Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Rybka Discussion / Eros Riccio wins against the cluster
1 2 Previous Next  
- - By Lukas Cimiotti (Bronze) [de] Date 2009-07-13 22:21
[Event "Friendly Game, 60m + 30s"]
[Site "Engine Room"]
[Date "2009.07.13"]
[Round "?"]
[White "Auryn"]
[Black "Rechenschieber, Rybka 3g55 cluster "]
[Result "1-0"]
[ECO "B12"]
[WhiteElo "1843"]
[Annotator "0.00;0.58"]
[PlyCount "151"]
[EventDate "2009.07.13"]
[TimeControl "3600+30"]

{My book 15.1.ctg, 2048 MB} 1. e4 {[%emt 0:00:02]} c6 {[%eval 0,0] [%emt 0:00:
00]} 2. d4 {[%emt 0:00:06]} d5 {[%eval 0,0] [%emt 0:00:00]} 3. e5 {[%emt 0:00:
13]} Bf5 {[%eval 0,0] [%emt 0:00:00]} 4. Nf3 {[%emt 0:00:34]} e6 {[%eval 0,0]
[%emt 0:00:00]} 5. Be2 {[%emt 0:00:02]} Nd7 {[%eval 0,0] [%emt 0:00:00]} 6. O-O
{[%emt 0:00:04]} Ne7 {[%eval 0,0] [%emt 0:00:00]} 7. c3 {[%emt 0:00:03]} h6 {
[%eval 0,0] [%emt 0:00:00]} 8. Na3 {[%emt 0:00:39]} c5 {[%eval 0,0] [%emt 0:00:
00]} 9. dxc5 {[%emt 0:03:20]} Nxc5 {[%eval 0,0] [%emt 0:00:00]} 10. Nb5 {[%emt
0:00:23]} Nc6 {[%eval 0,0] [%emt 0:00:00]} 11. Be3 {[%emt 0:00:18]} Be7 {
[%eval 0,0] [%emt 0:00:00]} 12. Nfd4 {[%emt 0:00:12]} Bg6 {[%eval 0,0] [%emt 0:
00:00]} 13. b4 {[%emt 0:00:09]} Nd7 {[%eval 0,0] [%emt 0:00:00]} 14. Nd6+ {
[%emt 0:00:09]} Bxd6 {[%eval 0,0] [%emt 0:00:00]} 15. exd6 {[%emt 0:00:00]} O-O
{[%eval 0,0] [%emt 0:00:00]} 16. Nxc6 {[%emt 0:00:14]} bxc6 {[%eval 0,0] [%emt
0:00:00]} 17. Qa4 {[%emt 0:00:01]} Nb6 {[%eval 58,21] [%emt 0:04:21]} 18. Qxc6
{[%emt 0:00:44]} Rc8 {[%eval 59,21] [%emt 0:00:49]} 19. Qb7 {[%emt 0:00:10]}
Na4 {[%eval 60,21] [%emt 0:04:07]} 20. Qxa7 {[%emt 0:00:16]} Nxc3 {[%eval 63,
21] [%emt 0:00:38]} 21. Rfe1 {[%emt 0:00:09]} Ra8 {[%eval 62,22] [%emt 0:04:38]
} 22. Qe7 {[%emt 0:03:13]} Qb8 {[%eval 61,22] [%emt 0:00:00]} 23. Bc5 {[%emt 0:
00:10]} Nxe2+ {[%eval 76,21] [%emt 0:04:29]} 24. Rxe2 {[%emt 0:00:15]} Qb5 {
[%eval 78,22] [%emt 0:03:43]} 25. Ree1 {[%emt 0:00:50]} Qa4 {[%eval 78,22]
[%emt 0:02:50]} 26. Qh4 {[%emt 0:03:37]} Bf5 {[%eval 77,21] [%emt 0:00:00]} 27.
Re3 {[%emt 0:01:07] (h3)} Rfe8 {[%eval 82,21] [%emt 0:04:48]} 28. Rg3 {[%emt 0:
01:31] (Qd4)} Kh7 {[%eval 77,21] [%emt 0:02:19]} 29. Qf4 {[%emt 0:04:41] (Qd4)}
Rec8 {[%eval 74,21] [%emt 0:03:26]} 30. Qd2 {[%emt 0:03:27] (h3)} Qd7 {[%eval
80,22] [%emt 0:04:30]} 31. a3 {[%emt 0:00:07] (Qb2)} f6 {[%eval 76,21] [%emt 0:
03:05]} 32. Rc1 {[%emt 0:04:42] (h3)} Be4 {[%eval 75,21] [%emt 0:05:15]} 33.
Rgc3 {[%emt 0:02:39] (Re3)} Rcb8 {[%eval 77,21] [%emt 0:02:52]} 34. Re3 {[%emt
0:01:22] (h3)} Bg6 {[%eval 80,21] [%emt 0:03:40]} 35. h3 {[%emt 0:00:57]} Rc8 {
[%eval 86,20] [%emt 0:02:33]} 36. Rce1 {[%emt 0:01:21] (Qb2)} Re8 {[%eval 81,
