Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Rybka Discussion / Kramnik-Anand Advanced finished
1 2 3 4 Previous Next  
- - By turbojuice1122 (Gold) Date 2007-11-26 18:56
The Anand-Kramnik advanced chess match has finished, with both games draws.  I haven't really checked them out yet, since I currently have my computer running a CEGT gauntlet to get Shredder 8 onto the 40/20 list (after 140 games, it is currently about 10 elo points ahead of Shredder 9!).  It seems that these guys were playing in styles similar to their normal OTB styles, though.

There is more discussion on http://hiarcs.net/forums/viewtopic.php?t=601 from where I got these games from the poster Ted Summers.  In particular, it seems currently unknown what engines these guys were allowed to use.

[Event "Tal - Advanced Chess"]
[Site "?"]
[Date "2007.11.23"]
[Round "2"]
[White "GM Anand, Viswanathan(IND)"]
[Black "GM Kramnik, Vladimir(RUS)"]
[Result "1/2-1/2"]
[ECO "C67"]
[WhiteElo "2801"]
[BlackElo "2785"]
[Annotator "Ubiyca"]
[PlyCount "47"]

1. e4 e5 2. Nf3 Nc6 3. Bb5 Nf6 4. O-O Nxe4 5. d4 Nd6 6. Bxc6 dxc6 7. dxe5 Nf5
8. Qxd8+ Kxd8 9. Nc3 Ne7 10. h3 Ng6 11. Bg5+ Ke8 12. Rad1 Be6 13. Nd4 Bc4 14.
Rfe1 Bb4 15. Bd2 Rd8 16. a3 Bxc3 17. Bxc3 Nf4 18. Kh2 c5 19. Nf5 Rxd1 20. Rxd1
Ne6 21. f3 Bb5 22. Rd2 h5 23. Kg3 Bc6 24. Kf2 {
Kramnik - Anand Advanced match final score: 1-1} 1/2-1/2

[Event "Tal - Advanced Chess"]
[Site "Moscow"]
[Date "2007.11.23"]
[Round "1"]
[White "GM Kramnik, Vladimir(RUS)"]
[Black "GM Anand, Viswanathan(IND)"]
[Result "1/2-1/2"]
[ECO "D47"]
[WhiteElo "2785"]
[BlackElo "2801"]
[Annotator "Ubiyca"]
[PlyCount "93"]
[EventDate "2007.??.??"]
[TimeControl "1500+10"]

1. d4 Nf6 2. c4 e6 3. Nf3 d5 4. Nc3 c6 5. e3 Nbd7 6. Bd3 dxc4 7. Bxc4 b5 8. Bd3
Bb7 9. a3 Bd6 10. O-O O-O 11. Qc2 a5 12. e4 e5 13. dxe5 Nxe5 14. Nxe5 Bxe5 15.
h3 Re8 16. Be3 Qc7 17. Rad1 Rad8 18. Ne2 Bd6 19. Nd4 Nh5 20. Rfe1 Nf4 21. Bf1
Ne6 22. Nxe6 Rxe6 23. f4 f6 24. Qb3 Qf7 25. Bb6 a4 26. Qa2 Rd7 27. Be2 g6 28.
e5 Bc7 29. Bxc7 Rxc7 30. Bg4 f5 31. Bf3 Kg7 32. Rd8 Ree7 33. Qxf7+ Kxf7 34.
Red1 c5 35. Be2 c4 36. R1d6 Be4 37. Kf2 Re6 38. R8d7+ Rxd7 39. Rxd7+ Re7 40.
Rxe7+ Kxe7 41. g4 fxg4 42. hxg4 Bd5 43. Ke3 b4 44. Kd2 c3+ 45. bxc3 bxa3 46.
Kc1 Ke6 47. Kb1 1/2-1/2
Parent - - By Nelson Hernandez (Gold) Date 2007-11-26 21:52
Honestly, I believe Anand + top-end octo with any software he wanted would not stand a chance against any of the top five Freestylers in a straight-up non-exhibition match, if they were allowed to use anything they had in their arsenals.
Parent - - By JohnL (***) Date 2007-11-26 23:00
Are you serious? Of course Anand would be huge favourite. Most probably he uses Rybka daily for analysis and knows very well its weaknesses and strengths.
And the top 5 freestylers would cry if they saw Anands opening preparation :-)
Parent - - By Uly (Gold) Date 2007-11-26 23:12
Why doesn't Anand join Freestyle and beat them all?

