Hydra v Adams

Main

Date: 29 Jun 2005 09:03:25
From: Peter Billam
Subject: Hydra v Adams

HYDRA beat Michael Adams 5.5-0.5.

http://www.chesscenter.com/twic/twic.html 's take on the result is

> Adams clearly didn't prepare properly for the match and was duly
> hammered (Kasparov and Kramnik spent months preparing for matches
> against Deep Fritz and Junior).

but I'm not sure that's fair. As Kasparov pointed out, how are you
supposed to prepare when you have no access to the opponent's games,
or opening repertoire ? Perhaps Hydra is simply much stronger than
DB, DF & J were ...

Will this be remembered as the moment when the strongest computers
became decisively stronger than the strongest humans ? It has to
happen sometime...

I like Hydra's choice of openings, choosing sensible mainlines
and then a reasonable little-played byline, just aiming for a
balanced position in which it can deploy its playing abilities.
None of the games were decided in prepared miracle-rook-sacrifices.

I wonder what the future holds for Hydra. Presumably its time could be
sold to GMs wanting to analyse more miracle-rook-sacrifices. Perhaps
some of its special-purpose hardware could be sold to users and patched
into a commercial chess engine, perhaps a dedicated palmtop device.

There may not be many more human volunteers for high-profile matches :-)
But on the other hand, if the top humans continue to refuse to play
against each other, perhaps it could be used as a standard opponent,
with World Champion being deemed to be that person who scores best
against it. Or if it becomes much stronger, then the World Champion
could be that human whose games it decides are of the best quality.
Or perhaps humans will completely bow out, and the World Championship
will be fought between rival corporations, like F1 ...

Regards, Peter

P.S. Is there a published analysis of the last game ? I know that
1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 a6 5. Bd3 Bc5 6. Nb3 Ba7
was at one time considered a bit dubious for Black...

--

Peter Billam, DPIWE/CIT/Servers, hbt/lnd/l8, 6233 3061

Date: 29 Jun 2005 07:29:02
From: Henri Arsenault
Subject: Re: Hydra v Adams

It would be interesting to see an analysis of the actual moves in the
games. Any novelties? Any new tactics? Any new theory?

A machine like Hydra could be useful to test new ideas (or old ones) in
the openings, and it would be fun to watch a game between two grandmasters
where both had assistance from Hydra. We would then see the best from both
humans and machines.

OTOH I find it dangerous to draw too many conclusions from a single match
between a single opponent and Hydra.

Henri

Date: 30 Jun 2005 11:20:21
From: Antonio Torrecillas
Subject: Re: Hydra v Adams

En/na Henri Arsenault ha escrit:
> It would be interesting to see an analysis of the actual moves in the
> games. Any novelties? Any new tactics? Any new theory?

Adams usually play "not main lines" and theree is not much theory in
them (well, shall does not count).

When humans play with engines the human can achieve good positions but a
single mistake changes the game. Maybe you undestand better the position
but a blunder cost you the entire point, ... that happens seldom.

Surprisingly in that match it seems HYDRA allways obtained safe
positions and it had no problems in any game and in any moment. That can
mean that some extra ply can produce "positional understanding". But
well, more extra games are needed as you point later.

> A machine like Hydra could be useful to test new ideas (or old ones) in
> the openings, and it would be fun to watch a game between two grandmasters
> where both had assistance from Hydra. We would then see the best from both
> humans and machines.

Search in chessbase site the freestyle tournament, ... very interesting!

The problem with brute force is that it only produces "good moves" but
not good explanations. You can compare engine analysis of a game with
some notes of an human (for exemple here you have Claus Jurgen ones), I
think last ones are much more interesting.

> OTOH I find it dangerous to draw too many conclusions from a single match
> between a single opponent and Hydra.
>
> Henri

Date: 30 Jun 2005 12:09:36
From: David Richerby
Subject: Re: Hydra v Adams

Antonio Torrecillas <[email protected] > wrote:
> En/na Henri Arsenault ha escrit:
>> It would be interesting to see an analysis of the actual moves in the
>> games. Any novelties? Any new tactics? Any new theory?
>
> Adams usually play "not main lines" and theree is not much theory in
> them (well, shall does not count).

Also, Hydra had a very shallow opening book -- I think I read that it
only goes down about ten moves in most cases.

Dave.

--
David Richerby Miniature Generic Atlas (TM): it's
www.chiark.greenend.org.uk/~davidr/ like a map of the world but it's just
like all the others and you can hold
in it your hand!