21] [%emt 0:02:23]} 37. Qd1 {[%emt 0:02:35]} Bf7 {[%eval 85,20] [%emt 0:00:00]}
38. Qd3+ {[%emt 0:00:47] (f4)} Kg8 {[%eval 82,21] [%emt 0:02:44]} 39. f4 {
[%emt 0:01:17]} Kh8 {[%eval 87,20] [%emt 0:00:54]} 40. g4 {[%emt 0:03:56] (Qb3)
} Bg8 {[%eval 117,20] [%emt 0:02:15]} 41. Kh2 {[%emt 0:06:10] (g5)} Rf8 {
[%eval 117,18] [%emt 0:01:21]} 42. Rg3 {[%emt 0:00:31]} Rae8 {[%eval 162,18]
[%emt 0:02:13]} 43. g5 {[%emt 0:01:11]} Bh7 {[%eval 179,17] [%emt 0:00:00]} 44.
Qa6 {[%emt 0:02:45]} fxg5 {[%eval 188,19] [%emt 0:00:00]} 45. fxg5 {[%emt 0:00:
12]} hxg5 {[%eval 216,18] [%emt 0:00:53]} 46. Qa7 {[%emt 0:00:16]} Qxa7 {
[%eval 220,18] [%emt 0:00:13]} 47. Bxa7 {[%emt 0:00:00]} Bf5 {[%eval 220,19]
[%emt 0:01:13]} 48. Bb6 {[%emt 0:01:06] (b5)} e5 {[%eval 228,20] [%emt 0:01:10]
} 49. Rxg5 {[%emt 0:01:02]} Bd7 {[%eval 236,22] [%emt 0:00:00]} 50. Rgxe5 {
[%emt 0:00:28]} Rxe5 {[%eval 258,24] [%emt 0:00:41]} 51. Rxe5 {[%emt 0:00:13]}
Kg8 {[%eval 258,22] [%emt 0:00:07]} 52. Bc5 {[%emt 0:01:02] (Kg3)} Re8 {[%eval
237,22] [%emt 0:01:15]} 53. Rxd5 {[%emt 0:00:22]} Kf7 {[%eval 243,22] [%emt 0:
00:31]} 54. Kg3 {[%emt 0:00:13]} Bc6 {[%eval 259,22] [%emt 0:01:51]} 55. Rd1 {
[%emt 0:01:55] (Rf5+)} Ra8 {[%eval 266,22] [%emt 0:01:08]} 56. Rd3 {[%emt 0:00:
45]} g6 {[%eval 266,23] [%emt 0:00:30]} 57. Rb3 {[%emt 0:00:56] (h4)} Rb8 {
[%eval 288,22] [%emt 0:01:25]} 58. Kf4 {[%emt 0:00:41]} Bd7 {[%eval 289,21]
[%emt 0:00:04]} 59. Rb2 {[%emt 0:02:33] (Rb1)} Kf6 {[%eval 294,20] [%emt 0:00:
34]} 60. h4 {[%emt 0:00:41]} Bb5 {[%eval 307,21] [%emt 0:00:00]} 61. Ra2 {
[%emt 0:00:37] (Rb1)} Ra8 {[%eval 304,21] [%emt 0:00:19]} 62. Ra1 {[%emt 0:00:
20]} Ba4 {[%eval 326,22] [%emt 0:00:24]} 63. Re1 {[%emt 0:00:26] (Rb1)} Re8 {
[%eval 318,22] [%emt 0:00:13]} 64. Rxe8 {[%emt 0:02:52] (Rb1)} Bxe8 {[%eval
238,23] [%emt 0:00:06]} 65. Bb6 {[%emt 0:00:00] (Ba7)} Ke6 {[%eval 353,25]
[%emt 0:00:43]} 66. Bc7 {[%emt 0:00:08]} Bc6 {[%eval 418,24] [%emt 0:00:24]}
67. Bb8 {[%emt 0:00:06] (Kg5)} Be8 {[%eval 488,23] [%emt 0:01:03]} 68. Ke4 {
[%emt 0:00:20]} Kf7 {[%eval 487,22] [%emt 0:00:09]} 69. Kd5 {[%emt 0:00:13]
(Kd4)} Bb5 {[%eval 511,18] [%emt 0:00:51]} 70. Kc5 {[%emt 0:00:09]} Bd7 {
[%eval 520,14] [%emt 0:00:02]} 71. b5 {[%emt 0:00:05] (Bc7)} Bc8 {[%eval 651,
14] [%emt 0:00:43]} 72. Kc6 {[%emt 0:00:07] (a4)} Ke8 {[%eval 1529,14] [%emt 0:
00:50]} 73. b6 {[%emt 0:00:10]} Bf5 {[%eval 1776,12] [%emt 0:00:19]} 74. Kc7 {
[%emt 0:00:13] (b7)} Kf7 {[%eval -28831,15] [%emt 0:00:50]} 75. d7 {[%emt 0:00:
06]} Bd3 {[%eval 2529,14] [%emt 0:00:04]} 76. d8=Q {[%emt 0:00:21]
Rechenschieber,Rybka 3g55 cluster A resigns} 1-0

