Author: guy haworth
Date: 01:38:28 09/15/99
Go up one level in this thread
L.B.Stiller's 6-man endgame results were not retained in files as far as I know.
His technique was to use the massive primary memory of the machine.
His thesis includes for the ~36 profiles he studied
the number of positions,
the % of wins for White, the stronger and attacking side as is usual
the maximum depth (to a won conversion)
the number of zugzwangs
I would have thought there was some use of disc for back-up and analysis at the
time but the files were certainly not retained. The machine was a SIMD
Connection Machine, similar in architecture to ICL's Distributed Array
Processor.
Assuming his results are confirmed, LBS discovered the deepest forced wins so
far known. They were in the KRBKNN (maxdepth 223m) and KRNKNN endings (maxdepth
243m), the first of which featured in Scientific American.
From an 'infallible' computer's point of view, these endings - despite their
depth and the FIDE 50m rule - would be endings to play for against a fallible
opponent, provided a theoretically won position could be arrived at.
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.