy playing u-max against it for a quantitative benchmark. At the very
least it will calibrate GetClub in real world computer match ELO score
if it is within 1800 +/- 400 after a few games.
Regards,
Martin Brown
JVMe...@aol.com
wrote:
>
> It would be amusing to see. Does someone with the required patience
> fancy playing u-max against it for a quantitative benchmark. At the very
> least it will calibrate GetClub in real world computer match ELO score
> if it is within 1800 +/- 400 after a few games.
The time control problem is still an issue. Until that is fixed, I
would guess that GetClub would lose at least half its games on time --
and losing ALL of them on time is not out of the question. Until both
engines are operating under identical settings and circumstances,
trying to determine a rating is pointless.
jm
Martin Brown
It would put an upper bound on it if GetClub was allowed whatever time
it took to run at the nominal setting. Has it ever beaten Jester yet?
How about making two estimates one with time controls and one without...
Regards,
Martin Brown
JVMe...@aol.com
wrote:
There is no such thing as a rating without an attached time control.
How would that be managed?
As for my engine, Myrddin, my latest version gets 280/300 on WAC at
10s per position.
I would be very surprised if GetClub (also at 10s per position) even
got 200/300. Too bad there's no way to find out.
jm
Simon Krahnke
> If you are able to actually have this match, and you count losses on
> time, you'll see how bad your engine is. Remember, GetClub will have
> to win the match just to be rated around 2050.
As far as I know, Getclub has never won a single match. Because the only
the way it can properly end a game is by giving up.
I've heard about a single instance where someone got a draw by
repetition. Was that help-bot oder Sanny himself?
mfg, simon .... l
Simon Krahnke
> I would be very surprised if GetClub (also at 10s per position) even
> got 200/300. Too bad there's no way to find out.
Well, there is a way. But it's probably not worth the effort.
mfg, simon .... l
Martin Brown
>> It would put an upper bound on it if GetClub was allowed whatever time
>> it took to run at the nominal setting. Has it ever beaten Jester yet?
>>
>> How about making two estimates one with time controls and one without...
> There is no such thing as a rating without an attached time control.
> How would that be managed?
Set time control long enough that even GetClub can complete a game in
the required time. For its nominally 10s per move allowing 3 hours per
game ought to do it (ie 1000x nominal time per move).
Engine results don't improve all that much with small linear increases
in time spent so it should not bias things by all that much. And if as
we both expect it will fail low then we have a quantitative result!
If it does win then we have to accept that we handicapped the opposition.
>
> As for my engine, Myrddin, my latest version gets 280/300 on WAC at
> 10s per position.
That is pretty good! Slightly better than Rybka 2.32 manages if memory
serves. Out of curiosity how do the ones your engine fails on compare
with the 21 that mine cannot get?
The tricky ones were
2, 71, 74, 80, 87, 100, 141, 146, 163, 196, 213, 222, 229, 230, 235,
247, 250, 262, 264, 265, 269, 274
Of which I think my engine got a marginally better or equal (but wrong
by the book) answer to #100, #235 and #264
>
> I would be very surprised if GetClub (also at 10s per position) even
> got 200/300. Too bad there's no way to find out.
ISTR he did try it at some point last year and scored something like 30%
at 30s per position.
Regards,
Martin Brown
Patrick Volk
I would submit that Sanny is referring to the Zelo (Zimbabwean Elo)
point system.
chip...@gmail.com
Yes, they made him play chess at GitClub and he couldn't take it for
very long.