Contempt was 15

I guess this was a book win ;)
Parent - - By Ray (****) Date 2009-07-13 22:44
Didn't take long for the cluster losses to be posted on the Hiarcs forum...
Parent - By Carl Bicknell (*****) [gb] Date 2009-07-13 22:51
Yes there's a certain unfortunate delight when it is beaten and yet they don't want to face it in the world championships. I could understand if it was HIARCS beating it, but it's not. Maybe they're trying to drum up enthusiasm for the centaur - cluster match...
Parent - - By yanquis1972 (****) [us] Date 2009-07-13 22:56
can we get some kind of schedule for cluster games up? even if it's rough that way we'll have some idea of when to tune in.
Parent - By Lukas Cimiotti (Bronze) [de] Date 2009-07-14 13:16
Dates for cluster games are in the first post of this thread.
http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?tid=11686
But if there is time for one or two more games when the annouced game is over and when a good centaur wants to play - then I play one more game.
Parent - - By Watchman (***) Date 2009-07-13 23:35

>I guess this was a book win


Bad opening ok... but certainly not a book win. :-P

>Yes there's a certain unfortunate delight...


What is unfortunate is attitudes likes these...

Maybe it's because there is some geniune/sincere interest in the top chess engine playing on a unique $30,000 "piece of hardware" if it can be beaten?  Ever stop to think of that?  Course not... it's always, "let's see how we can continue to insult Team Hiarcs."
Parent - - By Carl Bicknell (*****) [gb] Date 2009-07-13 23:42

> Maybe it's because there is some geniune/sincere interest in the top chess engine playing on a unique $30,000 "piece of hardware" if it can be beaten?  Ever stop to think of that?  Course not... it's always, "let's see how we can continue to insult Team Hiarcs."


your point of view is very one sided. Where is all the credit it deserves for winning when it does? Ever stop to think of THAT?
Parent - - By Watchman (***) Date 2009-07-14 00:30

>Where is all the credit it deserves for winning when it does? Ever stop to think of THAT?


Oh please... if I had a nickle for every time someone bloviated on RF about "the cluster" either winning or at the very least, drawing...  The expectation was it would win, otherwise, I would have the joy of saying, “Well, isn't that special?” everytime it did.

>It is great fun to insult Team Hiarcs


u u u... just u u u :-P
Parent - - By Carl Bicknell (*****) [gb] Date 2009-07-14 07:32

> Oh please... if I had a nickle for every time someone bloviated


bloviated. I have a pill for that.

I didn't actually realise it was Harvey who wrote the post on H.F until you pointed it out. But this is the RYBKA forum - of COURSE people are positive about Rybka, duh!

