[Retros] Fwd: [itaprob] computers

Richard Stanley rstan at math.mit.edu
Thu Apr 17 21:58:40 EDT 2008


> someone just wrote to me:


> > There are many chess problems

> > that are totally impossible for computers, but human can solve those.

> > Seeing beutiful things are not possible for computers, but human

> > can realize those - by seeing brightness/light. You cannot make as

> > complicated computer-program as your brains are.


> is it true ?

Good examples of such problems are "fortress problems."

W. Harston, New Scientist, 1999


White to play and draw

A reasonably good chess player can solve this problem in a few seconds
- just refuse to capture any Black pieces, and Black cannot break
through. However, the computer program Deep Thought, at that time
(1993) the world's best chess program, played 1.bxa5. This position is
discussed by Jane Seymore and David Norwood on page 25 of New
Scientist 139, #1889 (1993) in an article about the psychology and
intellectual abilities of good chess players, and by Roger Penrose on
page 104 of "The Large, the Small, and the Human Mind" in an essay
about the relationship of the mind to the physical world. Tim Krabbe
gives an example in item 223 of his "Open Chess Diary" of a fortress
position that occurred in an actual game played in 1951 which the
chess computer Shredder 7.04 had no difficulty analyzing.

Another fortress problem that requires a little foresight to set up

Anonymous, Zadachi i Etiudi, 1928


White to play and draw

I wonder how present-day computers would do on this problem.


More information about the Retros mailing list