Date: 30 Jun 2005 15:28:34
From: =?ISO-8859-1?Q?Claus-J=FCrgen_Heigl?=
Subject: Re: Hydra v Adams

David Richerby wrote:
> Also, Hydra had a very shallow opening book -- I think I read that it
> only goes down about ten moves in most cases.

In the two games I analysed (game 3 and 4) it was both times Adams who
made the first non-book move. In game 3 Adams deviated with 16...Bd8 and
in game 4 Adams played the new move 11. Qd3.

Claus-Juergen

Date: 30 Jun 2005 18:30:37
From: David Richerby
Subject: Re: Hydra v Adams

Claus-Juergen Heigl <[email protected] > wrote:
> David Richerby wrote:
>> Also, Hydra had a very shallow opening book -- I think I read that it
>> only goes down about ten moves in most cases.
>
> In the two games I analysed (game 3 and 4) it was both times Adams who
> made the first non-book move. In game 3 Adams deviated with 16...Bd8 and
> in game 4 Adams played the new move 11. Qd3.

It's entirely possible for Hydra to find book moves over the board, of
course.

Dave.

--
David Richerby Accelerated Smokes (TM): it's like
www.chiark.greenend.org.uk/~davidr/ a pack of cigarettes but it's twice
as fast!

Date: 30 Jun 2005 14:02:58
From: Antonio Torrecillas
Subject: Re: Hydra v Adams

En/na David Richerby ha escrit:
> Antonio Torrecillas <[email protected]> wrote:
>
>>En/na Henri Arsenault ha escrit:
>>
>>>It would be interesting to see an analysis of the actual moves in the
>>>games. Any novelties? Any new tactics? Any new theory?
>>
>>Adams usually play "not main lines" and theree is not much theory in
>>them (well, shall does not count).
>
> Also, Hydra had a very shallow opening book -- I think I read that it
> only goes down about ten moves in most cases.
>
> Dave.

That is difficult to asume, ... in that case any human could beat HYDRA
playing a Sicilina Najdorf or any similar extensively analized line
without playing any new move.

I think that opening book of most engines contains all known theory plus
some new ideas from book developers.

Another fact is that in some position new moves can be in 30th move and
in anothers can be in first ten moves (I remember Kasparov obtained "out
of book" with white in first moves playing Reti Opening in 1997 in his
match with DEEP BLUE)

AT

Date: 30 Jun 2005 18:29:07
From: David Richerby
Subject: Re: Hydra v Adams

Antonio Torrecillas <[email protected] > wrote:
> En/na David Richerby ha escrit:
>> Also, Hydra had a very shallow opening book -- I think I read that it
>> only goes down about ten moves in most cases.
>
> That is difficult to asume,

From http://tournament.hydrachess.com/faq.php :

``Is there some special opening preparation done?

``Not directly. The Hydra opening book is very short. Typically 10
moves. After 10 moves we let the monster from the leash and rely on the
playing strength of the program. We know the favorite opening lines of
GM Adams and try to play the most active variations. But there are no
attempts to develop some novelties or opening traps. We think it is
nowadays almost impossible to find real good alternatives within the
first 10 moves. Hydra shall find some new lines on its own after the
10th move.''

> ... in that case any human could beat HYDRA playing a Sicilina Najdorf
> or any similar extensively analized line without playing any new move.

Hydra would mince any human with either side of the Najdorf: it's just too
tactical. Most likely, the human would forget a piece of theory and Hydra
would work it out over the board; or Hydra would refute theory over the
board.

> I think that opening book of most engines contains all known theory plus
> some new ideas from book developers.

Most engine books do, yes.

Dave.

--
David Richerby Psychotic Sadistic Radio (TM): it's
www.chiark.greenend.org.uk/~davidr/ like a radio but it wants to hurt you
and it wants to kill you!

Date: 01 Jul 2005 03:16:52
From: Antonio Torrecillas
Subject: Re: Hydra v Adams

En/na David Richerby ha escrit:
> Antonio Torrecillas <[email protected]> wrote:
>
>>En/na David Richerby ha escrit:
>>
>>>Also, Hydra had a very shallow opening book -- I think I read that it
>>>only goes down about ten moves in most cases.
>>
>>That is difficult to asume,
>
> From http://tournament.hydrachess.com/faq.php :
>
> ``Is there some special opening preparation done?
>
> ``Not directly. The Hydra opening book is very short. Typically 10
> moves. After 10 moves we let the monster from the leash and rely on the
> playing strength of the program. We know the favorite opening lines of
> GM Adams and try to play the most active variations. But there are no
> attempts to develop some novelties or opening traps. We think it is
> nowadays almost impossible to find real good alternatives within the
> first 10 moves. Hydra shall find some new lines on its own after the
> 10th move.''

If I have read correctly, that mean that HYDRA creators concentrate in
finding "novelties" after 10th move and not before. That's very
different to say that opening book has no moves beyond 10th move.

a second view (Michael Adams opinion):
... from http://www.chessbase.com/newsdetail.asp?newsid=2485

"... not too much was known about Hydra really until very recently, when
it played games against Topalov, Ponoiov and Karjakin in the Man
against Computer event. Okay, people understood that Hydra was strong,
but since then they have managed to improve its performance and opening
preparation dramatically."

(...)

"... and also they were very clever in their opening choice. I mean that
made a big difference. The opening preparation of Hydra was completely
different to any other computer. You know in game one they came up with
this powerful novelty.

Were you surprised by 14.Rb1? I mean it has been played before�

Yeah but with a completely different idea to play b5 and� well, it was
just very powerful. It was obvious that it was prepared before the game,
and the creators of the program have said that it created the type of
position which is not possible for a human being to defend."

AT

Date: 01 Jul 2005 10:12:54
From: David Richerby
Subject: Re: Hydra v Adams

Antonio Torrecillas <[email protected] > wrote:
> En/na David Richerby ha escrit:
>> From http://tournament.hydrachess.com/faq.php :
>>
>> ``Is there some special opening preparation done?
>>
>> ``Not directly. The Hydra opening book is very short. Typically 10
>> moves. After 10 moves we let the monster from the leash and rely on the
>> playing strength of the program. We know the favorite opening lines of
>> GM Adams and try to play the most active variations. But there are no
>> attempts to develop some novelties or opening traps. We think it is
>> nowadays almost impossible to find real good alternatives within the
>> first 10 moves. Hydra shall find some new lines on its own after the
>> 10th move.''
>
> If I have read correctly, that mean that HYDRA creators concentrate in
> finding "novelties" after 10th move and not before. That's very
> different to say that opening book has no moves beyond 10th move.

I don't think you have read it correctly. The phrases, ``The Hydra
opening book is very short. Typically 10 moves,'' and, ``there are no
attempts to develop some novelties or opening traps,'' seem to make it
absolutely clear that Hydra does not have a deep opening book full of
prepared novelties. The phrases, ``After 10 moves we let the monster from
the leash and rely on the playing strength of the program,'' and ``Hydra
shall find some new lines on its own after the 10th move,'' seem to make
it absolutely clear that any novelties played by Hydra are its own, so to
speak: calculated by the program itself and not fed in by some openings
coach. The only opening preparation they admit to doing is looking at
Adams's games to work out what openings he likes and choose the most
active lines within them.

Dave.

--
David Richerby Sadistic Tool (TM): it's like a hammer
www.chiark.greenend.org.uk/~davidr/ but it wants to hurt you!

Date: 01 Jul 2005 16:26:56
From: Antonio Torrecillas
Subject: Re: Hydra v Adams

En/na David Richerby ha escrit:
> I don't think you have read it correctly. The phrases, ``The Hydra
> opening book is very short. Typically 10 moves,'' and, ``there are no
> attempts to develop some novelties or opening traps,'' seem to make it
> absolutely clear that Hydra does not have a deep opening book full of
> prepared novelties. The phrases, ``After 10 moves we let the monster from
> the leash and rely on the playing strength of the program,'' and ``Hydra
> shall find some new lines on its own after the 10th move,'' seem to make
> it absolutely clear that any novelties played by Hydra are its own, so to
> speak: calculated by the program itself and not fed in by some openings
> coach. The only opening preparation they admit to doing is looking at
> Adams's games to work out what openings he likes and choose the most
> active lines within them.
>
> Dave.

ok, we have different perceptions.

Best regards
AT

Date: 04 Jul 2005 00:07:35
From: AK
Subject: Re: Hydra v Adams

On Fri, 01 Jul 2005 16:26:56 +0200, Antonio Torrecillas
<[email protected] > wrote:

>En/na David Richerby ha escrit:
>> I don't think you have read it correctly. The phrases, ``The Hydra
>> opening book is very short. Typically 10 moves,'' and, ``there are no
>> attempts to develop some novelties or opening traps,'' seem to make it
>> absolutely clear that Hydra does not have a deep opening book full of
>> prepared novelties. The phrases, ``After 10 moves we let the monster from
>> the leash and rely on the playing strength of the program,'' and ``Hydra
>> shall find some new lines on its own after the 10th move,'' seem to make
>> it absolutely clear that any novelties played by Hydra are its own, so to
>> speak: calculated by the program itself and not fed in by some openings
>> coach. The only opening preparation they admit to doing is looking at
>> Adams's games to work out what openings he likes and choose the most
>> active lines within them.
>>
>> Dave.
>
>ok, we have different perceptions.