Parent - - By Watchman (***) Date 2009-07-14 13:19

>bloviated. I have a pill for that.


Then it would be very kind of you to share with many of your fellow posters here at RF.

>I didn't actually realise it was Harvey who wrote the post on H.F until you pointed it out.


I pointed it out??? Where did I do that btw?  Also btw, better to look for yourself and see who is doing the actual posting, than assume it is "so-and-so" imo.
Parent - - By Carl Bicknell (*****) [gb] Date 2009-07-15 15:35

> I pointed it out??? Where did I do that btw?


here: "let's see how we can continue to insult Team Hiarcs." I knew from the post above mine, which was the one I replied to, that A person on the hiarcs forum had posted the rybka cluster game, but didn't know who, because I didn't bother to check on this occasion. Then you said it was a member of the hiarcs team.

> Also btw, better to look for yourself and see who is doing the actual posting, than assume it is "so-and-so" imo.


I didn't assume anyone did it...
Parent - By Watchman (***) Date 2009-07-15 21:06
I don't mean to belabor this but...

>Then you said it was a member of the hiarcs team.


Right... then it could have been any number of people besides Harvey, including myself or Sebi Böhme or... and the list goes on...
Parent - By Banned for Life (Gold) Date 2009-07-13 23:49
It is great fun to insult Team Hiarcs, but does this have anything to do with Eros' game?
Parent - - By billyraybar (***) [us] Date 2009-07-13 23:52
Maybe not a book win, but It's hard to fathom how the black could come out of book with such a negative evaluation.
Parent - - By Uri Blass (*****) [il] Date 2009-07-14 00:50
I looked at some books that I have and did not find 8...c5
I could only see this move written with 0 games in mybook15.1.ctg

It is not in CompMasterBeta.ctg or in perfect15.ctg or in HS-masterbook.ctg or  in OM_master_book2.ctg or in Storm2.1.ctg

I wondered what is the reason that cluster rybka played as book move a move that seems to be a novelty based on all the books that I can find(mybook15.1 has 0 games and 0% probability).

If c5 is not a novelty then what is the reason that no book that I downloaded include moves after this move?
It is not only that this move is probably a bad move because even if it is a bad move I can expect a good book to include moves after it in case that it is not a novelty.

Uri
Parent - - By Kreuzfahrtschiff (***) [de] Date 2009-07-14 01:16
in pferdchen2 is 19. Qb5 the last move
Parent - - By Uri Blass (*****) [il] Date 2009-07-14 03:08
I do not know what is pferdchen2 but did some search in all the .ctg books that are installed in my computer and here is the result:

1)test.ctg does not include 1..c6 and only has 1...e5
2)TheTurk.ctg and OM_corr_2500.ctg do not have 7.c3 in the book
3)Junior9.ctg include 8...c5 but with 0 games and after 8...c5
9.Nb5 is the only book move
4)Mybook15.1.ctg also has 8...c5 with 0 games and has no moves after it
5)RybkaII.ctg,Rybka3.ctg,DeepFritz.ctg,Fritz8.ctg,compmasterbeta2.ctg,storm2.1.ctg,perfect15.ctg
HS-Masterbook.ctg,OM_masterbook2.ctg,Perfect2009.ctg,compmasterbeta3.ctg do not include 8...c5

I wonder which good book do I miss?
Parent - - By Jonas (****) [at] Date 2009-07-14 03:23
you miss tons of books... none of the books you mention (besides mybook 15.1 which is quite good and new) would actually perform well in a book tournament
Parent - - By Uri Blass (*****) [il] Date 2009-07-14 05:47
This is not correct
see the posts of Tom Casanovas
He claimed that compmasterbeta3 is the best book he tested.

here are his posts:

http://talkchess.com/forum/viewtopic.php?topic_view=threads&p=279833&t=28822
http://talkchess.com/forum/viewtopic.php?topic_view=threads&p=280032&t=28822

" I have tested a lot of books and at this moment I don't know one single book better than this. Amateur or Pro."

I can see names of many books that I do not know if they are free in
http://www.sedatchess.com/scct5.html

Note that the statistical error is too high to claim that mybook is better than compmasterbeta3

Uri
Parent - - By Jonas (****) [at] Date 2009-07-14 10:39 Edited 2010-01-24 21:04

> Note that the statistical error is too high to claim that mybook is better than compmasterbeta3


in sedats test yes, he tests with learning which can slightly destroy the results because it depends on the order of what books he test.

