Author: Bruce Moreland
Date: 11:08:13 06/15/98
Go up one level in this thread
On June 15, 1998 at 10:37:06, Don Dailey wrote: >You are always implying that you have some special insights and >that others are in some kind of conservative rut. But all your >ideas have been on an extremely high level of abstraction. Saying >"use more knowledge" or "make your program search fewer nodes" >is not a very coherent plan, and reveals no insight whatsover. I think he is right to say that programs need to learn to play more dynamically. But yes, he surrounds this with a lot of language that detracts from the issues. >You often use the time honored technique of being critical and >pointing to the problems and saying it loud enough that no one >notices you are not presenting any solutions. This makes it >appear that you are privy to a better way when in fact you are >not. > >But you have said nothing we do not already know. Every >programmer on this group knows what the most serious problems >of computer chess are. And each one of us is keenly interested >in solving them. Stop attacking us as "materialists" which is >a gross overstatement of what we actually do. We are simply >engineers and will always be writing the strongest programs >because we will always use the techniques that work best. What Thorsten wants is dynamism. The way he pushes for this is confusing, because he attaches political and personal baggage to this, and is wedded to some perceived attributes of specific implementation, and deathly opposed to some others. When I see phrases such as "techniques that work best" I think of the argument between brute force and whatever the AI people have invented. I don't think that what is going on here, I think at the root of all of this it's just an argument about playing style, and that he'd be happy with a program that played more dynamic chess, especially if you let him jump to erroneous conclusions about how much knowledge the program really contains. When everything else is stripped away or waded through, I think he has a point. Chess programs haven't been very dynamic, they play like oatmeal. This has been something of a blind spot for us, because and oatmeal program won't be able to take advantage of another program that is playing like oatmeal. I think that the more dynamic programs are more fun for sure, and that to a degree that is probably getting larger, more dynamic play can result in increased strength against other computers and against humans. bruce
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.