Author: ALI MIRAFZALI
Date: 08:48:15 12/03/01
Go up one level in this thread
On December 03, 2001 at 11:21:21, Gian-Carlo Pascutto wrote: >And now, for a point-by-point rebuttal: > >>First I like to say that Phalanx is much waeker than CM8000. > >What's much? Don't underestimate Phalanx. It's one of the strongest >free engines out there, and one of the 'Young Talents'. As far as >sacrificial attacking play is concerned, it will _certainly_ hold >its own. > >>Secondly I said chess as pratically solved ;which is a far cry from from it >>being actually solved. > >Blech. Don't use words or terms which will confuse the issue because >different people interpret them differently. > >>Thirdly the analysis by Phalanx proves nothing. > >Because it contradicts your point I presume? > >>At the slowest time control which is 40/2 the program has only an average of 3 >>minutes to see that Nxg6 actually wins . > >If a program spots a score change, it can and will use more time as needed. > >>It should have at least a plus 1 pawn evaluation or more after 3 min. > >Why? Why?? Why not a 1.1541*sqrt(7) advantage? > >>Also the full analysis of why Nxg6 is winning is 19 plys deep. > >Possible, I didn't check. > >>Even DeepBlue could not not reach 19 plys in 3 minutes. > >Wrong. On two counts. And you lose another time by bringing >up DeepBlue which is always a bad idea. > >DeepBlue reached 17-18 _nominal_ ply in middlegame nearly always, >so it's certainly possible it reached 19 ply in some positions. > >Secondly, there exist such a thing as extensions. You do not need >a 19 nominal ply search to see a 19 ply combination. Fritz can >easily reach 30 ply with extensions in less than a minute. > >>Fourthly you claim if Phalanx did reach a certain point if would play e6 >>and not exd6 but you offer no proof. > >Gee, very hard to crosscheck it. The program is freely available. > >>The point here is that programs DONOT understand the Position after 19.Nxg6. > >'Understand' > >Yes, this is the word to use so nobody can ever prove >you wrong. I always love it when someone uses it in this context. > >Programs dont understand anything. They count beans. They count enough >beans to outplay 99.99% of those who _do_ understand chess. > >>And hence do not see the winng lines all the way to the end . >>And this DOES MATTER. > >Sure honey. > >-- >GCP Counter rebuttal Where did you get the information on Deep Blue? According to the programmers it was 10 to 14 plys in most middle game positions Not 17.By "understand I mean seeing (calculating for a Computer that is the equivalent). that Nxg6 is winning.Phalanx could not come up with e6 in atournament game Simply because of the Depth required for seeing that e6 is good.You say that that Fritz can search 30 plies with extensions .I am very well aware of extensions.However in this case at least it does not lead to the expected results.Ironically it is Fritz to solve this position .It does not even come close
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.