...15.0 is stronger and 15.1 is sure stronger... but both books arent really up to date and dont belong to the current top 10 public books

they are stronger because they are newer. it might be correct that it was one of the best some time ago (alltough i havent seen any book test where it perfomed kinda well). theres no need to test it it will perform worse.

> He claimed that compmasterbeta3 is the best book he tested.


where are his results and what books did he test? i can only see he still using perfect 15 which is much too old.

but a strong public book lasts maximum a month after the authors got enough games from playchess and they get destroyed automatically.
its totally different than with engines.. the improvement runs alot faster and to have the strongest book you have to be up to date. novelties are released monthly. the main reason is that rybka 3 is stronger than the gms in the past. and what we have in our books is a big chaos which we try to put in an order slowly

take a look at my tests imo they are slighlty more accurate because they are tested without learning, but the time control and the number of games is smaller. but more games arent needed if learning is off and would result in duplicates. and the result doesnt change much with higher time controls. important is just not using too less time and give an increment to avoid endgame errors.
the most important thing is that the result is more up to date

http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?tid=11455;pg=1

almost all important free books are listed here: [Deleted] in downloads section.

you can also try Jonas.ctg: http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?tid=11614
it became first in my last bullet test http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?tid=11750
note that this is just an experiment on different testing conditions.

however you can test any of the public books i used under any equal conditions and i promise they will perform much better than compmaster 3.0 beta
Parent - - By Uri Blass (*****) [il] Date 2009-07-14 11:01
Thanks for the links

I will try to download them later
I am interested in downloading more books not for testing engine-engine games but because maybe the data may be useful for my correspondence games that part of them are still in the theory phase.

I have my doubts about your claim:"
"and the result doesnt change much with higher time controls."

It does not make sense to me based on common sense because it is clearly possible to tune books for blitz.

If you take public moves X and add to them many moves that rybka play in blitz to generate book Y
then it is clear to me that Y is going to perform better than X at blitz thanks to time advantage when I do not expect Y to perform better than X at long time control.

Uri
Parent - By Bouddha (****) [ch] Date 2009-07-14 11:25
Well, what you see in book competition is tested on playchess at 3min time but behind that there is work with more search time and everytime the books have to but updated. So if you enter a "bluf" move, its killed very quickly.

My recommendation for a CC player would be to download the latest opening books, look at the openings but check before the game is played because opening books is like everything. Its never perfect but its better everyday.

Playchess books & games are updated daily if you want to stay higher than 2750 rated.

And this brings me to the 2nd part. Openings are even more important than what I thought. Evenso Cluster Rybka 4beta is much stronger than Rybka 3, Rybka 3 is able to win a good position against the cluster.

This WE, the cluster played in a 16min tournament against other engines (most of the Rybka 3) and didn't finish in the first 5.
It didn't lost right after the opening, but when Rybka 3 had a good position, it was just able to play convincely to win it even against the Rybka 4 cluster.

regards
Parent - - By Jonas (****) [at] Date 2009-07-14 12:37
well the main thing i wanted to say is that the development of books is a never ending learning process (something like trying to solve chess from the wrong side)
so an old book is always weak
Parent - - By Bouddha (****) [ch] Date 2009-07-14 12:40
Agreed.

Unless you found the truth in some lines....
Parent - - By Jonas (****) [at] Date 2009-07-14 14:54
yes, unless you find a move that wins for sure.. and that one wins with any time control. that move is then not played again because new books will make a different move erlier after enough games are lost with that move
Parent - By Bouddha (****) [ch] Date 2009-07-14 16:26
Updated books yes :-)
Parent - - By Uri Blass (*****) [il] Date 2009-07-14 13:49
I disagree and I think that it is dependent on the time control.

If you can use few hours per moves(what happens in correspondence games) then it may be better to have a smaller book and give rybka to find hopefully superior move relative to the book moves.

