Author: blass uri
Date: 19:18:09 11/23/98
Go up one level in this thread
On November 23, 1998 at 20:13:16, Robert Hyatt wrote: >On November 23, 1998 at 18:47:20, Amir Ban wrote: > >>On November 23, 1998 at 17:14:28, Robert Hyatt wrote: >> >>>On November 23, 1998 at 11:50:01, Amir Ban wrote: >>> >> >>>I'm going to try to keep this simple. Here is a point-blank question: if you >>>really believe that nonsensical statement you wrote above, then how can you >>>reconcile that with a program that is searching at least 1,000 times faster >>>than you, yet only gets to depth 10-11 in the game? If they are not extending >>>far more than you could ever hope to then exactly *what* are they doing with >>>that factor of 1,000? And remember that they have a pretty simple quiescence >>>search and they toss out bummer captures as well, so the work is *not* in >>>looking at zillions of captures. >> >>Seems like you misunderstand another side of this, because you are mistaking me >>for a null-mover. >> >>To get to nominal ply 11, I would have to search at least to my (half-ply) depth >>21. If you've ever seen Junior play, you know that that's a lot for a middlegame >>position. I'm guessing it will take Junior at least a few days to get there. >> >>When I do get to that depth, though, I expect to see much more than Deep Blue >>sees at its ply 11, because I'm much more heavily extended, and of that, I'm >>sure. >> >>Amir > >That's my question, however.. They are *only* getting to ply 11. I can get to >ply 10 with crafty with null-move *off*. searching under 300K nodes per second. >So what do you suppose *they* are doing with that factor of 1,000 nodes? Must >be something *very* serious... IE I saw *Socrates "outsearch" deep thought by >a ply or two in Cape May (if you look at depth reported on the screen). And >Deep Blue isn't going any deeper today it seems. So those nodes are going >*somewhere* and the only place I can figure is extensions... They have always >been liberal (ie they allow 1 ply of extensions for *every* ply in the search >if you read their JICCA paper. You can't go beyond that and have a search that >terminates... and even 1:1 is on the ragged edge... > >That was my only point here... that they follow some things *very* deeply. >I sometimes see 24 ply variations from Crafty on a 12 ply search. But when >I watched Deep Thought I saw many 30+ ply variations in tactical cases, and >that doesn't touch the last 4-5 plies + captures that don't get displayed... Junior5 also can search very deep at some lines. I saw Junior5 display a 19 ply variation on ply 6 search(I translate depth 12 of Junior5 to ply 6 search). When Junior complete a 12 ply search(at depth 24) and it can take days it also calculate 30+ ply variations in tactical cases. Uri
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.