Not logged inRybka Chess Community Forum
Forum RybkaChess Help Search Register Login
Up Topic Rybka Support & Discussion / Rybka Discussion / Leiden Round 7: Rybka vs Critter
1 2 3 Previous Next  
- - By Ray (****) Date 2012-05-13 14:04
Move 15, Rybka as white to move
Parent - - By Ray (****) Date 2012-05-13 14:07
1. d4 d5 2. c4 e6 3. Nf3 c6 4. e3 Bd6 5. b3 f5 6. Bd3 Nf6 7. O-O Qe7 8. Bb2 O-O 9. Qc1 b6 10. Ba3 Bb7 11. Bxd6 Qxd6 12. Nc3 dxc4 13. Bxc4 Nbd7 14. Re1 b5 15. Bd3 b4
Parent - By Ray (****) Date 2012-05-13 14:11
1. d4 d5 2. c4 e6 3. Nf3 c6 4. e3 Bd6 5. b3 f5 6. Bd3 Nf6 7. O-O Qe7 8. Bb2 O-O 9. Qc1 b6 10. Ba3 Bb7 11. Bxd6 Qxd6 12. Nc3 dxc4 13. Bxc4 Nbd7 14. Re1 b5 15. Bd3 b4 16. Na4 c5 17. Nxc5 Bxf3 18. gxf3 Rfc8
Parent - By Nick (*****) [gb] Date 2012-05-13 14:11
14.Re1 is a novelty to me. I quite like white, black's d6 square might be a problem.
Parent - By Ray (****) Date 2012-05-13 14:15
1. d4 d5 2. c4 e6 3. Nf3 c6 4. e3 Bd6 5. b3 f5 6. Bd3 Nf6 7. O-O Qe7 8. Bb2 O-O 9. Qc1 b6 10. Ba3 Bb7 11. Bxd6 Qxd6 12. Nc3 dxc4 13. Bxc4 Nbd7 14. Re1 b5 15. Bd3 b4 16. Na4 c5 17. Nxc5 Bxf3 18. gxf3 Rfc8 19. Bc4 Nxc5 20. dxc5 Rxc5 21. a3 bxa3
Parent - - By Ray (****) Date 2012-05-13 14:24 Edited 2012-05-13 14:28
1. d4 d5 2. c4 e6 3. Nf3 c6 4. e3 Bd6 5. b3 f5 6. Bd3 Nf6 7. O-O Qe7 8. Bb2 O-O 9. Qc1 b6 10. Ba3 Bb7 11. Bxd6 Qxd6 12. Nc3 dxc4 13. Bxc4 Nbd7 14. Re1 b5 15. Bd3 b4 16. Na4 c5 17. Nxc5 Bxf3 18. gxf3 Rfc8 19. Bc4 Nxc5 20. dxc5 Rxc5 21. a3 bxa3  22. Qxa3 Nd5 23. Red1 Qc7 24. Kg2 Rd8 25. Bxd5 Rdxd5 26. Qa4 Qe7 27. Rxd5 Rxd5 28. Qxa7 Qxa7 29. Rxa7
Parent - By Arifur (**) [se] Date 2012-05-13 14:27
where are you watching the game?
Parent - - By Ray (****) Date 2012-05-13 14:32
1. d4 d5 2. c4 e6 3. Nf3 c6 4. e3 Bd6 5. b3 f5 6. Bd3 Nf6 7. O-O Qe7 8. Bb2 O-O 9. Qc1 b6 10. Ba3 Bb7 11. Bxd6 Qxd6 12. Nc3 dxc4 13. Bxc4 Nbd7 14. Re1 b5 15. Bd3 b4 16. Na4 c5 17. Nxc5 Bxf3 18. gxf3 Rfc8 19. Bc4 Nxc5 20. dxc5 Rxc5 21. a3 bxa3  22. Qxa3 Nd5 23. Red1 Qc7 24. Kg2 Rd8 25. Bxd5 Rdxd5  26. Qa4 Qe7 27. Rxd5 Rxd5 28. Qxa7 Qxa7 29. Rxa7 Rb5 30. Re7 Rb6
Parent - - By Highendman (****) Date 2012-05-13 14:41
After 30 moves - this looks totally drawish. I'd be amazed to see if Rybka can win this - I don't believe it's possible. Likely a draw :(
Parent - - By Ray (****) Date 2012-05-13 14:43
1. d4 d5 2. c4 e6 3. Nf3 c6 4. e3 Bd6 5. b3 f5 6. Bd3 Nf6 7. O-O Qe7 8. Bb2 O-O 9. Qc1 b6 10. Ba3 Bb7 11. Bxd6 Qxd6 12. Nc3 dxc4 13. Bxc4 Nbd7 14. Re1 b5 15. Bd3 b4 16. Na4 c5 17. Nxc5 Bxf3 18. gxf3 Rfc8 19. Bc4 Nxc5 20. dxc5 Rxc5 21. a3 bxa3  22. Qxa3 Nd5 23. Red1 Qc7 24. Kg2 Rd8 25. Bxd5 Rdxd5  26. Qa4 Qe7 27. Rxd5 Rxd5 28. Qxa7 Qxa7 29. Rxa7 Rb5 30. Re7 Rb6 31. Kg3 Kf8 32. Rd7 Rxb3 33. Kf4 Rb5
Parent - By Highendman (****) Date 2012-05-13 14:46
Yep, looks totally drawish. A strong bragging right for Critter getting to draw Cluster Rybka (unless something shocking happens later in this game).
Parent - - By Ray (****) Date 2012-05-13 14:46 Edited 2012-05-13 14:49
Probably a draw unless there is an an endgame bug in Critter. Did Rybka play safe with contempt and play for a draw ?

34. h4 h6 35. Kg3 Rb8 36. h5 Re8
Parent - - By Ray (****) Date 2012-05-13 14:55
37. Rb7 Re7 38. Rb5 Kf7 39. e4 Kf6 40. exf5 exf5
Parent - - By Ray (****) Date 2012-05-13 15:00
41. Kf4 Re2 42. Rxf5+ Ke6 43. Rb5 Rxf2 44. Rb6+ Kf7 45. Rb7+ Kf6
Parent - By Ray (****) Date 2012-05-13 15:03
Operators accepted a draw.
Parent - - By Venator (Silver) [nl] Date 2012-05-13 14:56
Did Rybka play safe with contempt and play for a draw ?

No, Rybka showed a white advantage during most of the game. Winning a pawn reaching a dead drawn rook ending was apparently the best she could squeeze from the position.
Parent - - By Highendman (****) Date 2012-05-13 15:08
Congrats on 6.5/7!

Can you please share what was the eval around move 30? Any decent human player without engine assistance can guess it's a total draw at that position. I'm curious if Rybka realized that too by that move?
Parent - - By Lukas Cimiotti (Bronze) [de] Date 2012-05-13 15:19
At move 30 the cluster had an eval of 0.07
Parent - - By Dr.X (Gold) Date 2012-05-13 15:32
I would have loved to seen this game boardcasted!
Parent - - By Lukas Cimiotti (Bronze) [de] Date 2012-05-13 15:34
I would have done it - but I found out the tool I wrote for getting positions out of the log files isn't fully compatible with this version of the cluster. And I didn't have the time to fix it.
Parent - By Dr.X (Gold) Date 2012-05-13 15:36
That's cool-I'm glad you tried!
Parent - - By h.g.muller (****) [nl] Date 2012-05-13 18:55

> I would have loved to seen this game boardcasted!


Well, it was actually 'broadcasted', because Critter was playing all its games through the Variant ICS, on which it was logged on as a bot. We just logged in from the tournament room to that same ICS to challenge the bot. So anyone that would have logged on to the ICS would have been able to follow the game live.