I am interested in books only as a source of information what to analyze and not as something that will give me the truth because it is possible that books have some good moves that rybka cannot find even after hours of search and analyzing with rybka for hours after the book moves is going to help to find a better move.
Parent - By Jonas (****) [at] Date 2009-07-14 14:52
you are right with that.. it depends on the time control... i just said the results doesnt differ that much.
Parent - - By Jonas (****) [at] Date 2009-07-14 22:57
ntw i tried about 10times to get an account on talkchess.com but i just dont get the activation email by a mod. what can i do?
Parent - - By Vempele (Silver) [fi] Date 2009-07-14 23:05
Never got an activation email myself. Just try to log in.
Parent - - By Jonas (****) [at] Date 2009-07-14 23:12
doesnt work..
Parent - - By Jonas (****) [at] Date 2009-07-14 23:13
could someone create an account called DeepJonasMP or something like that for me and give me the password so i can change it...
Parent - By Jonas (****) [at] Date 2009-07-14 23:14
if any of you can, please write a message so i can give you a nickname to create
Parent - - By Vempele (Silver) [fi] Date 2009-07-14 23:18
Unfortunately, changing an account's email address deactivates it.
Parent - - By Sunny (***) [us] Date 2009-07-15 02:46
I changed email address both on Rybka forum and Talkchess but never got deactivated. Both the forums again sent me an activation link. May be I was lucky enough :)
Parent - By Dadi Jonsson (Silver) [is] Date 2009-07-15 09:22
That's how it is supposed to work here on the Rybka forum. The option is there to be used if you need to.
Parent - By Milton (***) [us] Date 2009-07-15 02:28

>ntw i tried about 10times to get an account on talkchess.com but i just dont get the activation email by a mod. what can i do?


Hi,

I think the guy at talkchess you have to contact is Sam Hull.  Try sending him an email at:

tcadmin@verizon.net.

Regards,
Milton
Parent - By Jonas (****) [at] Date 2009-07-14 02:15
[Event "FIDE-Wch k.o. (Women)"]
[Site "New Delhi"]
[Date "2000.11.28"]
[Round "1.2"]
[White "Qin Kanying"]
[Black "Klinova, Masha"]
[Result "1/2-1/2"]
[ECO "B12"]
[PlyCount "88"]
[EventDate "2000.11.27"]
[EventType "k.o."]
[EventRounds "6"]
[EventCountry "IND"]
[Source "ChessBase"]
[SourceDate "2001.01.11"]

1. e4 c6 2. d4 d5 3. e5 Bf5 4. Nf3 e6 5. Be2 Ne7 6. O-O h6 7. c3 Nd7 8. Na3 c5
9. Nc2 Nc6 10. Nce1 Be7 11. Bd3 Bxd3 12. Nxd3 cxd4 13. cxd4 Qb6 14. Be3 a5 15.
Qd2 a4 16. Rab1 Qa5 17. Qd1 Nb6 18. b3 O-O 19. Nf4 axb3 20. axb3 Nb4 21. Nh5
Qb5 22. Ne1 Kh7 23. Qg4 Rg8 24. Qh3 Nd7 25. f4 f5 26. exf6 Nxf6 27. Nf3 Nxh5
28. Qxh5 Qd3 29. Bd2 Nc6 30. Qh3 Qg6 31. Rbe1 Qf5 32. g4 Qf7 33. Kh1 Ra2 34.
Be3 Rb2 35. Ng5+ Bxg5 36. fxg5 Qg6 37. gxh6 gxh6 38. Bc1 Rc2 39. Rf7+ Rg7 40.
Rf4 Nb4 41. Rf3 Nc6 42. Rf4 Nb4 43. Rf3 Nc6 44. Rf4 Nb4 1/2-1/2

[Event "Ðåéòèíãîâàÿ ïàðòèÿ, 3m + 0s"]
[Site "Engine Room"]
[Date "2009.06.30"]
[Round "?"]
[White "Ultima, Rybka 3 960"]
[Black "ÑåðãåéÊîæåâíèêîâ, Rybka 3"]
[Result "0-1"]
[ECO "B12"]
[Annotator "0.62;0.01"]
[PlyCount "104"]
[EventDate "2009.02.01"]
[TimeControl "180"]

