Author: Ratko V Tomic
Date: 21:23:39 11/01/99
Go up one level in this thread
>Illegality probably does not even add one bit. The reason is that >there are so few types of illegal positions since the kings next >to each other are handled. I think that's right. It probably isn't worth checking for the encoding purposes (other than simple stuff for kings+castle in 12 bits), unless you come a fraction of a bit over 160 and wish to squeze out every last air bubble. >> This goes against the grain, as it were, when one is in the >> "compressing" mind-set to make onself think up of ways to lenghten >>the encoding. > >I referenced this in my other post from a few minutes ago. The best way to >decrease the most bit intensive positions is to find algorithms that decrease >the number of bits for them while increasing the number of bits for bit >non-intensive positions (not necessarily an easy thing to do). The two SV >methods mentioned do this. It is funny that we both commented on the same against-the-grain aspect of your project at virtually the same moment and independently of each other (since you played with this problem for much longer, you probably have observed it earlier; for me, not having thought at all about this problem before, that observation struck me only as I was replying to Guido's post).
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.