Author: Uri Blass
Date: 17:11:57 12/31/02
Go up one level in this thread
On December 31, 2002 at 19:35:40, Antonio Dieguez wrote: >On December 31, 2002 at 19:30:52, Antonio Dieguez wrote: > >>On December 31, 2002 at 19:22:26, Uri Blass wrote: >> >>>On December 31, 2002 at 18:59:07, Antonio Dieguez wrote: >>> >>>>On December 31, 2002 at 17:49:52, Uri Blass wrote: >>>> >>>>>From Ed programmer stuff >>>>> >>>>>Killer-One [current ply] 110 >>>>>Killer-One [current ply-2] 108 >>>>>Killer-Two [current ply] 106 >>>>>Killer-Two [current ply-2] 104 >>>>> >>>>>I until today used only >>>>> >>>>>Killer-One [current ply] >>>>>Killer-Two [current ply] >>>>> >>>>>I am interested to know if using 4 killers is a new idea or maybe this idea is >>>>>known to be used by other programs. >>>> >>>>It is obviusly not a new idea, but a candidate to new idea is using exactly >>>>those 4 killers, and the exact details of the rest of Rebel ordering. >>>> >>>>I use >>>> >>>>Killer-One [current ply] >>>>Killer-Two [current ply] >>>>Killer-Three [current ply] >>>>Killer-Four [current ply] >>>>Killer-One [current ply+2] >>> >>>You say that you use ply+2 and not ply-2 >> >>yes. >> >>> or maybe you say that you the first 4 >>>are like Ed as "exactly those 4 killers suggest. >> >>I don't understand this but surely not. >> >>>>When a killer causes a cut-off it win two places. >>>>When a killer is the best move (with exact score) it win 3 places. >>> >>>Do you say that a killer that give something above alpha but not above beta >>>win 3 places in case that you find nothing better than it? >> >>yes. >> >>>>Also when the recommended move of the hashtable causes inmediate cutoff it win 3 >>>>places. >>> >>>I thought that every time a move cause cutoff it is considered as the first >>>killer move for that ply and the question if it is because of hash or because of >>>other reasons is not relevant. >> >>I though the hash was something apart. >>And as I see in Ed's page, he has it apart too, with 127 points. > >oops I didn't answer well. >Forget those two sentences. > >my answer is that at least for me it is relevant. > >And when a cutoff ocurres it is not so obvius that promote to the first place or >even second is the best. You may be right and I agree that it may not be best to promote a move that cause a cutoff every time to the first place but the original idea of killer move is to promote killers always to the first place(they can go down only one place but go up only to the first place). I understand that if a move is not in the list of the first 5 move then promoting 3 places mean going to the 3th place and there are basically 6 places and every new killer move that is not in the best 5 is ranked 6th. I have ideas that I did not try about move ordering of killer moves but this is not one of them so I may try some combination of my ideas with this idea. Uri
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.