{Rybka 3 960 (2 CPUs): 14.7 plies; 102kN/s AMD Athlon(tm) 64 X2 Dual Core
Processor 4800+ 2400MHz, My book 15.ctg, 64 MB} 1. e4 {0.00/0 0} c6 {0.00/0 0}
2. d4 {0.00/0 0} d5 {0.00/0 0} 3. e5 {0.00/0 0} Bf5 {0.00/0 0} 4. Nf3 {0.00/0 0
} e6 {0.00/0 0} 5. Be2 {0.00/0 0} Nd7 {0.00/0 0} 6. O-O {0.00/0 0} h6 {0.00/0 0
} 7. c3 {0.00/0 0} Ne7 {0.00/0 0} 8. Na3 {0.00/0 0} c5 {0.00/0 0} 9. dxc5 {
0.62/11 3} Nxc5 {(Nc6) 0.00/0 0} 10. Nb5 {0.63/10 2} Nc6 {0.00/0 0} 11. Nbd4 {
0.59/12 7} Bh7 {(Bg6) 0.01/12 9} 12. Nxc6 {(Be3) 0.00/0 0} bxc6 {0.02/13 2} 13.
b4 {(Be3) 0.00/0 0} Nd7 {0.03/13 5} 14. Qa4 {0.39/10 5} a5 {-0.08/13 0} 15. b5
{0.14/11 7} c5 {-0.08/15 0} 16. Be3 {(Qf4) 0.14/11 4} g5 {-0.12/11 2} 17. Rfe1
{(Rad1) 0.12/12 9} Bg7 {-0.30/12 4} 18. Rad1 {0.13/12 0} c4 {(0-0) -0.39/12 3}
19. Qa3 {(Bd4) 0.03/9 5} Qe7 {(Rb8) -0.48/11 2} 20. Qb2 {(Qxe7+) -0.23/11 8}
O-O {-0.52/12 3} 21. h4 {(a4) -0.33/11 0} g4 {-1.00/12 2} 22. Nd4 {
(Nh2) -0.52/11 5} Qxh4 {-1.02/12 3} 23. Bf4 {-0.73/12 1} Bxe5 {-1.22/14 2} 24.
Qd2 {(Qc1) -0.72/12 1} Bxf4 {-1.29/14 3} 25. Qxf4 {-0.88/14 0} e5 {-1.29/16 14}
26. Nf5 {(Qxg4+) -1.23/14 3} exf4 {-1.29/13 0} 27. Nxh4 {-1.21/15 5} Nf6 {
-1.82/15 2} 28. Rd4 {-1.21/14 15} Rfe8 {(g3) -300.02/15 2} 29. Rxf4 {-1.30/12 3
} Kg7 {-2.12/15 9} 30. Kf1 {(g3) -1.30/13 5} Re5 {-2.37/13 5} 31. g3 {
-2.22/11 0} Rb8 {(h5) -2.32/14 2} 32. a4 {-1.95/10 1} h5 {-2.20/14 2} 33. Rd4 {
-1.95/12 0} Rbe8 {-2.20/15 4} 34. Ng2 {-1.95/12 9} Ne4 {-2.20/16 0} 35. Rc1 {
-2.67/12 3} Nxf2 {-300.02/15 11} 36. Nf4 {(Kxf2) -2.94/14 0} Ne4 {-2.21/14 0}
37. Kg2 {-2.94/13 11} Nc5 {-3.22/14 0} 38. Bd1 {(Kf2) -3.24/13 4} Nb3 {
(Kh6) -4.21/12 12} 39. Bxb3 {-3.24/10 2} cxb3 {-4.14/14 0} 40. Rd2 {-3.56/12 4}
Re3 {-300.02/15 7} 41. Kh2 {-3.98/12 0} Be4 {-300.02/14 10} 42. b6 {
(Ng2) -3.98/13 0} h4 {(Rb8) -300.02/13 9} 43. gxh4 {-4.83/11 4} Rh8 {
(Rb8) -300.02/13 0} 44. Kg1 {-4.60/9 1} Rg3+ {(Rxh4) -5.53/11 1} 45. Ng2 {
-4.65/9 7} Rb8 {(Rxh4) -5.93/13 0} 46. Rb2 {-5.39/9 3} Rxb6 {-6.11/12 0} 47.
Kh2 {-6.03/10 6} Rh3+ {-6.46/12 0} 48. Kg1 {-5.78/8 0} f5 {-6.42/10 1} 49. Rf1
{(Re1) -6.23/9 1} Rxc3 {-7.26/10 1} 50. Nf4 {-7.15/10 1} d4 {(Rb4) -7.77/10 1}
51. Rff2 {(Rd1) -7.39/9 4} Rb4 {(Rc4) -7.98/9 1} 52. Rf1 {(Rfd2) -7.63/8 2}
Rxa4 {Ultima,Rybka 3 960 abbandona -7.98/9 1} 0-1
Parent - By Bouddha (****) [ch] Date 2009-07-14 08:54
Because the books you mention are highly optimized for playchess server and based on these games.

