Author: Walter Faxon
Date: 02:20:03 01/05/06
Go up one level in this thread
On January 05, 2006 at 04:00:45, Uri Blass wrote: >On January 05, 2006 at 03:54:18, Walter Faxon wrote: > ...snip... >>Also IIRC, Hsu wrote that adding a term for "the initiative" hurt Deep Blue in >>testing. A good search already shows you what any initiative can do. > >I disagree. > >Hsu probably did not evaluate initiative correctly. > >A good search cannot show you everything. >There are leaf positions that you need to evaluate. > >Uri Hello, Uri. You evaluate initiative in movei? Enlighten us! Initiative appears to be a dynamic feature; you need search to demonstrate it. But the search itself shows you what the initiative could achieve. By additionally rewarding "initiative" along the way you would be doubly-rewarding certain lines, which would lead to evaluation instability, I suspect. If you could invent a static method of measuring "initiative" at terminal nodes -- a method independent of mobility -- that would be a big step forward. Please take some time to consider this and give us your thoughts. Respectfully, Walter
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.