Author: Uri Blass
Date: 12:37:47 09/18/00
Go up one level in this thread
On September 18, 2000 at 15:06:58, Gian-Carlo Pascutto wrote: >On September 18, 2000 at 13:50:51, Uri Blass wrote: > >>I believe that big part of the results without dynamic null move pruning are >>different. >>The main problem is not zunzwangs but cases when the null move pruning prune >>line that has a deep threat. > >I don't think this is true at ply depths this large. >Even with the most aggressive pruning the lines still only go from >16->11 ply. If a threat doesn't show up in those eleven plies it >is nearly sure it can be avoided with an extra move. I agree about it but the problem is that it is not correct for small number of plies. It is possible that the program will prune lines like 1.xx yy 2.zz tt 3.ww aa 4.rr uu 5.ff because of the fact that 5.ff is not a threat of 3 plies when 5.ff is an important move because it is a threat of 5 plies and if you search 16 plies without pruning you will find that 5.ff is good. These arent >exactly zugzwang situations either, but very active ones. The more >active, the less the chance for a nullmove error.(sortof) > >In short, I think aggressive dynamic nullmove pruning is accetable >at these depths in these positions. If I thought it weren't, I wouldn't >have used it. I do not say that they are not productive. If you can search 16 plies with null move faster than 14 plies without null move and see usually at least the same things than they are productive. > >>I also believe that extensions are important and 15 plies with the normal >>extensions may be better for analysis than 16 plies when you tune down >>extensions and razoring. > >This is true. But reaching 15 ply with the standard search is damn HARD. > >I didn't tune down razoring btw but re-added it. Bob got rid of it in 16.18 >because the gain of 20-30% wasn't significant enough related to the risk. >Again, at ply depths this large things are different. > >>If you want good analysis of the position the best thing to do is to ask the >top programs to use a long time and not to goto big depth. > >Probably true. However Dan is right about the fact that extensions are >gambling. Sometimes they see it, sometimes they don't. Using a large >standard search always sees it, save the nullmove/razoring failures, which >are really small. Not extending is also gambling because you can miss a long tactical idea. I prefer the gambling of extensions and programmers know that extensions are productive. Chess is a tactical game and I am almost sure that 14 plies with extensions are better than 16 plies without them. 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.