I didn't check but e.g. I would not be surprised if most of these books dont have line for the King gambit as there is no reason to include them if your book doesn't play 1...e5 on 1.e4

The same applies to this CK line.
Parent - - By keoki010 (Silver) [us] Date 2009-07-14 14:11
Look in Jeroens book.  8...c5 shows 166 games and success of 45%.
Parent - By Nelson Hernandez (Gold) [us] Date 2009-07-14 23:35
That's interesting.  Clearly he has extensively tested that line with private games.
Parent - - By ernest (****) [fr] Date 2009-07-14 23:57

> Look in Jeroens book.  8...c5 shows...


??? strange! My (original) Aquarium Jeroen_book shows no 8...c5
Parent - - By keoki010 (Silver) [us] Date 2009-07-15 13:14
I've modified mine slightly. Maybe that's where it comes from?
Parent - - By ernest (****) [fr] Date 2009-07-15 17:15
Then you should know where your 8...c5 comes from     :evil:   ... not from Jeroen!
Parent - By keoki010 (Silver) [us] Date 2009-07-15 18:14
You know I went back and looked and it's not there!  Also I can't find it in any of my other trees except for "WideBook" where it shows 4 games from d2m. I've forgotten what I was looking at when I posted the reply.  I'll have to search around on my computer and rediscover what I did.
Parent - By Nelson Hernandez (Gold) [us] Date 2009-07-14 23:31
Definitely a book win.  Eros was in book through white's 21st move, in a position that has always (100%) won before.  At that point, according to the cluster, white was +0.62.  You don't give Eros Riccio that kind of an opening advantage and survive.  Not very often.
Parent - - By Vasik Rajlich (Silver) [pl] Date 2009-07-14 06:20
Congrats to Eros. Bloody day yesterday.

In these ultra-sharp lines like the 3. e5 Caro-Kann, testing Eros' book is pretty dangerous. This is another problem without a clear solution. My Rybka doesn't like the variation almost from the beginning - 8. .. c5 is a big score drop, although the move looks logical and if it's backed up by good preparation I can see it working. The book we used yesterday is (AFAIK) decent and completely new.

White's technique was also good. Black had a light-squared blockade and white had to play accurately to take advantage of his chances.

Vas
Parent - - By mikal (**) [us] Date 2009-07-14 06:57
That is unforgivable you know. That cluster should be more than 3600 elo now and should not even lose a single game. Heck, rybka alone non-cluster should not even lose a single game versus these centaurs in such a very short set of games, if you were to make the same improvement from 2.3.2a to version 3. I hope not, but I am smelling a failure. I bet you gonna delay it and release it like november and december or maybe 2010 and not august like last year because you just fail at some parts of it :-)

Maybe you should get tortured for every game the cluster loses, so you get motivated more.
Parent - By Bothkill (*) [ro] Date 2009-07-14 07:51
LOL
Parent - - By Uri Blass (*****) [il] Date 2009-07-14 08:39
It seems that the cluster is hanicapped by a book.
8...c5 is a positional blunder based on rybka.

My opinion is that the book for the cluster should be different.
There should be 2 books.
1)A small book when every move in it was analyzed earlier by the cluster for significant time to verify that it is not a blunder(or alternatively by rybka3 for a long time)

Moves in the small book do not have to be best but the difference in the evaluation should be less than 0.1 pawns in most cases.
There may be possible exceptions too the 0.1 pawn rule but humans need to check them carefully before allowing the cluster to play them to see that the cluster does not perform worse with the book move relative to games without the book move.

2)A Bigger book when the cluster did not check the moves earlier.

Only moves in the small book should be played in 0 seconds.
moves in the big book should be analyzed but not played unless the cluster finds that they are best.

Uri
Up Topic Rybka Support & Discussion / Rybka Discussion / Eros Riccio wins against the cluster
1 2 Previous Next  

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill