Author: Normand M. Blais
Date: 14:54:47 01/08/03
Go up one level in this thread
On January 08, 2003 at 13:28:15, Charles Roberson wrote:
>
> NoonianChess uses the time given from the gui at each move. This makes
> time management much easier. Lots of time controls with and without
> move increments.
>
That is what I'm doing at the moment.
> From a quick look at a middlegame position, you can not ascertain time
> requirments.
>
> An algorithm that old versions of NoonianChess used was to check the
> time at each ply of the iterative deepening. If more than 50% of the
> estimated time per move has passed then make move otherwise go another
> depth. This is simple and works. It also allows going over the estimated
> time in some positions. With most effective branch factors <= 4, 50% allows
> the program to consume 2x the time. In the types of positions that this
> happens, it is probably good.
>
> My current algorithm is based on the fact that I search the PV first.
> Just let the program run until the time runs out. Check the time
> periodically but not too often during the search. When the time is out,
> set a variable and return. Whatever is at the top of the PV is the move.
>
> Charles
Your idea is welcome. Thank you.
NMB
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.