There is nothing here about "perception." Is English your first
language? It's very clear on Hydra's web site (see the FAQ
http://tournament.hydrachess.com/faq.php) that the opening book of
Hydra is around 10 moves deep (unless you don't understand English I
don't see how you can interpret that differently.)

Date: 04 Jul 2005 18:46:24
From: David Richerby
Subject: Re: Hydra v Adams

AK <[email protected] > wrote:
> Antonio Torrecillas <[email protected]> wrote:
>> En/na David Richerby ha escrit:
>>> I don't think you have read it correctly.
>>
>> ok, we have different perceptions.
>
> There is nothing here about "perception."

Au contraire. Antonio and I have read exactly the same text so our coming
to different conclusions from it can only be a matter of perception. :-P

Dave.

--
David Richerby Disposable Laser (TM): it's like an
www.chiark.greenend.org.uk/~davidr/ intense beam of light but you never
have to clean it!

Date: 04 Jul 2005 19:09:00
From: AK
Subject: Re: Hydra v Adams

On 04 Jul 2005 18:46:24 +0100 (BST), David Richerby
<[email protected] > wrote:

>
>AK <[email protected]> wrote:
>> Antonio Torrecillas <[email protected]> wrote:
>>> En/na David Richerby ha escrit:
>>>> I don't think you have read it correctly.
>>>
>>> ok, we have different perceptions.
>>
>> There is nothing here about "perception."
>
>Au contraire. Antonio and I have read exactly the same text so our coming
>to different conclusions from it can only be a matter of perception. :-P

Only if he doesn't speak English or has some magic abilities to read
between lines.

Date: 30 Jun 2005 20:05:41
From: Ron
Subject: Re: Hydra v Adams

In article <7JD*[email protected] >,
David Richerby <[email protected] > wrote:

> Hydra would mince any human with either side of the Najdorf: it's just too
> tactical. Most likely, the human would forget a piece of theory and Hydra
> would work it out over the board; or Hydra would refute theory over the
> board.
>

This may be going a little far. I suspect that Hydra's "ten moves deep"
are steering away from some of the biggest theoretical minefields. I
don't think you want to get into certain open sicilians, or, say,
semi-slav positions with a computer without an opening book.

-Ron

Date: 30 Jun 2005 22:50:25
From: David Richerby
Subject: Re: Hydra v Adams

Ron <[email protected] > wrote:
> David Richerby <[email protected]> wrote:
>> Hydra would mince any human with either side of the Najdorf: it's just
>> too tactical. Most likely, the human would forget a piece of theory
>> and Hydra would work it out over the board; or Hydra would refute
>> theory over the board.
>
> This may be going a little far.

Possibly. But I'd be very surprised if a human went in for that kind of
position against any strong computer, let alone Hydra.

> I suspect that Hydra's "ten moves deep" are steering away from some of
> the biggest theoretical minefields.

That's a very good point, yes.

Dave.

--
David Richerby Poisonous Soap (TM): it's like a
www.chiark.greenend.org.uk/~davidr/ personal hygiene product but it'll
kill you in seconds!

Date: 29 Jun 2005 12:41:52
From: Ray Gordon
Subject: Re: Hydra v Adams

> A machine like Hydra could be useful to test new ideas (or old ones) in
> the openings,

A machine like FRITZ can be very useful for that purpose as well. Don't
forget that Hydra got its butt kicked 2-0 in a correspondence match earlier.

What computers do is punish GMs who can't play strong openings. Once they
get an advantage, it's usually game-over.

Date: 29 Jun 2005 01:59:41
From: =?ISO-8859-1?Q?Claus-J=FCrgen_Heigl?=
Subject: Re: Hydra v Adams

Date: 30 Jun 2005 09:12:26
From: Peter Billam
Subject: Re: Hydra v Adams

Peter Billam wrote:
> P.S. Is there a published analysis of the last game ? I know that
> 1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 a6 5. Bd3 Bc5 6. Nb3 Ba7
> was at one time considered a bit dubious for Black...

In <[email protected] >, Claus-J�rgen Heigl wrote:
> I searched for it also and found nothing in the web. When I'm
> finished with game 4 I plan to do an analysis myself.

The game was:

[Event "Man-Machine"]
[Site "London ENG"]
[Date "2005.06.07"]
[Round "6"]
[White "Adams,Mi"]
[Black "HYDRA"]
[Result "0-1"]
[WhiteElo "2737"]
[EventDate "2005.06.21"]
[ECO "B42"]

1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 a6 5. Bd3 Bc5 6. Nb3 Ba7 7. O-O Ne7
8. c4 d6 9. Nc3 Nbc6 10. Qe2 O-O 11. Be3 e5 12. Rad1 Nd4 13. Bxd4 exd4 14.
Nd5 Nc6 15. f4 Qh4 16. Nd2 Be6 17. Nc7 Rac8 18. Nxe6 fxe6 19. g3 Qe7 20. a3
e5 21. f5 Nb8 22. Kg2 Nd7 23. b4 Kh8 24. Bc2 Nf6 25. Rc1 Rc7 26. Bb3 Rfc8
27. Rc2 a5 28. Rfc1 Qe8 29. h3 a4 30. Ba2 Re7 31. c5 dxc5 32. bxc5 Rec7 33.
Be6 Rd8 34. Qd3 g6 35. Kh2 Qc6 36. Qf3 Rf8 37. g4 Qb5 38. Qg3 Qe2+ 39. Qg2
Qe3 40. Qg3 Rxc5 41. Qxe3 dxe3 42. Nf3 Nxe4 43. Kg2 Kg7 0-1

This 5...Bc5 line seems to be known as the "Polugaievsky variation" :-)

But, writing in 1986, Polu says of 5...Bc5 "This setup has lost its
former poularity" and gave:

7. Qe2! Nc6 8. Be3 Bxe3 9. Qxe3 Nf6
{ also after 9... Nge7 10. Nc3 O-O 11. O-O Black has problems }
10. Nc3 d6 11. O-O-O b5
( after 11... O-O 12. f4 Qc7 13. g4 b5 14. g5 Nd7 15. f5!
White's kings-side attack is dangerous }
12. Rd2 O-O 13. Rhd1 Qc7 14. f4 b4 15. Ne2
{ White has centralised his pieces and is ready to attack.
Byrne-Larsen (Biel 1976) continued: }
15... e5 16. fe5 de5 17. Rf1 Nd7 18. Ng3 Nb6 19. Nf5 f6 20. Qc5
{ with White advantage }

and indeed my database gives 7. Qe2 57% for White, but 7.O-O only 52%

However, Black seems to have done better with 8...Nf6 or 8...Nge7,
rather than 8...Bxe3. Maybe these moves has revived Polu's line ?
Disclaimer: I'm not a Kan man or a Paulsen person, so feel
free to disregard...

--

Regards, Peter

Peter Billam, DPIWE/CIT/Servers, hbt/lnd/l8, 6233 3061

Date: 29 Jun 2005 05:48:01
From: Philip Feeley
Subject: Re: Hydra v Adams

Claus-J�rgen Heigl wrote:
> Peter Billam wrote:
>
>> P.S. Is there a published analysis of the last game ? I know that
>> 1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 a6 5. Bd3 Bc5 6. Nb3 Ba7
>> was at one time considered a bit dubious for Black...
>
>
> I searched for it also and found nothing in the web. When I'm finished
> with game 4 I plan to do an analysis myself.
>
> Claus-Juergen

Chessgames.com lists 216 games with these moves. Replies for move 7 were
Qe2, 0-0, Nc3, c4 or Qg4. I don't know how many were analysed.

Phil

Date: 29 Jun 2005 16:25:43
From: =?ISO-8859-1?Q?Claus-J=FCrgen_Heigl?=
Subject: Re: Hydra v Adams

Philip Feeley wrote:
> Claus-J=FCrgen Heigl wrote:
>> Peter Billam wrote:
>>
>>> P.S. Is there a published analysis of the last game ? I know that
>>> 1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 a6 5. Bd3 Bc5 6. Nb3 Ba7
>>> was at one time considered a bit dubious for Black...
>>
>> I searched for it also and found nothing in the web. When I'm finished=
=20
>> with game 4 I plan to do an analysis myself.

> Chessgames.com lists 216 games with these moves. Replies for move 7 wer=
e=20
> Qe2, 0-0, Nc3, c4 or Qg4. I don't know how many were analysed.

True. Perhaps I was not clear enough that I meant to not have found an=20
analysis of the match Hydra - Adams, game 6.

Claus-Juergen

Date: 29 Jun 2005 15:35:26
From: Peter Billam
Subject: Re: Hydra v Adams