Author: Don Dailey
Date: 15:42:30 01/13/98
Go up one level in this thread
On January 13, 1998 at 17:22:50, Jay Scott wrote: >On January 13, 1998 at 16:02:53, Don Dailey wrote: >>The reason such an obviously good algorithm like square of pawn is not >>as big a win as you might intuitively think is that > >Is that... ? Is that by the time they're near the root the search >can solve them anyway? I'm not sure this is true for KN-K. :-) >I'm on the edge of my seat! > > Jay Hi Jay! I started to add this to my post but decided against it. I forgot to clear it. But since I started it .... The reason such an obviously good algorithm (square of pawn) is not as big a win as you might intuitvely think is that a whole lot of the time your program is playing correctly anyway. For instance you have a pawn outside the square and what will you do? Most programs will start pushing the thing anyway. But lest I get blasted I want to make it clear I strongly believe in the power of this algorithm. I wish we had more like it. If you construct it correctly it is right almost all the time and can save you many plies of searching. I view it more like the king and pawn database, it's real benefit is to help you get into the right ending, less so to play it correctly once you have it. Computers seem to have good pretty good technique at this point, the real problem is some critical decision point that requires them to go this way, or that. Most of the time they do fine if this single decision is correct. The decsion rarely leads to a quick tactical loss or crushing combination. - Don
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.