Author: Eric Oldre
Date: 07:36:40 07/21/04
Go up one level in this thread
On July 21, 2004 at 07:56:17, Dan Honeycutt wrote: >On July 21, 2004 at 05:43:57, Aivaras Juzvikas wrote: > >>On July 21, 2004 at 05:38:14, Uri Blass wrote: >> >>>On July 21, 2004 at 05:33:23, Aivaras Juzvikas wrote: >>> >>>>On July 21, 2004 at 05:27:31, Uri Blass wrote: >>>> >>>>>On July 21, 2004 at 03:18:02, Aivaras Juzvikas wrote: >>>>> >>>>>>has anybody got any good ideas what to do with the time saved, when you have >>>>>>only one legal move at root position and instead of launching a search, you just >>>>>>make that move instantly. >>>>>>meybe it would be wise to add the saved time to the next move so you search it >>>>>>for twice as long as you would normally. >>>>>>im askin this question because its very hard to test it (it rarely happens). >>>>> >>>>>I do not understand your problem. >>>>> >>>>>Every move you get a new time control and decides about the time management >>>>>based on the new time control and the position. >>>>> >>>>>If you had 34 seconds for 4 moves and played a move instantly then now your new >>>>>time control is 34 seconds for 3 moves. >>>>> >>>>>Using constant time for moves is also not logical and it is important to use >>>>>more time after fail low. >>>>> >>>>>I also do not use constant time per move for other reasons and I try to stop in >>>>>most cases at the end of the iteration and the idea is that at the end of the >>>>>iteration the program cannot change it's mind quickly because it needs to search >>>>>the move that it plans to play first so if the program needs a long time to >>>>>change it's mind then it is good to play immediately and not waste time. >>>>> >>>>>Uri >>>> >>>>correct me if im wrong but having only 1 single legal move is usually because >>>>you're in check, thats a threat and more time in this case would certainly help. >>>>if its not a threat then oh well, its not like you lose much. >>> >>>It is usually because the side to move is in check but I do not see a reason >>>that it is espacially important to use time in the move after it. >>> >>>The rule should be to use more time when you believe there is a good chance to >>>change your mind and to use less time when you are almost sure that you will not >>>change your mind. >> >>good point, however i dont see how it could hurt if one did what im proposing >>here. > > >I do. You want to allocate the extra time on the next move for a threat that >may or may not exist. Better to allocate it, like Uri says, for a fail low when >you know you have trouble. > >Dan H. Dan, I'm not sure I know what use mean by a fail-low in this context. Does that mean a fail low on an aspiration window? Eric
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.