Author: Tony Werten
Date: 00:06:56 07/08/05
Go up one level in this thread
On July 07, 2005 at 06:49:04, Ed Schröder wrote: >On July 07, 2005 at 03:50:21, Alvaro Jose Povoa Cardoso wrote: > >>Hi, I'd like your opinion on a modification to the first reduction explained on >>your webpage. >>The reduction reduces a full ply if we are behind alpha by a certain amount. >>Also you make sure you aply this only once in every possible line in the tree. >>What I propose is instead of a full ply reduction to reduce say 1/16th of a ply >>and alow it to kick in more than once in every line in the tree. >>If we do it like you explain and we are at a line with a sacrifice the engine >>will reduce a full ply and might not get it. >>But if we do it like I explained we still make a reduction on a line with a >>sacrifice but it will be a 1/16th reduction and the engine might see the >>sacrifice is sound. >>Now I could be wrong, that's why I would like your thoughts on this. >> >>Note: I know you use 1/4 of a ply as the smallest fraction possible to do >>reductions/extensions, so you might need to expand this to 1/16th or higher. > > >I have used this technique in the past, it even has a name -> fractional >extensions or in this case fractional reductions. It's my opinion that it gains >nothing. But note that others might disagree. Probably because you don't do fractional extensions either (or hardly) or other fractional reductions. Without those, reducing 1/16 of a ply x times (1<=x<=16) is exactly the same as reducing 1 full ply once. Well, except for hashtable replacements that is. Cheers, Tony > >Good luck with your engine. > >Ed
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.