Maybe he did, anonymously, and lost.
Parent - By lkaufman (*****) Date 2007-11-26 23:13
Because the prize isn't $100,000 or so.
Parent - - By Nelson Hernandez (Gold) Date 2007-11-27 01:23
Of course Anand would be a huge favorite.  And I suppose the subsequent upset would rank as one of the biggest of all time.

A lot depends on what help Anand was getting during the match.  If it is just him alone and one strong computer against any of the top five Freestyle teams, he loses.  That's my proposition.
Parent - - By billyraybar (***) Date 2007-11-28 17:14
Sometimes the current and past Freestyle champions seem so arrogant.  Why is this?  They lead everyone to believe that not even God could beat them in Freestyle chess.
Parent - - By Nelson Hernandez (Gold) Date 2007-11-28 19:12
I apologize for giving you that impression.  I have little to be arrogant about.  Nobody is invincible.
Parent - - By billyraybar (***) Date 2007-11-28 20:22
No problem, Nelson.  I'm not speaking of you in particular.  All the past champions seem to be a tad arrogant. I probably would be too. 
Parent - - By turbojuice1122 (Gold) Date 2007-11-28 20:28
It's easy to do--after all, each Freestyle tournament is the strongest chess tournament in history.
Parent - - By Uly (Gold) Date 2007-11-28 21:20

> It's easy to do--after all, each Freestyle tournament is the strongest chess tournament in history.


I still have problems understanding this. If the time control on Freestyle is several times shorter that Correspondence chess, from where the additional "strength" comes from?
Parent - By Banned for Life (Gold) Date 2007-11-28 21:47
High level correspondence games are significantly higher level than top freestyle games. This is not surprising given the huge disparity in time controls (1-1/2 minutes a move vs 4 days a move).

Alan
Parent - By turbojuice1122 (Gold) Date 2007-11-28 22:51
I think that you and Alan are correct, and that I'm wrong here--I had forgotten about that "small detail" :-).
Parent - - By stephenNJUKI (**) Date 2007-11-30 16:12
The difference between you and a top flight GM like Anand is the later understands a lot more about the development of pieces on the board. The fact that Anand rarely loses in open competition means he knows what he is doing..  so if you were to play him in freestyle, he would not be Rybka's slave..  He is a Grand Master. 
Parent - - By Banned for Life (Gold) Date 2007-11-30 16:27
There are a reasonable number of GMs playing in the freestyle events, although obviously none of them are near Anand's level. As a group, they have not done particularly well. Probably the major problem that many GM's have, is that they believe they should not be Rybka's slave because they are GMs. Nokamora summed it up best with his unforgettable quote: "My brain is better than Rybka 6 days a week".

I still believe that a GM who has trained to play centaur style games (not just perform computer assisted analysis) could have a significant advantage over a centaur with no appreciable OTB skills. However, this advantage is not nearly enough to compensate for a GM's tendency to overconfidence in his own ability relative to a top engine in most circumstances.

Regards,
Alan
Parent - By stephenNJUKI (**) Date 2007-11-30 17:11
Yeah perhaps there over confidence has led to poor results. You know Kramnik was quoted ' I only accept responsibility for the quality of my games, not the result. ' You see it seems these guys are at another level. A win is nice but first you need to grasp what you are doing. Movement and co-ordination of pieces is something the understand deeply as per their respective repertoires. Thats what I as a novice can speculate..
Parent - By Vasik Rajlich (Silver) Date 2007-12-02 12:34
Actually, I think it's obvious that good OTB skills are useful in a freestyle setting. It's like saying that 3-point range is useful in a basketball setting.

Vas
Parent - - By Nelson Hernandez (Gold) Date 2007-11-30 17:33
If we were "Rybka slaves" we wouldn't have much of a chance of winning Freestyle tournaments because other Rybka slaves with more hardware could routinely beat us.  Think about it, it's perfectly logical.

So there must be some value being added by the team.  What could it be?

1.  Our opening book is broadly designed to keep us out of trouble rather than finding/playing killer book lines.  It is defensive in strategic character though we do swiftly punish reckless ideas.  If we can exit book with an equal position against book experts we're not unhappy.  In competition it is far better to avoid mistakes than it is to have an arsenal of traps which may or may not get sprung.
2.  Our decision-making processes integrate more data in a more sophisticated way than most people.  This is our biggest strength in relative terms.  That's the reason we're happy with mere equality out of book.

We are not infallible, nobody is.  But we are certainly less fallible than Rybka playing alone and I daresay that's true for at least the top twenty or thirty Freestyle teams.  So...please abate the slavery remarks, none of us Freestyle aficionados merits such an accusation.
Parent - - By stephenNJUKI (**) Date 2007-11-30 17:55
In playing computer chess, There are different ranks of 'slaves'..  we are learning every day so some are more ' ' than others. Point is freestylers need Rybka's analysis to play. Anand can go some distance without it. Rybka is always calculating.. Where as GMs, though not infallible, understand what will eventually happen in key positions of their repertoire. Sorry for any offence caused.
Parent - - By stephenNJUKI (**) Date 2007-11-30 18:28 Edited 2007-11-30 19:01
In retrospect your argument of the sum being greater than the parts is valid. But presently we have GMs who play specific opening lines and ofcourse analyse them extensively. So to think you would have a field day against them would be a fallacy.   Looking at the whole picture, understanding movement of pieces and there co-ordination, is more important than winning.
Parent - - By turbojuice1122 (Gold) Date 2007-11-30 18:51
I think that his team's openings specifically tend to avoid lines that GMs have analyzed extensively.  Furthermore, if a good Freestyle team has a decent opening surprise, this puts the GM+computer at a significant disadvantage.
Parent - By stephenNJUKI (**) Date 2007-11-30 19:29
I think understanding and playing mainlines is a better exploit than playing novelties. Novelties may get you a favourable result..  especially under tight time controls, but a postmortem analysis usually unearthes considerable flaws.
Parent - - By davidwhite (***) Date 2007-11-30 20:03
            " Our decision-making processes integrate more data in a more sophisticated way than most people.  This is our biggest strength in relative terms.  That's the reason we're happy with mere equality out of book."

            Nelson, there must be some time control  below which the team(really,any team) wouldn't
be able to function in this manner.What would you estimate is the minimum time control needed
for the Cato Freestyle team to perform at a higher level than its engine operating without human
assistance?

           Regards,
           David
Parent - By Nelson Hernandez (Gold) Date 2007-11-30 23:22
I really can't answer that question, too many variables.  Without a doubt, though, the shorter the time control the less opportunity to calculate, digest data streams and add value.  At some point centauring a whole game with little time on the clock when you have little or no innate chess skill is too much.  What Anson and Wing's breaking point is, I am not sure.  I am guessing he'd be very good down to 30 minutes and no increment.  Less than that and it gets increasingly dicey--a GM with an extensive memorized repertoire could have an edge, particularly if our side was dumb enough to follow main lines.
Parent - - By Kapaun (****) Date 2007-12-01 13:59
And the top 5 freestylers would cry if they saw Anands opening preparation

Certainly by far the best opening preparation you can get is being in the machine room on a more or less regular basis. I doubt that Anand has ever been there. Getting a sound disadvantage out of the opening will even give Anand quite a few problems to solve...
Parent - - By turbojuice1122 (Gold) Date 2007-12-01 17:27
It's not clear-cut.  Clearly, Freestylers have far more extensive opening preparation for playing against machines than Anand could ever hope to have.  However, it's also true that top grandmasters are better at coming up with nice opening moves than are machines, and they will often come up with moves and sequences that the machines will address as bad, but then suddenly "change their minds" 10 moves later, after which it is, of course, too late.  This will not happen in every game, of course, which is why I say that it's not clear-cut who has the advantage in this--on some days, the Freestylers would have a clear advantage, and the Freestylers would definitely win some games in a 10-game match.  However, Anand may also win some games based on those days where he comes up with some really nice "human novelties".
Parent - - By Nelson Hernandez (Gold) Date 2007-12-01 19:50 Edited 2007-12-01 19:53
Exactly right.  It is entertaining and interesting to pit an "engine games only" book against an "FM+ games only" book and watch what happens.  In my observation, the human book stymies the engine book a lot, forcing the latter to exit book much earlier than you see when engines play each other.  The sooner this happens, typically the more unclear the final outcome.

Another way of observing this is to just pick any GM game off of TWIC at random and follow it in your engine book.  Time and again the humans exit book early.  They have a much lower propensity to follow main lines than you'd think.

Lesson: human OB trees are shaped quite differently than engine OB trees for the obvious reason that engine games tend to be in-bred while every human being is a unique chess engine with his or her own preferences, strengths and weaknesses.  As a result of this it isn't hard at all for a human to force an engine out of book with positional equality.  (In fact it is trivial with white pieces.)  It is what happens after that which makes all the difference.
Parent - - By turbojuice1122 (Gold) Date 2007-12-01 20:02
Okay, but how about for the more direct topic: what happens when you follow two Freestyle games in which you play neither side?  Do the Freestylers tend to exit your book earlier or later than the GMs?  Earlier or later than the computers?
Parent - By Nelson Hernandez (Gold) Date 2007-12-01 20:43
It depends at least in part on the properties of your book, doesn't it?

But aside from that different Freestylers have different approaches, resulting in a very broad continuum.  It is impossible to generalize about the species as a whole. 

We saw players try some really daring lines against us just in this last tournament, knocking us out of book really early (though not with advantages).  Others (like Alan) specialize in only a few lines and build up quite a bit of expertise.  Others (generally newbies) follow main lines slavishly.  Some (like Dagh and Eros) do fanatic analysis in a variety of lines and secretly bank deadly novelties for future use, creating a kind of chess minefield.  (We stepped on one of those mines last time.)  Some play human lines.  Some play engine room lines.  Some create new lines.  It's a jungle out there.
Parent - - By Vasik Rajlich (Silver) Date 2007-12-02 12:39
Actually, it would be really interesting to have Jeroen's book go up against Anand's book. I'm not sure what the result would be. My only prediction is that Jeroen would make a more realistic estimate of the outcome than Anand. Computer chess is still to a large extent living in the dark.

Vas
Parent - - By Nelson Hernandez (Gold) Date 2007-12-03 02:41
Actually, it would be really interesting to have Jeroen's book go up against OUR book.  A nasty little nip-and-tuck affair at close quarters it would be!

Someday.
Parent - - By Vasik Rajlich (Silver) Date 2007-12-04 12:55
It sure would, wouldn't it? :) Somehow, I suspect that both sides would claim a victory from the last clash - Jeroen's book gave us a clear advantage, while your book got you a draw.

Vas
Parent - By Nelson Hernandez (Gold) Date 2007-12-04 17:11
The book didn't give us a draw.  The book had nothing to do with it, which is the whole point!  The book certainly didn't tell us to play the offbeat and not entirely sound h6 instead of the standard a6! 

We accomplished all our strategic goals with the black pieces, facing a very formidable opponent who knew who he was playing.  We avoided your preparation and survived.  Had you blundered we might have even won.  There are bigger considerations than just achieving a favorable book-exit!

Give us white next time.  :)
Parent - - By lkaufman (*****) Date 2007-11-26 23:13
What on earth leads you to believe this? Surely the fact that 2500 level GMs haven't done well in freestyle has no bearing on this question. They are simply too far below the level of Rybka for their chess strength to be a major factor. But Anand (or Kramnik or Kasparov) is strong enough for his chess skill to be a significant contributor to the combined strength of human + machine.
Parent - - By rivaldo (***) Date 2007-11-27 00:32
as I understand nelson hernandez is part of a freestyle winner. if he states, that kramnik or anand wouldn't have a chance against his team it shows, that at least he is not a serious chessplayer. such a ridiculous claim makes me think, that the freestyle games are a bit overrated. but I must admit, that I didn't analyse them. kramnik or anand will hardly lose 1 out of 50 games against any top five team!
Parent - By lkaufman (*****) Date 2007-11-27 00:44
Well, I wouldn't go that far either. The strength of the team is mostly determined by the computer(s). Having Anand or Kramnik helping will make a real difference, but even an unassisted Rybka on the same hardware will get in some wins with the White pieces, and with a prize winning team operating I should expect a score of only around 65% or so for Anand & Rybka.
Parent - By Nelson Hernandez (Gold) Date 2007-11-27 01:15
You're absolutely right.  I am not a serious chessplayer.  I am not even an unserious chessplayer.  That's the whole point of my ridiculous claim.
Parent - - By Nelson Hernandez (Gold) Date 2007-11-27 01:11 Edited 2007-11-27 01:25
Larry, my contempt for GMs only mirrors their contempt for computer chess circa ten years ago.  They thought that humans had the upper hand then.  I think computers have the upper hand now by a wide margin, and if used properly that margin can be extended considerably.  Anyone, rated or not, who knows how to use computers properly--i.e. to their fullest potential--has an enormous advantage over GMs who are mostly relying on skill and experience, or use the computer passively.  Go ahead, ask Vas if he really disagrees with me.

My sense is that GMs lack some fundamental computer skills and playing insights that Freestylers--even non-OTB players--have in surfeit.  They also have a serious inherent disadvantage: their sense that we're not worth taking seriously, that what we do is a carnival act.  They are in complete denial that the world has turned and the top Freestylers are currently playing the strongest chess in the world, aside from the top correspondence players.

I am not saying that GMs couldn't learn some skills and improve.  But currently they are not inclined to do so as there is no financial or prestige incentive to do so.  Freestyle isn't their gig, just like OTB isn't ours. 
Parent - - By lkaufman (*****) Date 2007-11-27 01:42
What you say may be true for the majority of GMs, but I'm sure there are a decent number who understand perfectly well that computers are stronger in more than 95% of positions, and a decent number who do understand pretty well how to use computers fairly well. Also, grouping ordinary GMs (around 2500) with Anand and Kramnik (2800) is as wrong as considering Anand and Kramnik to be in the same class as Rybka. GMs at 2500 who rely on their judgement over Rybka are usually making a mistake, but when Anand or Kramnik disagrees strongly with Rybka he is likely to be right.
Parent - - By Nelson Hernandez (Gold) Date 2007-11-27 01:55
Being one of the top players in our geographical region for quite a number of years, personally familiar with many of the world's top players, and a Rybka insider/co-author to boot, I will respectfully yield to your judgment.  Just the same, it is too bad none of us will ever get the chance to test my brash hypothesis.  There would be no upside at all in swatting down Freestyle gadflys, moreover were the underdogs to prevail with merciless efficiency it could shake the foundations of chess.
Parent - - By turbojuice1122 (Gold) Date 2007-11-27 03:38
I don't see the hypothesis as unreasonable at all--in fact, I think that it's "obviously" true: Anand+his best hardware/software would not stand a chance in, say, the Freestyle finals tournament.

However, something that I think is also obviously true: give him another top centaur or two on his team (along with their computers/engines/opening books), and he will get first or second after they've had some practice working together.
Parent - - By Nelson Hernandez (Gold) Date 2007-11-27 13:01
Thanks for the support, Turbo.  Your second paragraph--right again.  I have no doubt that Anand could play a lot better with a little specialist help.  My premise was Anand playing alone on my turf with a top-of-the-line computer.  My hypothesis is that this scenario would really be an ambush as the Freestylers bring more hardware that is more creatively used, more opening prep in the relevant lines (operative word here is "relevant"), more eyeballs and a nothing to lose attitude. 

Of course, in a two-game exhibition anything could happen.  The Freestylers might choke, being so excited and honored to play the world champion.  But....nah.
Parent - - By Hetman (*****) Date 2007-11-27 15:24
If you are sure enough of your opinion concernig V.Anand chances vs freestylers then  top freestylers can send him the challenge as Vas has sent to Fide.
Rgds
Hetman
Parent - - By Nelson Hernandez (Gold) Date 2007-11-27 16:16
Not a bad idea.  But first we need to settle who is the strongest Freestyle team which could best represent the rest of us.  I don't think one tournament definitively answers that question, but the seven tournaments we have had establishes who is in the top group.  The obvious, popular choice at this point would be two-time champion and perennial contender Team Rajlich, but before that conclusion is made in haste perhaps we should have one more tournament...and if that still doesn't settle things then they should face the top challenger one-on-one.  :)  Let's see how the next tournament goes...I hear rumors it will be January-February.

Personally I would love to see Anand face the best Freestyle team.  It would be intensely interesting.  But the problem is that I doubt he would take it seriously unless there was a major financial incentive, which seems about as likely as a rejuvenated Elvis appearing in Las Vegas, singing "Suspicious Minds".
Parent - - By Hetman (*****) Date 2007-11-27 18:12
I am wondering about conditions of such a match, I doubt that VA will want to play remotely so the contenders will have to be in the same room and using equal machines. It looks obvious.

I am not sure if we were speaking about the best freestyler or the best freestyler team for the match vs Anand ?

I will be interested in match in advanced (freestyle) chess between best freestyler vs GM Joel Benjamin for the time being. :-).
The result of such the match is not clear for me.

Rgds
Hetman
Parent - - By Banned for Life (Gold) Date 2007-11-27 18:49
I think that would be an excellent match. I would suggest having GM Benjamin play against a single opponent, probably either Dagh or Anison, on equal hardware (maybe a single quad or octal machine) with whatever preparation they wanted to bring in with them. Given that GM Benjamin has accepted matches against Rybka that he was not likely to win for an ~$11K purse, it might be possible to do this with a purse of $5K or even less (since his chances of winning would be significantly higher).

I'm not sure how I would handicap this. At the risk of pissing a lot of people off, I'd probably give GM Benjamin a slight edge under these conditions.

Regards,
Alan
Parent - - By Nelson Hernandez (Gold) Date 2007-11-27 19:47
Slight edge?  How about a clear advantage to Benjamin? 

What you are laying down is wildly divergent from the conditions I proposed.  One centaur on a single machine?  Right, just take triple-digit ELO points off of the team's strength!  If Cato/Ibermax plays a GM, the TEAM is playing, which means three people (one of them alternately napping and smoking cigars, i.e. nearly useless) and all the high-end computers we can usefully handle. 
Parent - - By Hetman (*****) Date 2007-11-27 21:39
NH  have you started with that ?

"Honestly, I believe Anand + top-end octo with any software he wanted would not stand a chance against any of the top five Freestylers in a straight-up non-exhibition match, if they were allowed to use anything they had in their arsenals."

One of top five ... it does not mean team but one ?
Rgds
Hetman
Parent - - By Nelson Hernandez (Gold) Date 2007-11-27 22:42
Good catch.  When I said "Freestylers" I meant teams, not individuals.  My fault for being imprecise.  My point was that WC Anand playing Advanced chess on a top machine is at a disadvantage against a collection of relative nobodies playing Freestyle--which implies teams, arrays of computers, whatever the latter can pull together.  Put in simple terms, my contention is that the value-add of a Super-GM is less than the value-add a top Freestyle team puts together for an ordinary tournament.

Of course such a match would be problematic, not just for financial reasons.  A top Freestyle team is, by its nature, a virtual entity that isn't very portable on account of the extensive computer gear it uses, which is set up in a particular way which doesn't easily permit replication somewhere else.  The team members engage in continual shorthand conversation and complex routines that have been developed over a period of time.  You couldn't put such an entourage on the same stage with Anand--it would be distracting to both sides, but probably more so for the Freestylers, who require a certain chemistry and atmosphere to work well together. 
Parent - - By Banned for Life (Gold) Date 2007-11-28 04:21
If what you have meant all along is that a team of people with an order of magnitude more processing hardware can beat a top GM working with a single machine, I guess I don't find that to be either surprising or impressive in any way.

Alan
Up Topic Rybka Support & Discussion / Rybka Discussion / Kramnik-Anand Advanced finished
1 2 3 4 Previous Next  

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill