Author: Uri Blass
Date: 19:21:34 09/16/00
Go up one level in this thread
On September 16, 2000 at 20:13:46, Dann Corbit wrote: >On September 16, 2000 at 03:59:26, Uri Blass wrote: > >>On September 15, 2000 at 21:04:53, Dann Corbit wrote: >> >>>On September 15, 2000 at 17:57:47, Gian-Carlo Pascutto wrote: >>> >>>>On September 15, 2000 at 17:13:57, Dann Corbit wrote: >>>> >>>>>Sorry. Only 24 coming. This is only solved to 15 ply. Checkmate or not, you >>>>>must complete the ply to claim the "prize". >>>>>;-) >>>> >>>>Quote: >>>> >>>>Here are a set of tough positions to search deeply. Just finding a mate is not >>>>good enough, uless you can *prove* it is the shortest mate. >>>> >>>>>1.f6+ Kh8 2.Qh6 axb2+ 3.Kb1 Ne6 4.dxe6 Qxa2+ 5.Nxa2 Rg8 6.hxg6 >>>>> +- (#9) depth: 15/21 00:00:18 436kN >>>> >>>>If you want a _proof_ I guess CHEST is the only option. >>>> >>>>Now, before you start writing those poems realize that this >>>>analysis is done by a Solution-Tree-Cost search and *not* by >>>>an alpha-beta searcher. >>>> >>>>The ply depths listed are totally uncomparable to alpha-beta >>>>ply depths. >>>> >>>>If you think this is cheating, consider that that measuring >>>>ply depth, via any algorithm at all, is a bogus measure. >>> >>>So! I've been bamboozled. I think that pruning that leads to the same >>>solutions is valid. Every sort of pruning will throw away some correct >>>solutions. But NULL move is only very rarely going to lead to problems. >> >>I believe that recursive null move may often lead to problems. >> >>My guess is that if you compare Crafty's result after 10 plies without null move >>pruning and with null move pruning you will find at least 10% different moves(I >>think it will be interesting to do this experiment) >> >>Take some games of Crafty tell it to search all position to 10 plies without >>null move pruning and with null move pruning and see the number of different >>moves in 1 ply,2 plies....10 plies. >> >>Another interesting experiment will be to play fixed depth matches(x plies of >>the regular Crafty against x-1 plies without pruning) >>I am not sure which side is going to win and my guess is that when x is big >>enough x-1 is going to win. >> >>I do not say that recursive null move is counter productive but there is a >>significant price that you pay for searching deeper. >> >>The price is smaller than the advantage and this is the reason that programmers >>use null move. > >I wish someone would do this experiment, since I think it is an important idea. >Unfortunately, unless someone can find a better way to search faster I think >people will continue to use it anyway, even if it is shown to be clearly >harmful. The reason is that as long as the deeper search provides more value >than the missed possibilities, they are going to use it anyway. > >I have always thought it funny that the cost to ignore is exactly one tempo. >Does that seem strange to anyone else? Why one tempo and not some other cost? > >Consider also when I have 200 possible choices. Maybe 100 of them are just >plain stupid. But far less than that will be affected by null move. > >Consider when I only have 3 possible choices to move. One of them might put me >in a zugzwang position but afterward lead to the only successful solution. The most common problem with the null move is not zunzwang but cases when there is a threat and the threat is too deep to see because null movers analyze threats to smaller depth. programs can prune moves when it search 16 plies because there is no threat of 12 plies inspite of the fact that there is a threat of 13 plies. The situation may be even worse when programs use recursive null move pruning. 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.