Author: Christophe Theron
Date: 09:48:09 04/21/98
Go up one level in this thread
On April 20, 1998 at 16:23:19, Thorsten Czub wrote:
>On April 20, 1998 at 15:33:57, Christophe Theron wrote:
>>Any volunteer to help the understanding of this topic? Thorsten?
>>Fernando? Enrique? Torsten? ... ... ...
>>
>>
>> Christophe
>
>Still I believe you make a mistake relating genius depths.
>
>My old Richard Lang program ROMA is often only doing M2 searches.
>This cannot be real !
>
>It has an 68000 CPU running 12 Mhz and I cannot believe that this
>program (assmbler) is only doing brute-force 2 plies !!
>Also with M1 Sn it does a better search and has huge main lines.
>No -
>Genius uses the Mx Sy in a different way than other programs.
>
>M is search-brute-force
>S is search-depth with extensions/captures etc.
>but x only 2 ?????
>and the main line 5 or 7 or more plies and accurate ?
M=2 means the program tries to find a 2+4=6 plies PV. This can indeed
give you 5, 6 or even 7 plies main lines. Not strange.
>The Richard-Lang PC-software does not give us relevant material to find
>out about. IMO we can only study the 680x0 machines because here Richard
>is PURE and not hiding anything since it was so early nobody discussed
>it heavily, like now.
>
>IMO we should make a 40/120 game with my Mephisto ROMA and i write down
>main-lines, evaluations and search-depth and
>we try to replay with genius and you will see:
>the 2 values genius offers us (M and S) or 6/18 is not what e.g. The
>King offers with 3/7. It is not 6 plies (before extensions)/(after
>extension).
>
>Mephisto Roma was IMO the best positional program he ever had.
We already discussed about that, and I just can tell anything about old
dedicated machines as I dont have them.
But the fact that your Roma does only 2 plies brute force searches does
not seems strange to me. That means (maybe) 2 plies brute force followed
by 4 plies selective. I would say that it is equivalent, in quality, to
a 4-5 plies brute force search. I believe it fits what machines were
able to do at that time with slow processors, isn't it?
Maybe it did more selective plies than current Lang's program do, so it
is equivalent to a 5-6 force brute search.
Extensions apply both in the brute force and in the selective part. So
it's not strange to get lines longer than M+4. Especially in lines where
you find checks, recaptures, mate threats, ...
Christophe
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.