Author: John Coffey
Date: 14:07:08 09/28/98
Go up one level in this thread
The gracious response I got earlier from Dr. Hyatt would seem to answer the question both ways ... -------------------------------------------------------------------------- when talking about material, we are testing the side that is *up* in material, and saying "if I give my opponent two moves in a row, can he recover the material he is down?" If the answer is yes, then we have to search this node normally. If the answer is no, then we say "two moves in a row didn't let him recover the material, so if I play a move in between his two best moves it is even *less* likely he can recover it, so we can stop searching this path. >I have asked repeatedly if we only do the null move check when up material, >and I have been told repeatedly that it is done everywhere. this is correct... -------------------------------------------------------------------------- And then the responses I got a few days ago confused me even more ..... -------------------------------------------------------------------------- Posted by Robert Hyatt on September 23, 1998 at 16:38:57: In Reply to: Re: Mr. Hyatt, could I get a clarification on this? posted by John Coffey on September 23, 1998 at 15:14:58: On September 23, 1998 at 15:14:58, John Coffey wrote: >On September 22, 1998 at 22:37:24, Dave Gomboc wrote: > >>On September 21, 1998 at 18:35:42, John Coffey wrote: >> >>> >>>>I try them _everywhere_ in the search, before trying any other move. The idea >>>>is that if your opponent can't take two moves in a row and crush you, your >>>>position is **overwhelming** and doesn't need any further searching to prove that >>>>it is probably winning... >>> >>>The "**" above are mine. >>> >>>So you only try them when ahead in material? I just needed to clarify >>>that because you also say "everywhere." >>> >>>Thanks a lot. >>> >>>John Coffey >> >>He really means **everywhere**. Actually, he might disable them in a few >>low-material endgame positions if tablebases aren't present. Or he might not. >>:) >> >>But yeah, basically **everywhere**. >> >>Dave Gomboc > > >What is confusing to me is that I thought that null-move search is only done >when ahead in material? Why do them when the material is even or when one >is behind in material? > >John Coffey Because it works even in non material-ahead positions. IE I would venture that I could beat almost any player in the world, if at some point in the game, I am allowed to make two moves in a row. Just once per game. Material ahead will obviously cause the null-move search to fail high if your opponent has no threats to regain that material. But it will also fail high if you have a serious positional edge that your opponent can't neutralize. Or if your opponent has a serious positional weakness that he can't neutralize with two moves in a row. It works *everywhere*. There is one short trick, written up by the Deep Blue guys several years ago... when you do a probe into the hash table, and you get a "hit" but the draft is too low, it still might be a good clue whether or not the null-move search will fail high or not, because you will be searching two plies less deeply. If the draft is enough to cover that, even though it can't match the current depth requirement, then you can use that hash entry to say "don't try a null-move here, it will fail low" and save some time doing so... Other than that, I do them *everywhere* except at the root, and not in the q-search at all of course.
This page took 0.01 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.