is Hydra Invincible ?..
Moderators: Harvey Williamson, Steve B, Watchman
Forum rules
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
is Hydra Invincible ?..
Ok, I'm going to start a new topic (the topic about the Dr. Donninger became burning), Terry said that Hydra isn´t Invincible, and for me at least , no human being can beat Hydra, and about the programs... Rybka or Zappa could gain some games, but not a match, I believe that nobody can beat Hydra. Zappa in 8 CPU was enough to beat Rybka, so Rybka have no chances vs. Hydra it runs on a 64CL and evaluates about 200,000,000 chess positions per second, so it can demolish any thing,
Hmmm, my 2 cents
I'd agree that no human could beat Hydra in a proper World Champ style match say over 20 games or so, but for that matter I also believe that Zappa, Hiarcs and Rybka would do the same if they were on half decent hardware.
The win of Zappa over Rybka was, in part anyway, due to the small number of games played. Statistically 10 games is absolutely nothing, it takes many, many thousands of games to prove that one program is stronger then another. I see practical examples of this in my testing of Colossus for Martin Bryant, often even up to as many as 300 games you think you've got a real breakthough only to find by the 1000 game mark the "improvement" has dwindled to almost nothing and may even be nothing but statistical "noise". The rest of the Zappa win was delivered by some very clever preparation and by thinking on their feet when it mattered.IMHO.
As for the number of positions analysed this is pretty much meaningless. Fritz 5.32 churns through 2.5-3 million positions per second on my machine, single core Hiarcs is maybe 300 thousand positions any match between the two would be rather one sided!. Not all positions analysed are equal, the search tree explodes exponentially and without using special search techniques you can waste probably 90% of your effort by analysing lines that any Patzer human would avoid easily.
Hydra is/was very interesting but chess won't yield to any current hardware from a brute force point of view. It's the guided search and evaluation from the software that is king and here Hydra is behind the Rybkas, Zappa's and Hiarcs.
As for any Rybka - Hydra match, assuming we could get it over a decent number of games and Rybka was on a octal or greater then I'd have a large wager on Rybka.......
Just my opinion!
I'd agree that no human could beat Hydra in a proper World Champ style match say over 20 games or so, but for that matter I also believe that Zappa, Hiarcs and Rybka would do the same if they were on half decent hardware.
The win of Zappa over Rybka was, in part anyway, due to the small number of games played. Statistically 10 games is absolutely nothing, it takes many, many thousands of games to prove that one program is stronger then another. I see practical examples of this in my testing of Colossus for Martin Bryant, often even up to as many as 300 games you think you've got a real breakthough only to find by the 1000 game mark the "improvement" has dwindled to almost nothing and may even be nothing but statistical "noise". The rest of the Zappa win was delivered by some very clever preparation and by thinking on their feet when it mattered.IMHO.
As for the number of positions analysed this is pretty much meaningless. Fritz 5.32 churns through 2.5-3 million positions per second on my machine, single core Hiarcs is maybe 300 thousand positions any match between the two would be rather one sided!. Not all positions analysed are equal, the search tree explodes exponentially and without using special search techniques you can waste probably 90% of your effort by analysing lines that any Patzer human would avoid easily.
Hydra is/was very interesting but chess won't yield to any current hardware from a brute force point of view. It's the guided search and evaluation from the software that is king and here Hydra is behind the Rybkas, Zappa's and Hiarcs.
As for any Rybka - Hydra match, assuming we could get it over a decent number of games and Rybka was on a octal or greater then I'd have a large wager on Rybka.......
Just my opinion!
- Steve B
- Site Admin
- Posts: 10146
- Joined: Sun Jul 29, 2007 10:02 am
- Location: New York City USofA
- Contact:
hi guys
as a preemptive warning
this is an interesting topic for the members and as long as it remains about computer chess it will of course be left untouched by moderator hands
HOWEVER...
IF any posts rear their ugly heads in the thread not about Hydra or other chess computer topics then they will IMMEDITALY be DELETED
Tiger Grenadier Regards(only dedicated computer ever released to contain a Donninger engine)
Steve
as a preemptive warning
this is an interesting topic for the members and as long as it remains about computer chess it will of course be left untouched by moderator hands
HOWEVER...
IF any posts rear their ugly heads in the thread not about Hydra or other chess computer topics then they will IMMEDITALY be DELETED
Tiger Grenadier Regards(only dedicated computer ever released to contain a Donninger engine)
Steve
Re: is Hydra Invincible ?..
Hi, "Invincible" implies 'unbeatable', in *any* game. This does'nt apply
to Hydra, as it's bound to drop points and half points along the way against
the likes of Rybka.
In any case, discussing Hydra is not to discuss chess 'programming',
because Hydra is a software/hardware combination, as I understand.
There can't be much interest in it, simply because it is not commercially
available, and can never be, except perhaps to the ultra rich.
What does Hydra do besides play chess? I'm surprised someone was
found to finance such a project. Where's the return on investment?
all the best
Larry
to Hydra, as it's bound to drop points and half points along the way against
the likes of Rybka.
In any case, discussing Hydra is not to discuss chess 'programming',
because Hydra is a software/hardware combination, as I understand.