It was not very interesting, though, as we decided to switch off Critter's PV kibitzing. We had connection problems during the first two games of the tourney, Critter losing the connection with the ICS several times. After we switched off the PV kibitzing this did not happen again. This might have been pure coincidence, but we did not feel like using the tourney to find that out.

But because of the sub-standard functionality I did not announce the game anywhere. Sorry about that.
Parent - By Dr.X (Gold) Date 2012-05-13 20:13
Oh, well, perhaps may be the next time around a more studied attempt to broadcast these games can be made. I think the interest in there.
Parent - By Venator (Silver) [nl] Date 2012-05-13 15:22
Thanks! But the congrats should go to Vas, Jiri, Lukas and Hans :-).

I phoned Hans around move 20 and at that point the score was something like +0.43.
Parent - - By Lukas Cimiotti (Bronze) [de] Date 2012-05-13 15:45
We played with a very low contempt. But I would have preferred a win :wink:
Parent - - By Ray (****) Date 2012-05-13 15:57
Yes, the game broke your perfect record in the tourney !
Parent - - By Chaotic Chess (****) Date 2012-05-13 16:27
On what hardware critter was playing?
Parent - - By Ray (****) Date 2012-05-13 16:28
AMD Phenom II X6 @ 3.7GHz
Parent - - By Chaotic Chess (****) Date 2012-05-13 16:34
It's a big achievement for critter! :wink:
Parent - - By Barnard (Bronze) Date 2012-05-13 16:42
and show us that even running on the cluster,rybka has been overpassed by today top engines rubbin on a merely 6 cores AMD,since they can tie with rybka from 6 AMD cores to 100 cluster core,that never at past (just 3 years ago) no one person can imagine that an engine can draw a game against rybka running on a 6 core against 100

as i said,im sure that if rybka ran in equal hardware,critter will beat her
Parent - - By Ray (****) Date 2012-05-13 17:14

> as i said,im sure that if rybka ran in equal hardware,critter will beat her


I don't think so. If Rybka had been playing say on an overclocked quad Intel Sandybridge, I think the tournament result would have been the same.
Parent - - By Barnard (Bronze) Date 2012-05-13 17:28
maybe not,since rybka never will played on a cluster and maybe made some draws,the same than critter,and will be tied the tournament and shared the 1 position

that is why they played with the cluster,to asure 100 % wins against the weaker engines,and even if rybka draw against critter or other top engine,that 100 % wins against the weakers makes rybka winner of the tournament,because the other top engines will surely make some draws against the weakers
Parent - - By Ray (****) Date 2012-05-13 17:30
Critter drew against Spark and Pandix, so yes I understand what you are saying. But would Rybka on a quad have been enough ? I think it would have, but we will never know.
Parent - - By kgburcham (***) [us] Date 2012-05-13 17:49
would Rybka on a quad have been enough ? I think it would have
I assume you are saying that the Rybka version that played in this tournament on 100 threads, would have the same results if it had played on 4 threads.
Parent - By Barnard (Bronze) Date 2012-05-13 17:50
4 cores my friend,4 cores...isnt the same than 4 threads
Parent - - By Ray (****) Date 2012-05-13 17:53

> I assume you are saying that the Rybka version that played in this tournament on 100 threads, would have the same results if it had played on 4 threads.


Yes. Apart from Critter, the field was not that strong
Parent - By Barnard (Bronze) Date 2012-05-13 17:56
yes...apart from critter,were only patzers
Parent - By kgburcham (***) [us] Date 2012-05-13 18:13
Yes. Apart from Critter, the field was not that strong

Ok but I was saying that if Rybka was willing to give up its cores and enter with 4 cores its chances of a draw was probable as was Critter, thus diff results.
Parent - By Barnard (Bronze) Date 2012-05-13 17:56
well,i think that they thought rybka wouldnt have enough,and that is why they played with the cluster...if they thought they will have enough with a quad,they will played with a quad...

a good form to asure the win is play with a too much better hardware than your opponents;for my point of view,that only shows that you not truly trust in your engine 100 %,and you only want to asure the win by other forms...legal?yes of course.honest with yourself?of course,not...if  you need suplement your engine with a 100 core cluster,is that you internally know that your engine needs that ''extra help'' to win against other engines running on a too much weaker hardware and you with that ''extra help'' only want to asure what your engine cant obtain alone
Parent - - By Lukas Cimiotti (Bronze) [de] Date 2012-05-13 18:07

>that never at past (just 3 years ago) no one person can imagine that an engine can draw a game against rybka


As usual you know nothing but make a lot of noise. To find out I'm right you might follow this link: http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?tid=18957;pg=1. But probably you'll find one more lame excuse for your ignorance.
Parent - - By Barnard (Bronze) Date 2012-05-13 20:55
again,you are answering me only what is your interest

my first question that i asked you remain not asnwered:

do you think that rybka fake her kn/s calculation,and why?

maybe im an ignorant,in fact,if a smart guys says that,im sure an ignorant,but im not a bad-mannered person like you
Parent - - By Lukas Cimiotti (Bronze) [de] Date 2012-05-14 06:19

>do you think that rybka fake her kn/s calculation,and why?


I've answered this alrerady, but it looks like you forgot my answer: http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?pid=402442

>im not a bad-mannered person like you


:lol:
I'll put you on my ignore list - I hope Uly will check your posts in the future.
Parent - By Uly (Gold) [mx] Date 2012-05-14 07:26

> I hope Uly will check your posts in the future.


Barnard has some preconceived ideas about some issues, so I'm no longer going to answer his questions.
Parent - By Barnard (Bronze) Date 2012-05-14 11:53
nice to know that,from now,too much easy task for me to know what mod moderate my posts...

so thanks for adding me to your ignore list :lol:
Parent - By Highendman (****) Date 2012-05-13 21:57
+1
Parent - - By Carl Bicknell (*****) [es] Date 2012-05-14 04:53 Edited 2012-05-14 04:57
This isn't right. It drew ONE game. Even if the cluster is 300 elo stronger the odd draw is to be expected.

You think just because the cluster draws one game it has been surpassed?
Parent - By Ray (****) Date 2012-05-14 05:45

> This isn't right. It drew ONE game. Even if the cluster is 300 elo stronger the odd draw is to be expected.


Exactly.
Parent - By Ray (****) Date 2012-05-14 10:05

> You think just because the cluster draws one game it has been surpassed?


No mention of that fact that Critter had 3 draws in the tournament....
Parent - - By Barnard (Bronze) Date 2012-05-14 12:05
of course,not

if one player is 300 elo points stronger than other,the lower player expect make some draws against the strong player

i just point that if an engine plays in a lets say an Intel 32 core machine (like Houdini is able to play),cluster wont be more than about 80-100 elo stronger,not more,and that with the 296 cores

and if Houdini 3 is released and is 40 elo points stronger than Houdini 2,cluster wont be more than 40-60 elo points than the Intel 32 core computer running Houdini 3

and at least for me,that means that Houdini averpassed and a lot rybka,because running on a weaker coputer,is near the 'power' of the cluster

and the lists like ipon,ccrl,etc,shows that im right,Houdini overpassed rybka in near 60-70 points,Houdini 3 will be more than 100 elo points stronger than Rybka

so in general terms,i think that if i can have similar (not equal,but similar) strength in a 32 core machine than in a 296 core machine,the cluster isnt giving the power for that it was designed because the engined was clearly overpassed
Parent - - By Harvey Williamson (*****) Date 2012-05-14 12:21
You should also factor in that a good book maker like Jiri or Jeroen can add another 50 elo.
Up Topic Rybka Support & Discussion / Rybka Discussion / Leiden Round 7: Rybka vs Critter
1 2 3 Previous Next  

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill