|
< Earlier Kibitzing · PAGE 1 OF 43 ·
Later Kibitzing> |
| Sep-11-04 | | fgh: Fear this monster!!! |
|
Sep-11-04
 | | chessgames.com: We're missing one of its victories over Shredder--if anybody has the score to that game please submit it via the PGN Upload Utility. Thanks. |
|
| Sep-11-04 | | bunti: Does anyone know if hydra is a program that will become commercially available for personal computers. |
|
| Sep-12-04 | | snowfalcon: Hydra beat the top ranked Shredder (2818 on the April SSDF rating list) 5.5 to 2.5! It didn't lose a single game. It also didn't lose a single game in beating all the worlds best at the 2004 Paderborn computer championships - it scored 6.5 out of 7. I'd like to see it take on Kasparov after they upgrade their hardware at the end of the year. |
|
| Sep-12-04 | | Lawrence: <bunti>, Hydra was running on 16 processors when it beat Shredder, which was running on "only" 4. It'll be some time yet before any of us have a computer with 16 processors. I have a hunch that Hydra on one processor would be no better than Fritz, Shredder, Junior, etc. and perhaps worse. |
|
| Sep-12-04 | | yoozum: i'm praying for the very unlikely event of anand/kasparov/kramnik playing hydra in another man vs. machine championship. |
|
| Sep-12-04 | | fgh: <Lawrence>, Hydra was running on 16 processors and Shredder on 4, but the deepety of the calculation doens't increase by the number of processors. |
|
| Sep-12-04 | | fgh: For example, if we have Hydra on 12 processors playing against Hydra on 14 processors, it's search isn't increased by 24 ply, but more like 1 ply. |
|
| Sep-13-04 | | Lawrence: <fgh>, could you explain that again? If moving from 12 processors to 14 increases the depth of search by 1 ply, wouldn't moving it from 4 to 16 increase it by, say, 6 ply? |
|
| Sep-13-04 | | percyblakeney: The overall record given for Hydra at the moment is +3 -1 =4, but I guess it really should be +4 -0 =4. |
|
| Sep-13-04 | | snowfalcon: percyblakeney - you are correct - in fact, I have not been able to find any game that Hydra has lost (to a computer or a human). If the funding behind this monster continues, and the hard-working team keeps improving it, it might very well become the first true 2900 elo "player" in history. |
|
| Sep-13-04 | | Lawrence: <snowfalcon>, welcome to this amazing site. Nice name. Are you an enginephile? If so, which one do you have? |
|
| Sep-13-04 | | snowfalcon: <Lawrence> Thanks for the welcome and kind words. Yes, I love computer chess programs and play Chessmaster 10th edition and Fritz 8 the most.I admire all the current top chess players and can't get enough of man vs. machine contests. I think in the next ten years we will start to see top players get destroyed by machines. (Hydra just beat up a 2628 GM 3.5 to 0.5) As Kasparov himself said - in the future it will be about if we can win one game. I just hope to be alive to see many of the contests. |
|
| Sep-13-04 | | snowfalcon: I finally found one loss by Hydra this year - it was running on its 16 processors. It lost to Shredder in April at the 4th Int'l CSVN Tournament in the Netherlands in 31 moves. I will try to get the game posted. It looks like Team Hydra learned a lot from that loss and improved the program so that it did not happen again in its eight-game match with Shredder in August. |
|
| Sep-13-04 | | fgh: <Lawrence>, I cannot say if the difference between Hydra on 16 and 4 processors is 6 ply or more, what I'am saying is that to reach one more ply, the program will need to calculate always more and more positions. |
|
| Sep-13-04 | | clocked: <Lawrence> your logic assumes the problem is linear (which it is not) |
|
| Sep-13-04 | | John Doe: But it's still running on a better machine, which is unfair... What is it, about Hydra' engine, that makes it better than other search algorithms? Does it have better heuristics or something? |
|
| Sep-14-04 | | Lawrence: <John Doe>, that's what I'm trying to figure out. As <fgh> and <clocked> rightly point out, 16 processors are not 4 times better than 4 processors, (but they must be better to a certain extent). I'm still not convinced that Hydra's engine is better than Shredder's: I'd like to see what happens when both engines run on the same number of processors. |
|
| Sep-18-04 | | John Doe: Well it is possible that Hydra, because of its greater positional understanding, could beat Shredder... But all the heuristics slow the computer way down. I bet its search engine is the same. With limited processing capacity I bet Shredder would win. However, if they are both runnning on a large number of processors, (maybe 20), the difference in speed might not overcome Hydra's better positional understanding. |
|
| Sep-18-04 | | Lawrence: <John Doe>, but couldn't Hydra's better positional understanding be attributed simply to the 1 or 2 or 3 (or whatever) ply advantage it has over Shredder due to its extra processors? |
|
| Sep-18-04 | | LOUDERMILK: A computer having multi processors would not nececarily make it a better chess playing machine, but will only make a comuter think faster. It doesn't have the advantage as far as understanding a possition. |
|
| Sep-18-04 | | John Doe: No, extra ply gives a tactical advantage. However, since chess is of now unsolvable, computers have to rely on heuristics also. |
|
| Sep-19-04 | | Lawrence: Heuristics--isn't that the same as what Edison used to call "cut-and-try"? (Just keep mucking around until you find something that works.) |
|
| Sep-22-04 | | fgh: Pretty good winning percentage, I say. |
|
| Sep-23-04 | | acirce: <Hydra is in the team of machines that are going to play against the team of humans which includes Ruslan Ponomariov, Veselin Topalov and Sergey Karjakin where as Hydra has Junior and Fritz in its team. the tournament is one of the event of International Chess Festival in Bilbao, Spain starting from 6th October 2004.> http://www.hydrachess.com/
http://www.ajedrezbilbao.com/Festiv... |
|
 |
|
|
< Earlier Kibitzing · PAGE 1 OF 43 ·
Later Kibitzing> |
Kibitzing
