Easy Level as good as Fritz 8 [Start a Challenge]
Sanny
Anyone having Fritz ready to play 5 games beween Fritz 8 & Easy Level?
Fritz 8 [5 sec / move] vs Easy Level [32 sec / move] Will be an
interesting match.
Which side will you take if there is game between Fritz[5 sec] vs Easy
Level?
I think both will play equally good.
Try and play 5 games: http://www.GetClub.com/Chess.html
and let us know who wins the majority of the games?
GetClub only uses single core So do not give multiple cores to Fritz.
Since the games are recorded we can watch each game played between
Fritz [5 sec/move] Vs Easy Level [32 sec / move]
I hope GetClub will play equally good to Fritz 8.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
Sanny
Then we can have a match between Easy Level [32 sec / move] vs Crafty
[5 sec / move]
And also Easy Level [32 sec / move] vs ChessMaster [5 sec / move]
Also if possible Easy Level [32 sec / move] vs Rybka [5 sec / move]
I ask you to give 5 sec to Commercial Programs, while 32 sec to
GetClub Chess as GetClub is a new program. So It needs 6 times more
time to compete with the strong programs.
I hope by giving 32 sec / move to Easy Level, GetClub will be able to
compete with the Commercial Programs.
Bye
Sanny
Martin Brown
> Now Easy Level of GetClub plays as good as Fritz 8.
>
> Anyone having Fritz ready to play 5 games beween Fritz 8 & Easy Level?
>
> Fritz 8 [5 sec / move] vs Easy Level [32 sec / move] Will be an
> interesting match.
Even Fritz 5.32 the now prehistoric free engine will destroy GetClub
totally. As will any of the the other toy engines on the free download
section of CB with the possible exception of Turing. And even then with
Turing I would not like to bet on Getclub being able to win.
The only engines you have a chance of beating are weaker amateur engines
that have been suggested before. What does GetClub score on a
standardised test of tactical ability like WAC these days?
My old VAX chess program that I am resurrecting from prehistoric source
code manages 234/300 on that test at 10s per position, and that is still
a rubbish level of tactical play for a chess engine.
>
> Which side will you take if there is game between Fritz[5 sec] vs Easy
> Level?
Fritz8 5 - 0 Getclub
And I would predict the same result for any modern commercial engine at
any thinking time that allowed it to reach at least 6 ply. You might
perhaps stand a chance against Sargon II or something of that era.
>
> I think both will play equally good.
In heavens name why?
>
> Try and play 5 games: http://www.GetClub.com/Chess.html
>
> and let us know who wins the majority of the games?
>
> GetClub only uses single core So do not give multiple cores to Fritz.
>
> Since the games are recorded we can watch each game played between
> Fritz [5 sec/move] Vs Easy Level [32 sec / move]
>
> I hope GetClub will play equally good to Fritz 8.
You have been saying this for years now. Why not do a bit of testing
before making these outlandish comments about engine srength?
Regards,
Martin Brown
Sanny
>
> You have been saying this for years now. Why not do a bit of testing
> before making these outlandish comments about engine srength?
I tested GetClub against Jester and GetClub Beginner Level now plays
as good Jester at 30 sec / move.
That is my basis of expactation.
I wish players having the commercial Engines test and provide the
complete details. As the games are recorded we can see how GetClub
played against each Engine.
I have been playing against IVAN -1800+ and Find GetClub plays much
better than IVAN.
I agree a few months back GetClub was much weak. But this month
GetClub was improved a lot.
So I think nothing can be hidden. Those who do not believe me.
Test Fritz 5.0 [5 sec / move] vs Easy Level [32 sec / move]
Try and play 5 games: http://www.GetClub.com/Chess.html
If you have Crafty/ Rybka / Chess Master etc
Instead of just assumptions I am looking forward for real challenges.
I will be happy to see the results, wether they go in GetClubs favor
or any other Chess Engines favor
LETS HAVE REAL CHALLENGE and we see who wins?
Other Engines [5 sec / move] vs GetClub Easy Level [32 sec / move]
Bye
Sanny
zzz
"challenge".
Rather than using a commercial engine though, I went to this chess
engine testing site: http://www.computerchess.org.uk/ccrl/4040/rating_list_all.html
and downloaded the WORST engine on the list: Dabbaba 2.62, ranked
number 180.
As asked, I ran the engine on a single core CPU at 5 seconds per move
(any extra time = operator time).
I had no doubts about what the outcome was going to be, and I fully
realize that Sanny only posts these rubbish boasts to draw people to
his website, but as I said: I really had nothing better to do ;-).
This is the game:
White: ChessChallenger7, played by Dabbaba 2.62
Black: GetClub, Easy Level
1. d4 Nf6
2. c4 e6
3. Nf3 d5
4. Nc3 Be7
5. Bf4 O-O
6. e3 Nc6 {blocking the c-pawn, making it hard for black to fight for
the centre}
7. Be2 Bb4 {losing time moving the same piece twice in the opening}
8. O-O dxc4 {surrendering the centre for no reason}
9. Bxc4 Bxc3 {giving up the bishop pair for no reason}
10. bxc3 Ne4 {wasting more time, this knight will soon be kicked back}
11. Qc2 f5
12. Bd3 Nf6
13. Rab1 Nd5
14. Bg5 Qd6
15. e4 fxe4
16. Bxe4 Rb8 {giving up a vital pawn for no reason}
17. Bxh7+ Kh8
18. Nh4 Rf7 {anything but this, now it's mate in 4}
19. Ng6+ Kxh7
20. Nf8+ Kg8
21. Qh7+ Kxh7 {an illegal move to finish things off} 1-0
A totally hopeless, clueless game by GetClub against the number 180 on
the "40/40" ranking list.