There can't be much interest in it, simply because it is not commercially
available, and can never be, except perhaps to the ultra rich.
What does Hydra do besides play chess? I'm surprised someone was
found to finance such a project. Where's the return on investment?
all the best
Larry
Re: is Hydra Invincible ?..
I disagree with you.DarkSide wrote:Ok, I'm going to start a new topic (the topic about the Dr. Donninger became burning), Terry said that Hydra isn´t Invincible, and for me at least , no human being can beat Hydra, and about the programs... Rybka or Zappa could gain some games, but not a match, I believe that nobody can beat Hydra. Zappa in 8 CPU was enough to beat Rybka, so Rybka have no chances vs. Hydra it runs on a 64CL and evaluates about 200,000,000 chess positions per second, so it can demolish any thing,
The fact that as far as I know hydra does not play in the last year suggests that it is probably weaker.
I think hydra is going to lose in a match against rybka or zappa.
I am not impressed by the number of positions because hydra probably has problems to make an efficient search because of hardware designs
so I guess that the effective number of nodes is smaller(same problem was also with deep blue that simply could not use hash in the last plies that the hardware searched based on my memory)
Uri
- Peter Grayson
- Member
- Posts: 208
- Joined: Sat Aug 04, 2007 3:23 pm
- Location: South Wales, UK
Given the cost of the Hydra hardware it would not have added too much expense to the project to buy an 8 or 16 thread PC together with Hiarcs, Rybka and now Zappa and tested its performance against the commercial engines.
The dramatic step increase in playing strength that came with Rybka's appearance must have put it within touching distance of Hydra and most people in Hydra owners position would have been too curious as to just how close it was so running a private match must have been inevitable.
I suspect the odds are this has been done. In that context, the absence of a public challenge to one of the commercial engines gives good weight to Uri's comment and the probability the Hydra team already know the outcome otherwise it would have made sense to allay all of the speculation of which is stronger and the result would be made known to the chess computer world. Perhaps it is too close to call.
PeterG
The dramatic step increase in playing strength that came with Rybka's appearance must have put it within touching distance of Hydra and most people in Hydra owners position would have been too curious as to just how close it was so running a private match must have been inevitable.
I suspect the odds are this has been done. In that context, the absence of a public challenge to one of the commercial engines gives good weight to Uri's comment and the probability the Hydra team already know the outcome otherwise it would have made sense to allay all of the speculation of which is stronger and the result would be made known to the chess computer world. Perhaps it is too close to call.
PeterG
- Dylan Sharp
- Senior Member
- Posts: 2431
- Joined: Fri Aug 10, 2007 12:07 am
Well, for the same reasons I think Zappa would win a match against Hydra, even if Hydra was stronger.Wardy wrote:The rest of the Zappa win was delivered by some very clever preparation and by thinking on their feet when it mattered.IMHO.
And, Hydra is indeed invincible! If you can't play against Hydra you can't beat it!
There is not an engine invensible
There is not an engine invensible
- Terry McCracken
- Senior Member
- Posts: 1300
- Joined: Tue Jul 31, 2007 5:49 pm
With your last comment you hit the nail squarely on the head!Dylan Sharp wrote:Well, for the same reasons I think Zappa would win a match against Hydra, even if Hydra was stronger.Wardy wrote:The rest of the Zappa win was delivered by some very clever preparation and by thinking on their feet when it mattered.IMHO.
And, Hydra is indeed invincible! If you can't play against Hydra you can't beat it!
Come Out and Play; regards,
Terry
Hi , I think, which percentage in a chess game is determined by the tactics?.. small combinations (as capablanca used to say), or Mammoth combinations lik Tal played, I guess more than 60%, the games that are solved with pure strategy are very few compared with which they are solved with tactical reasons, most of the world-wide champions they have been great and successful by its aggressive and tactical game (Morphy, Steinitz, Lasker, Alekhine, Kasparov, Fishcer, Thal). In this line I can say that Hydra is a tactical demon, with algorithms that seem to be make to complicate the game and to put to the rival in a tactical fight in where it has the best arms and can beat to anyone. certainly Rybka or Zappa also Hiarcs (great a tactical one) can gain some games to Hydra, but I believe that in a Match Hydra is invincible. In addition to everything it is not necessary to forget that the Hydra project was advised by a Great Teacher (Lutz) so that Hydra does not have to be so bad in strategic positions. Lamentably Hydra has died and this can´t be corroborate, but I am sure that the reason by which Hydra does not play but is not by fear to lose or so that he is weak tha Zappa or Rybka , none, simply their owners consider a cost who no longer wish to do. to make play Hydra does not report gains to them and they only it that wished are to become famous defeating the champion of the world, them advance Fritz and everything finished
I think this last sentence is better but unfortunately we still do not know because Hydra will never face Rybka or zappa, so all the discussion we can have here belongs to the imagination and is not for real, like somebody said before "you have to just come out and play" and see if our imagination is right or wrong. Everything else is not true.
- turbojuice1122
- Senior Member
- Posts: 2315
- Joined: Thu Aug 23, 2007 9:11 pm