Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Rebel Shows GM strength once AGAIN(draws Baburin)

Author: Stephen A. Boak

Date: 17:26:41 12/05/99

Go up one level in this thread


On December 05, 1999 at 18:53:32, James Robertson wrote:

>On December 05, 1999 at 15:14:08, Peter Hegger wrote:
>
>[snip]
>
>>I think a fair ratio would be 2150 positionally and 2750 tactically. To say they
>>are only 1600 positionally is absurd in my opinion.
>
>[snip]
>
>Why is there even this discussion? You guys are just picking numbers out of the
>air.
>
>James

Lighten up, James!  Abstract thinking (including formation of opinions) based on
concepts not reduced to concrete calculations is the fertile background for
creative ideas that eventually may bear fruit.  Don't stifle the free thinking,
please!

Actual program improvements start with ideas, even the unproven ones.  They
start with judgements, even the unfounded ones.  They start with the free
exchange of ideas, both the good and bad, the empirically deduced as well as the
gut feelings.

When a programmer creates or improves a program, doesn't he often judge (not
calculate) in what aspect the program is weakest and then spend his time trying
to improve that aspect--even if the surmise is just 'picked out of the air'
without exact calculation but only gut feel.

For example, does the programmer have to know whether the lack of coding
handling Bishops of opposite color endings or the lack of coding handling KQP vs
KQ endings (before tablebases, that is!) has caused his program exactly 50 lost
rating points or 60 lost rating points, in 2% or 13% of its games, in order to
determine which is worth pursuing next?

No, he only needs to have an opinion on how the program is weak and a feeling as
to whether the coding effort will be worth it (to him, the programmer) to tackle
right now.

Even the most sophisticated of chess programmers started with simplistic ideas
in their attempts to understand chess and chess programming on a computer.
Don't try to stop basic discussion even if you hope more for the Hyatt, Theron,
Ban, and Schroder type expert programmer input.  CCC readers include a broad
gamut of types and interests, from the beginner programmer to the sophisticate.

Have a heart, man.  There are many more threads in this forum worth complaining
about--that waste more bandwidth and don't contribute even a single opinion to
the computer chess subject.

--Steve Boak



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.