There really is no point in playing any further games between these
two. It's like a toddler taking on a heavyweight boxer.
JVMe...@aol.com
>
> And also Easy Level [32 sec / move] vs ChessMaster [5 sec / move]
I would do this, but I know that Chessmaster will win every game
easily -- even at ONE second per move -- probably even against your
Advanced Level, if I had a few months to play through one game.
Sanny, you do nothing on this forum but lie. Why don't you just go
away, please?
jm
Sanny
> the "40/40" ranking list.
> There really is no point in playing any further games between these
> two. It's like a toddler taking on a heavyweight boxer.
Thanks, for playing a game and analysing the mistakes of GetClub.
I hope your analysis will help in improving the GetClub Chess further.
GetClub is quite new So I expacted such results. I think It is still a
long way for GetClub to come in Top 100.
I feel your Commenary on the games will be quite helpful.
Kelsey Bjarnason
On Wed, 23 Dec 2009 00:29:40 -0800, Sanny wrote:
> Now Easy Level of GetClub plays as good as Fritz 8.
...
> Fritz 8 [5 sec / move] vs Easy Level [32 sec / move]
If Easy level is "as good as" Fritz 8, why does it need more than a 6:1
time advantage? If it is, in fact, "as good as" Fritz 8, it should hold
its own at equal time settings.
By your logic, GetClubbed should be stronger than Deep Blue - just give
DB 100 moves per minute and give GetClubbed 100 minutes per move. Wow,
imagine, it might even hold its own - when the challenged is rigged.
Simon Krahnke
> I tested GetClub against Jester and GetClub Beginner Level now plays
> as good Jester at 30 sec / move.
You mean Getclub still loses every single game, but you just jump the
conclusion they are as good.
> That is my basis of expactation.
Surprise!
> I wish players having the commercial Engines test and provide the
> complete details. As the games are recorded we can see how GetClub
> played against each Engine.
After you somehow fix the list of moves Getclub fucked up.
> I have been playing against IVAN -1800+ and Find GetClub plays much
> better than IVAN.
How, does it win? Or, at least, claim win?
> I agree a few months back GetClub was much weak. But this month
> GetClub was improved a lot.
Sure, it now can fuck up the list of moves.
> Then we can have a match between Easy Level [32 sec / move] vs Crafty
> [5 sec / move]
I know I'm boring, but I stick to the old Rybka version, Getclub lost
many games to.
> Also if possible Easy Level [32 sec / move] vs Rybka [5 sec / move]
Impossible. I just set it to search for one ply, which is the minimum.
[Date "2009.12.24"]
[White "Getclub easy"]
[Black "Rybka 2.2 32 bit"]
[Result "0-1"]
1. e4 Nf6 {(Nf6) +0.02/1 0}
2. Nc3 Nc6 {(Nc6) +0.08/1 0}
3. Nf3 d5 {(d5) +0.02/1 0}
4. Nxd5 Nxd5 {(Nxd5) +0.21/1 0}
5. exd5 Qxd5 {(Qxd5) +0.24/1 0}
6. d3 Bf5 {(Bf5) +0.27/1 0}
7. Bf4 e5 {(e5) +0.35/1 0}
8. Be3 O-O-O {(0-0-0) +0.63/1 3} { immediately back to e3, wtf? }
9. a3 Bc5 {(Bc5) +0.75/1 0}
10. Bxc5 Qxc5 {(Qxc5) +0.77/1 0}
11. Be2 e4 {(e4) +1.01/1 2}
12. d4 Qe7 {(Qe7) +1.46/1 0}
13. Nd2 Nxd4 {(Nxd4) +1.42/1 0} { wooh, that's actually less than the previous score }
14. O-O e3 {(e3) +1.54/1 0}
15. fxe3 Qxe3+ {(Qxe3+) +1.70/1 0}
16. Kh1 Nxe2 {(Nxe2) +3.21/1 0} { BAMM! }
17. Rxf5 Rxd2 {(Rxd2) +3.95/1 0}
18. Qb1 f6 {(f6) +4.76/1 0}
19. Ra5 a6 {(a6) +5.32/1 0}
20. c4 Nf4 {(Nf4) +5.31/1 0}
21. Qf5+ Kb8 {(Kb8) +5.54/1 0}
22. Rf1 Qe2 {(Qe2) +5.57/1 0}
23. Qxf4 Qxg2# {(Qxg2+) +M1/1 0} { How come Getclub didn't see that coming? }
0-1
It took Rybka 8 seconds (it somehow spent 3s on less than 600 nodes) and
23 moves to win this game.
It seems Rybka plays best, when just on quiescence search. ;->
Straightforward positional play is usually the way to go against really
weak engines like Getclub's. Don't overthink it, straight to the win!
To answer your question on how Rybka sees exchanges when looking only 1
ply deep: I don't know about Rybka, but conventional wisdom is to make a
quiescence search, that only does the violent moves like captures and
promotions, before evaluating a position.
> I ask you to give 5 sec to Commercial Programs, while 32 sec to
> GetClub Chess as GetClub is a new program. So It needs 6 times more
> time to compete with the strong programs.
Since how many years has Getclub been new?
What about teaching Getclub to actually win and lose games? What's
wrong with Getclub, that that ain't possible?
mfg, simon .... l