Author: Michael Yee
Date: 16:42:32 08/22/05
Go up one level in this thread
On August 22, 2005 at 08:39:33, Dann Corbit wrote: >On August 21, 2005 at 20:00:31, Michael Yee wrote: > >>On August 21, 2005 at 19:46:45, Uri Blass wrote: >> >>>On August 21, 2005 at 19:38:08, Brian Richardson wrote: >>> >>>>On August 21, 2005 at 02:48:31, Ed Schröder wrote: >>>> >>>>>The silence in CCC is deafening... >>>>> >>>>>This is an historic WCCC, the first time a non commercial winning the WCCC since >>>>>1992, it seems their domination period is over. Not to speak about the micro >>>>>WCCC which has been dominated from the very beginning by the commercials. >>>>> >>>>>Congratulations Anthony, you did the impossible and so convincing. >>>>> >>>>>Ed >>>> >>>>Let me add my congratulations. It has only been a couple of years since Tinker >>>>was sometimes minimally competitive with Zappa. The tremendous improvements you >>>>have made in that short time are clear with Zappa's absolutely stunning WCCC05 >>>>performance. >>>> >>>>I would also like to acknowledge Fruit's very strong performance too. >>>> >>>>Brian >>> >>>I think that Fabien's achievement is clearly better than Anthony's >>>achievement. >>> >>>Fruit is younger than zappa and also free source code so Anthony could learn >>>from Fabien's code but not the opposite. >>> >>>Uri >> >> >>Well, I wouldn't go so far as to say "clearly better". Even though it's possible >>Anthony looked at other source code, it's not clear that he necessarily did. >> >>In fact, I think there are some programmers that enjoy the challenge of seeing >>how far they can go with their own implementation before looking at what others >>have done. > >True, but they are not very successful. >And each and every one is still looking at what others have done in some >respects or they are complete idiots. > >Is there anyone who does not use some form of alpha-beta? And did they invent >alpha-beta? How about null move? How about pvs search or MTD(f) search? > >Everyone is free to invent the wheel. It just isn't a smart way of doing >things. > >That's on the one hand. But I think that the new crop of expert programmers are >doing something that the earlier crop who tried to be competitive are doing >something different. They are actually understanding the new methods. The old >method was to steal a strong program, make a few tweaks with a screwdriver, slap >your name on it, and enter it in a contest. Then, get caught and skulk away in >shame. > >The new method includes understanding the fundamental ideas and incorporating >the new ideas into your program as fundamental principles. There is also >innovation. Fabian has some fresh ideas. Anthony Cozzie once said that the >best program is the one that incorporates the most tricks, but I expect that he >has done some real innovation to succeed as well as he has. I agree with you 99.9% (give or take a few percent :). The only thing I would add is that there's some line between algorithms and implementations of algorithms that some people aren't comfortable crossing (e.g., Reinhard). For myself, part of the challenge would be coming up with correct and efficient implementations of known algorithms first, in addition to experimenting with new ideas. But I guess an argument could be made that pseudo-code and compilable code (which obviously has more/all details worked out) are ultimately very similar. Michael
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.