Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Rybka Discussion / Let Father play.
1 2 Previous Next  
- - By h1a8 (***) Date 2008-01-07 11:03
Why not organize a match with Father to play Rybka?
Father may do this either for free or a small amount of money.
The match can either be straight up and or draw odds (only when rybka plays white).

It has been said that it takes Father several games just to get a good result against Rybka (a draw).
So if this is true then Rybka would most likely win the match between the two.

If Rybka wins then it would be quite an achievement against anti-computer play.
Parent - - By turbojuice1122 (Gold) Date 2008-01-07 12:18
But Joel used anti-computer play when he could in the match with the exception of only one game of which I can think the day before yesterday.  The problem is that this Rybka version didn't provide great opportunities for Father methods except for game 4, which Joel exploited beautifully and easily--game 4 showed, I think, that Joel is more adept at using Father-like methods than is Father.

Nonetheless, I would like to see a match between the Father and Rybka under the same conditions--"just to be sure".
Parent - By Permanent Brain (*****) Date 2008-01-07 13:11 Edited 2008-01-07 13:14
Father (Pablo Restrepo) is a specialist of anticomputer blockade strategy under Fritz/Playchess conditions and at small time controls like 3m+0s. That means:

1. The interface automatically declares (imposes) draw if a position occured for the 3rd time, or if the 50 move rule is claimable. The human does not need to spend time on checking for repetitions and for claiming a draw (which may be complicated because not just the same two pieces are shuffled but maybe three or more over different squares, on each side).

2. Sometimes, engines will lose on time due to the x+0 time control, in 'desparation'.

3. The small time controls will cause that the engines recognise the danger of an inavoidable repetition or 50 move draw later, than with long time controls.

Nevertheless, a match against Rybka could be interesting but not with conditions identical to the Benjamin match. - I think it could be done via Internet, because due to his unusual and unique style, there is no supervision required for Father (in the sense of using an engine, because engines are not capable to play like that).

He is an anticomputer hero and had remarkable successes against almost all engines (typically some days after new releases even), nevertheless we need to keep in mind that the majority of his anticomputer games are losses. So it would not make sense to make match rules which expect 50% or more from him, even if draws count like wins. Maybe medium time controls should be considered, 4 or 6 games per day, and he wins if the gets something like 3 draws from 12 or 18 games.

1.5/18 against 3100 Elo would be a ~2710 performance.
Parent - - By ernest (****) Date 2008-01-07 13:15
You are the only ones not to know that "Father" plays tens (hundreds?) of games against Rybka before he gets Rybka to fall in one of his tricks.
"Father" will never accept to be ridiculed by accepting a public match against Rybka.
Parent - By Permanent Brain (*****) Date 2008-01-07 13:29
We would see if you'r right, if a decent offer including detailed conditions and prize money would be made.

I say he would play, because he certainly is NOT somebody afraid of losing. He also doesn't win the majority of games he plays in public, on Playchess. But he enjoys the draws and wins he DOES achieve, against computer opponents rated hundreds of points higher.
Parent - By turbojuice1122 (Gold) Date 2008-01-07 14:50
Typically he gets about one draw every 10 games on average.  You make strange assumptions about what I know :-).  I have already mentioned this before...
Parent - By Uri Blass (*****) Date 2008-01-07 13:33
Father has no chances under similiar conditions.
Maybe he can get draws against commercial rybka but benjamin did not play against default rybka.

Uri
Parent - By ebutaljib (****) Date 2008-01-07 13:36
I believe Father only plays blitz/bullet and for every game that he draws (or wins on time) he loses at least 5 or 6 times ( I have watched him on playchess). So he would have no chances.
Parent - - By lkaufman (*****) Date 2008-01-07 16:17
I will be glad to operate Rybka in an internet match against Father at draw odds (or draw and white odds) at some fast time control with a small increment, like 4'+2" for example. It could be either private or public, I don't care. My interest would simply be to see if we have truly resolved the problem of foiling such anti-computer tactics with our latest changes and contempt factor. Of course there is no sponsorship for a blitz match with an untitled player, but if he draws over half the games I'll send him $50 myself.
Parent - - By turbojuice1122 (Gold) Date 2008-01-07 16:39
Come to think of it, the player known as Maslakkostia might be a good choice for this, too--he is an IM also, and has experience in using the same tactics that Father uses against computers.
Parent - By Vempele (Silver) Date 2008-01-07 18:20
Also known as Konstantin Maslak :)
Parent - - By Nelson Hernandez (Gold) Date 2008-01-07 17:15
There are several thousand Father games in the public domain that can be had for the asking.  His opening repertoire and general pattern is predictable and thus should be fairly easy to frustrate.  It would only require Rybka playing some strange-looking opening moves to break up his entire plan.  I think such a contest would be practically over before it began.
Parent - - By JhorAVi (***) Date 2008-01-07 17:59
An 8 game match between Father and Rybka wherein father needs to DRAW only ONE game inorder to win the match... that would be fair concidering fathers inferior strength :)
Parent - - By Graham Banks (*****) Date 2008-01-07 18:30
What a joke.
Drawing one game gains you a match win.   :-)
Parent - - By Dr.Wael Deeb (***) Date 2008-01-07 19:48
Don't worry Graham,he won't even get this one draw,believe me :-)
Parent - By Nelson Hernandez (Gold) Date 2008-01-07 20:38 Edited 2008-01-07 20:41
Yeah, it would be like sticking Vas somewhere in empty Redskins stadium, marching blindfolded Pablo out to mid-field with a handgun, and telling him "you have eight shots to hit your target from this location".  With Vas free to move around in reaction to where he was pointing.
Parent - - By Gaмßito (****) Date 2008-01-07 21:37 Edited 2008-01-07 21:54
Yes, it doesn't have much sense to play against an untitled player. The player has to be GM. 

I told to Larry that give a try to GM Dzindzi - the most experienced player against computers that I have ever seen-.    
   
Dzindzi loved to play against machines and he spent many many hours on ICC playing only against computers. He played many hundred of games on ICC some years ago, against the best programs at that moment, and he won more than 150 and also draw a bit more than 250 games. He really knew what positions to play for this. He also got the most highest rating on this site (ICC) because of this, although, at the end he was accused to play against computers only for this purpose and he decided not to play anymore against computers in the site.
 
Many of his games were splendid, and Robert Hyatt can corroborate this; although, I don't know what has happened to Dzindzi in nowadays and I also do not know if he is now in quite good level or in good health to play against Rybka. I hope that he can be fine. And I also think that he does not play more in Tournaments. Does he already be retired? Anyway, I guess that an Odds draw match with him could be very dangerous for Rybka.

I think that Larry is the appropriate person to contact him now. It would be good to know Dzindzi opinion about this. If he is able to play, it would be quite good to give him an opportunity like this.

Regards,
Gaмßito.
Parent - - By turbojuice1122 (Gold) Date 2008-01-07 23:39
I strongly disagree about the need to play against a GM, though I certainly think that the prize money should be less for a non-titled player.  Things have changed greatly since the time of Dzindzi, and if he ever held any ratings records on the ICC, they have long been surpassed by very large amounts.  In this respect, Nakamura would, among the titled players on the ICC, have the greatest chances.
Parent - - By Banned for Life (Gold) Date 2008-01-08 00:55
No way. Kamsky would have the greatest chances and has already played many blitz games against Rybka running on very fast hardware on ICC.

Regards,
Alan
Parent - - By turbojuice1122 (Gold) Date 2008-01-08 01:42
I considered Kamsky when making my statement, but I think that Nakamura is better in this--he seems to have more luck against Rybka, at least circumstantially, and he also holds all of the records for highest human rating.
Parent - - By Banned for Life (Gold) Date 2008-01-08 01:51
I think you are being unduly influenced by his statement "My brain is better than Rybka six days a week"! :-) Too bad these types of matches can't really be set up. I don't usually get the chance to bet on such sure things.

Regards,
Alan
Parent - - By turbojuice1122 (Gold) Date 2008-01-08 01:58
Oh, no, I'm one who has often made fun of him for that statement :-).  Nonetheless, I have seen some of his games, and he seems to get lucky more often than Kamsky and actually has had a few nice wins.
Parent - - By Roland Rösler (****) Date 2008-01-08 03:05
Hey Turbo,
comparing Kamsky with Nakamura is real bullshit. Nakamura is a gambler and I´m sure he will never see the 2700 points, although he is so young. Okay, against Rybka you can roll the dice, but the dice will not help you!
Parent - By turbojuice1122 (Gold) Date 2008-01-08 04:26
This seems contradictory.  I recall from your previous posts that "real bullshit" is a good thing?
Parent - - By Gaмßito (****) Date 2008-01-08 04:42 Edited 2008-01-08 04:49
Naka is a wild player with a very sharp play, and I believe he has not chances to win the challenge against Rybka.
I think that we need a very cautious and solid GM. I mean, in the same line of Kramnik style.

Kamsky will be a very interesting option of course, despite that his style is sharp, but for the reason that he is playing at very high levels in nowadays, although who can pay the enormous quantities of money that he will request by the match? Perhaps Convekta?

A non titled player could play for very low prizes or even for free, but who really can give to Rybka the lunge? Be serious, I can't see a non-titled player having some chances against Rybka, but if you have some names, please put here.

Regards,
Gaмßito.
Parent - By turbojuice1122 (Gold) Date 2008-01-08 04:55
I would guess that Kamsky would end with a higher score, but Nakamura would end with more wins.  His style is very sharp, but it will occasionally have some luck and work, as has happened on occasion on the ICC.
Parent - - By josh Date 2008-01-08 04:52
lets see you do it Graham
Parent - By Graham Banks (*****) Date 2008-01-08 22:58
It's not my thing to play thousands of bullet games and then publish my handful of wins on time as masterpieces. Sorry.
Parent - - By lkaufman (*****) Date 2008-01-07 22:02
My only interest in such a match would be to see if we have made the PROGRAM immune to such drawing schemes. We have no control over what opening books people use with Rybka, so it won't serve any purpose for me to make a book that frustrates such play. You are surely right that this would be easy to do.
Parent - - By Nelson Hernandez (Gold) Date 2008-01-07 23:36
We differ.  As a proponent of the scientific method, you see this as an experimental process through which you gain new insights and incremental improvements can be made.  Call me Pattonesque but I believe the objective is always to defeat your opponent as quickly, decisively and overwhelmingly as possible.  Just imagine the mind-bending consternation in this forum if you faced Pablo in a contest and he scored even ONE draw.  I would be beside myself, as there would never be an end to his pidgin-English anti-computer triumphalism.
Parent - - By lkaufman (*****) Date 2008-01-07 23:45
Sure, but if I relied on a book to do this then when Rybka 3 came out he would be scoring draws against her if the book he happened to play against didn't have the right lines. I want the program to refuse to allow hopeless blockades unless behind. If we were playing some important match I would also make a book to be safe, but I wouldn't waste time on this for a blitz match with an untitled opponent.
Parent - - By Nelson Hernandez (Gold) Date 2008-01-07 23:58
Heh, beware of those untitled opponents, Larry.  :)  They're crazy enough to try anything.  They don't know any better.
Parent - By lkaufman (*****) Date 2008-01-08 00:16
Well, if you are hinting at yourself, your team has a title now, so I would give you plenty of respect in your field of freestyle chess.
Parent - - By Vasik Rajlich (Silver) Date 2008-01-08 15:01

> the objective is always to defeat your opponent as quickly, decisively and overwhelmingly as possible


The one real objective is to leave as little as possible to chance while expending minimal effort. Everything else is a means, not an end.

Vas
Parent - By Nelson Hernandez (Gold) Date 2008-01-08 15:26
Spoken like a true general. 
Parent - - By Richard Stickles (*) Date 2008-01-09 13:47
Pablo is most definatly a very great stone wall player, In playchess i run  Comp Fritz&Comp Shredder agasint Humans, I have played some great Gm's,Im's,Fm's etc.... But quite simply to beat the stone wall it only takes a very hi contempt factor, and some opening book tweaking,with over 5000 games played agasint humans the win ratio is only maybe 2% for humans, And let me say this, They we're very highly rated GM's that beat Comp Fritz....The Gm's do not try to play the stone wall,only the lower rated players try for the draw.....So when i am playing a high rated gm i will load a better book,As i have almost lost a few games versus GM's with my Anti Pablo book:-)

Pablo has a zero score versus Comp Fritz

If anyone's interested in the Comp Fritz games drop me a note i will send them no problem......

Regards
Richard

PS, Happy new year & Best luck to Rybka team in 2008
Parent - By Vasik Rajlich (Silver) Date 2008-01-10 23:00
Thanks Richard, have a nice 2008 yourself :)

Vas
Parent - - By Permanent Brain (*****) Date 2008-01-08 08:40
Father's strategy cannot be conquered with unusual opening moves.
Parent - - By Nelson Hernandez (Gold) Date 2008-01-08 13:04
As I have noted before, the key thing in a Pablo game is whether he successfully sets up a pawn blockade or not.  If not, he loses.  If he does, then an opposing engine enters the middle-game with uncertain chances.  So it seems to me the most critical period to frustrate his designs is the opening, and the most direct way to do that is to deviate from usual theory.  The reason Pablo is sometimes successful is because a) opening books do not make the necessary moves to put a crack in his pawn wall, and b) having failed to do so, they do not make the material sacrifices needed to breach his defense because these create significant evaluative disadvantages.

With what aspect of this analysis do you disagree? 
Parent - By billyraybar (***) Date 2008-01-08 13:51
I've watched Father play many many games against engines on Playchess.  Even when he does successfully set up a pawn blockade he loses the vast majority of the time.  There must be a luck factor involved.
Parent - - By Permanent Brain (*****) Date 2008-01-08 14:05
There are no predefined move sequences to prevent a blockade strategy. Father could play 1.a3, 2.h3 and still go for a pawn wall. For a book, you need variations. But there are zillions of (new!) move sequences possible, with nightmares of transpositions, if one side goes for a pawn wall instead of the typical "normal" opening play. A book cannot stop that.
Parent - By Nelson Hernandez (Gold) Date 2008-01-08 15:34
You make a very good point and I don't disagree that there are countless ways to set up a pawn wall, many of which have never been tried.  But in Pablo's specific case he does follow certain predictable moves quite often.  Humans are creatures of habit. 

I think you, Larry and Vas have adequately responded to my comments, though.  Ultimately Rybka development is not about opening book trickery EXCEPT in tournaments and matches, and for that they have Jeroen and Dagh.  R&D is about creating an engine that can respond to emerging or established blockade positions optimally when out of book.
Parent - - By josh Date 2008-01-08 04:50
I would just die if some of you ppl who post here actually know how to play chess period.
Parent - By Graham Banks (*****) Date 2008-01-08 08:13
It would be sad to see you end it all!   :-)
Parent - - By billyraybar (***) Date 2008-01-08 15:48
Huh.  I see many top notch chess players post in this forum.
Parent - - By josh Date 2008-01-08 20:09
Yeah right. Only a handful. Rest are computer jockeys that don't know squat except what Rybka prints out and how to configure thier system.Dare I say more?....
Parent - - By Graham Banks (*****) Date 2008-01-08 20:14
I doubt that anybody would have an interest in computer chess without being able to play chess.
How good they are at playing chess is irrelevant to this hobby, but there are a lot of very useful players here.
Parent - - By josh Date 2008-01-08 20:19
Well Graham , I wasn't naming names and i won't.( For all I know you yourself could be close to GM status OTB.)But at times , it gets deep in here IF you catch my drift.
Parent - By Graham Banks (*****) Date 2008-01-08 20:22
For all I know you yourself could be close to GM status OTB

I wish!   :-)

But at times , it gets deep in here IF you catch my drift.

Yes, I understand what you're alluding to.

Regards, Graham.
Parent - - By onursurme (***) Date 2008-01-08 21:54
I don't know how to play chess period.
What is chess period ?
The game which I know how to play is called "Chess", not "Chess period".
I'm serious.
Up Topic Rybka Support & Discussion / Rybka Discussion / Let Father play.
1 2 Previous Next  

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill