Greengard's ChessNinja.com

London Massacre

| Permalink | 42 comments

Things have been relatively quiet at ChessBase.com because the main editor was in New York for the Accoona match and staying with the other site contributor, me. We were busy, what can we say? So many bagels to eat.

Meanwhile, Mickey Adams is being obliterated by Hydra in their London match. (Their website has been rather overwhelmed. Maybe they have a Spanish web designer.) He has a draw and two really brutal losses. I thought he would lose but fare better than this. Kasparov said that Adams wouldn't have a chance. As Mickey himself put it, Hydra is stronger than Junior and Fritz and Kasparov is stronger than him, so by rights he should have a rough time.

But the transitive theory has never done well in chess, and less so when computers are playing humans. A very well-prepared 2600 would have chances as good as a 2700 without good anti-computer preparation. Two or three extra ply is lethal against another computer, but not nearly as relevant in the clash of style against humans.

That said, Hydra has totally dominated one of the world's best players. It looks like the only way the Hydra folks are going to learn anything is to keep playing in Advanced Chess (aka "Freestyle") events, which are actually a good way for programmers to learn. When strong (or even fairly weak) humans don't blunder, they consistently beat even the strongest machines. Let's just cross humanity's collective fingers and hope Adams can notch a win.


In the third game, you get the feeling that Black's lost tempos with the queen's bishop combined with the novelty 16... Bd8?! unleashed that monster Hydra! The black squares in Adams camp were exploited precisely by the program. 28.Bxh6! the final blow, revealed those weaknesses to the fullest.

Very few people, even armed with programs, saw Hydra's attack coming. I think it wasn't a bad game by Adams at all apart from his failure to play real anti-computer chess, and yet he went down in flames. Scary!

I'm glad to see Adams getting his ass kicked.

If Adams doesn't recover, maybe Kasparov and others will finally get off their soapboxes and agree that there just isn't anything worth debating about computers' chess strength anymore. The guys in UAE estimate Hydra to be about 2950 ELO, and I believe them.

Chess is no longer considered a metric for AI. Watching Hydra kick Adams's ass is like watching Tom Gordon's Porsche kick Carl Lewis's ass. Who cares? Chess computers are machines built to perform a single task, and computers like Hydra are the ultimate at this task; so sorry Mr. Kasparov, there is no "thinking" involved.

Hydra is to Fritz as a Porsche is to a Toyota.

Perhaps freestyle chess, or advanced chess will take on a new competitive following, but I think it's time we focus back to human versus human chess, and the competitive sport and human enjoyment that this provides.

Howard Goldowsky

"If Adams doesn't recover, maybe Kasparov and others will finally get off their soapboxes and agree that there just isn't anything worth debating about computers' chess strength anymore. The guys in UAE estimate Hydra to be about 2950 ELO, and I believe them."

Adams is not exactly a Kasparov or a Kramnik (in the context of man-computer matches that is; the latter was still 2800+ when he played Deep Fritz) *and* he hasn't been in his best form lately. And while I think he hasn't played badly except in game 1, it's also a matter about how well you master anti-computer techniques.

Even if he loses all the remaining games it would be no proof that machines have taken over. Even for the world's best player one match wouldn't prove it anyway. Let's see Anand, Leko, Kramnik, Topalov all take on Hydra and evaluate the overall result.

So much hype about Hydra.

The way Adams has played this match he would have been kicked by other PC softwares also. I checked all 3 games with Shredder 9 on my modest PC. Shredder plays the same key moves as Hydra or sees Hydra’s moves as 2nd or 3rd alternative. No big deal about Bxh6, Shredder 9 finds the move very quickly.

The truth is that Adams is playing under his level. He has played too much recently and did not have enough time to recover or prepare. If Adams would have got same amount of $$$$ as Kasparov or Kramnik (when they played the comps), he would have taken the match more seriously.

Of course it finds Bxh6. The point is the preceding play.

Jon Speelman did!

Oh shut up.

The lesson to be drawn from this match?

If you want your chess machine to beat a super GM, you must write to Heavy Metal or some other sci-fi magazine and hire their best graphic designer. He will draw the most intimidating looking beast imaginable.

You then name the machine accordingly and create its image around the monster poster. Then you build its reputation up so that eveyone thinks that this hideous beast sees everything and will devour all comers.

Finally, you organise a match against the smallest super GM you can find, and seat him facing the poster:

Results guaranteed.

Seem like the website folks could have made a better looking chess set. I kept thinking the queen was a rook when using their site to play over the games.

Off to a new challenge! This one is done.

Perhaps marketing the first PDA stronger than any human player?

Actually, if not now, very soon, computers will have to give handicaps to humans to make it any interesting. Maybe, allowing take back moves, or something.


I think the best player against computers is Anand. Not even Kasparov. Maybe they should play him against Hydra.

Strategy against Hydra: Lose with Black and draw with White.

You can still make some money that way :-)

I hope Adams doesn't complain that GM Lutz played some key move in game 3 causing him to lose the game badly and demoralize :-))

It's unfortunate Adams isn't playing anti-Computer chess. That is the only realistic way to beat computers nowadays.

Like Kasparov, Adams had no time to prepare anti-computer chess for this match. If atleast 3-6 months are needed just to prepare for World Championship match, how much more you need to play the monster computer? Its high time, the GMs should develop anti-computer strategies. Adams is loosing due to fear. Play the anti-computer chess and we can see how Hydra can perform.

Sorry, Kasparov had more time and resources to prepare for his match.

Adams has had months of notice, but he also played a busy tournament schedule. The question is whether or not he did much preparation. I've been talking to Garry about this match (and the freestyle tournament) a lot in the past few days. He wants to emphasize how important it is to really study the computer opponent. The "good" news is that while Hydra is undoubtedly stronger than the Fritz and Junior machines he and Kramnik played, this terrible result shows how hard Kasparov and Kramnik had to work to get those drawn matches. Preparation is the key to it. Kasparov will touch on the match in his next New In Chess article, including the freestyle tournament.


A GM who might be interesting to watch against computers is Kamsky. He seems to play well against them on ICC and taught one of my friends several things about playing against computers. (When a 2000 player consistently draws or beats Crafty, this is quite an accomplishment and a testament to the teaching ability of Gata, as well.) He still understands the game at 2700 level, evidenced by some of his quick analysis of games in the recent US Championship (instantly spotting a winning idea for Nakamura against Kudrin and also recommending, out-of-hand, a queen sacrifice IM Lapshun could have played against him to save a draw in their game), and it seems that only his knowledge of modern opening theory prevents him from holding this form once again. (This is similar to a great boxer whose footwork is not back to speed to allow him to show the force of his right hook.) I don't know if Gata would play, but it would certainly be interesting even to have him on a team which prepares a GM for such a match as this one.



Though on surface it looks as if Hydra is mauling Adams game after game, Adams has at times played almost spectacularly, like on the 4th game, i only think the pressure was too strong, and Adams was unable to keep up the intensity for perfect defense, and he went wrong not taking the d5 pawn when he had a chance but went for h7...i hope Adams' best day is still to come and when it comes and he plays as flawlessly as he can, i'm certain he can get a victory.
A draw against a computer as strong as Hydra is already a telling result that computers don't exactly dominate despite raw results, because the computer plays consistently, not better on one day and worse the next.

In any case, as Adams has not been playing anti-computer chess, this match has not been embarassing for him at all, even though looking at the pure results one might think so. Great games, and at times Adams has defended as well as Leko could.

And the massacre continues: 4,5-0,5 now !

It doesn't look like Adams did (or had time to do) any serious opening preparation for this match. I don't know who will play Hydra next -even though I would love to see Kasparov playing again to defend humans-, but I think that any super-GM with a focused opening preparation for this match could have done much better than Adams. Adams didn't seem like he was in top form in general, but he never got out of the opening with anything-except maybe in Game 2 and that ended in a draw.

Hydra's creator, Chrilly Donninger says "[Adams] is a far more difficult opponent than Kasparov, and Adams has played extremely well.".

Seems to me, the Hydra team is itching to provoke Garry into accepting a match! Nice trap :-)

If there's a human out there who could tame this beast, his name is Kramnik - not the 2004-05 edition, but the one who came close to beating Fritz...

I have a feeling that if Adams would play, say,
20-30 games against this opponent he would eventually learn the best strategy against Hydra. I think Hydra would eventually lose one or two games. Now, it's still unclear how to beat it. But maybe he has already learned ... the next game Adams as white will be interesting.

The best post is the one from dingding...I totally agree.

More seriously I think everything the top GM size can try and say about this subject is not possible for a human to win such a computer if they sit and face it in the tournaments condition and not in the human more confortable condition for chess.

I explain: the tournament rules (if u touch a piece u must move, the time control set to 2hx40 moves)exist to the aim of organizing interesting humans vs humans events.
They are not organize to let the GMs express their absolute maximum chess strenght but to express their best level RELATIVE to the rules of the tournament.

The computer has free access to all the tools of modern tecnology, humans can use only his brain.
Humans are limitated by rules in using eyes and hands for example...instead we can say that the computer uses not only the electronic correspondent of a human brain but also the electronic correspondent of eyes and hands.

It seems me that if from another planet ET people match earth inhabitants in a car race, the best human car against the best ET car we will use our best possible car.
We'll ask our engineers to build the best possible car (may be a 20000cc and 1200 km/h, who knows?).
We will not be satisfied if engineers say us:
"Why work on this, we can use a F1 car, they are quick despite their only 3000cc..." and face the probable ET monster with Shumacher and his Ferrari needing pit stops, petrol limit, 350 km/h max and all the F1 limitative rules....to accept at this condition will be a suicide (like Adams one).

Now that Adams' embarrassing 5.5-0.5 loss is behind, I'll say it again: the *only* way to beat a computer as strong as HYDRA is to use anti-computer chess.

It's almost like Adams was trying to lose. When was the last time *any* human beat a top computer in an open Sicilian? Yet that was what Adams tried to do. To beat a computer you need to play dirty, cagey, closed, tricksy chess, not the open classical chess Adams was playing.

I should add - Adams is too much of a nice boy. To beat a computer, you need a ruthless, intelligent, pragmatic, cynical player who's not interested in "fighting the good fight." Someone like Evgeny Bareev.

Am I the only one who didn't pay this match the least bit of attention? With players like Ivanchuk (who won a nice game against Karjakin today, btw), Radjabov, Dreev etc., playing at the same time as the Adams-Hydra match, I chose the human-human games. More fun and easier to understand.

I must say...

The thrashing of Adams gives me an idea of how Larsen and Taimanov must've felt after losing 6-0 to Fischer... that's got to be totally demoralizing.

Human-human games are more fun to watch, I agree. And a player like Ivanchuk is always a delight - no quick draws for him; he goes for a win every time, black or white.

That said, the HYDRA match still has a fair bit of interest.

I hope Kasparov does accept a match against Hydra - the ultimate chess showdown - the highest rated player of all time against the highest rated computer. I read somewhere that the Hydra team estimates its elo currently at 2950. Unfortunately for Garry, if he does play it in the next year, it will be stronger by then. As soon as they put in the other 32 processors, its speed is supposed to improve by 80%. With a team that is constantly upgrading, tweaking, and improving the machine, we can expect Hydra to be playing around 3000 or more by then with Garry maybe playing around 2800.
If I were a gambling man, I would put my money on the machine.

No offense to the HYDRA team, but the 2950 figure sounds like it was pulled straight out of an orifice.

*Unless* they are basing it on Hydra's results vs. other computers. But this has not historically been the a reliable measure of performance vs. humans.

Regardless of the result (I would bet on Hydra, too), GK would have to be insane to do something like this after his retirement from professional chess. (Which also implies that he would play for free, doesn't it?)

why does it imply that Ling? Btw, Chuky rocks, one of the greatest ever.. or to be more precise, one of the greatest never so to speak, unless he can hold it together in the Argentine thingy..

Just replace "insane" with "incredibly well paid" and you're probably on to something!

The Hydra team will have no reason to keep plowing money into it to make it faster if it scores like this. That's the problem; when one side wins all the time the experiment is over.

" And a player like Ivanchuk is always a delight - no quick draws for him; he goes for a win every time, black or white."

I hope you are being sarcastic. Was that Linares 2003 where Ivanchuk was playing one short draw after another - the only exception being against Topalov?

Actually, it was Wijk ann Zee 2003 where he drew alot of short games. In fact, his loses and wins were also around 30 moves in that tournament...

Ivanchuck is the greatest crazy genius in chess. But, he is too emotional to play really well against computers( or humans). I think Anand with his tactical vision and cold logical style would be great.

Macuga "When was the last time a human beat a computer in an open sicilian" Sergei Karjakin-Deep Junior Bilbao 1-0.

Ha, ha. Just curious, why do you call Ivanchuk 'the greatest crazy genius in chess'?

freddy i think both his games and his persona contribute to that image.. Even the one and only Kasparov has said that Ivanchuk is the opponent who surprised him most otb. Add to that mannerisms like "gazing at a spot 6 inches over your left shoulder" and you have the idea..

Ahem, I'd just like to point out this was a gross error and I was not in fact talking to myself - the 'Oh shut up' was directed at someone else on a different site i happened to be reading at the time and I accidentally typed it here too. Oops.

Hydra may have beaten Adam, however Man is still the victor.Science said that we humans only use 3% of our brain. We have yet to use the other 97%.
The best players would be able to beat Hydra if they just use another 2%. I have the feeling that many masters do not want to master the anti-computer strategies as it will upset their usual style. They will end up being poorer players against other humans. Howevr I do believe a group of anti-computer specialists will emerge.

"Science said that we humans only use 3% of our brain. We have yet to use the other 97%."

I'm afraid this is a persistent myth. Science certainly does not say this. You'll see figures of 3%, 10%, 20% but they're all wrong. Just as we don't use all our muscles at once we won't use all our brain at once. Doesn't mean you're not using it.

Each area of the brain seems to have a use from thinking to moving to feeling as well as maintaining life. In general, damage to any part of the brain will result in some form of disability or noticable change.

For some more info, see


Twitter Updates

    Follow me on Twitter



    About this Entry

    This page contains a single entry by Mig published on June 24, 2005 4:32 AM.

    Nakamura Wins Samford was the previous entry in this blog.

    Anand Chat is the next entry in this blog.

    Find recent content on the main index or look in the archives to find all content.