Website founded by
Milan Velimirović
in 2006

21:25 UTC
ISC 2021



Remember me

Forgot your
Click here!
to create your account if you don't already have one.

Rating lists


MatPlus.Net Forum Internet and Computing Proof Games
You can only view this page!
(1) Posted by Geoff Foster [Tuesday, Jun 5, 2007 11:57]

Proof Games

I believe that some proof games have been created using a computer program. How is this possible? Surely there are too many possible positions? How many possible positions are there after 10 moves? After 20 moves?

Also, how do proof game solving programs work? Do they play backwards from the diagram position, attempting to reach the initial game position?
(Read Only)pid=851
(2) Posted by Hauke Reddmann [Thursday, Jun 7, 2007 10:50]

I doubt that they have been *created* (only *verified*).
Since I never tried to decompile NATCH :-) I can only assume
that they work just like a human - counting moves,
taking care for collisions...only more reliable.

Computers usually have a hard time with "meta" style reasoning.
Don't see offhand an example for that in the case of proof games.

(Read Only)pid=858
(3) Posted by Hans Gruber [Saturday, Jun 16, 2007 15:20]

Hi Geoff,

some years ago, Norbert Geissler and Gerd Wilts wrote a program to compute massacre proof-games.

Their main goal was so investigate whether there exists a unique proof-game leading to a 2-men position without additional condition. They increased the number of single moves step-by-step. (30 captures needed, but no capture possible in the first two plies. Thus: Starting with 32 single moves, then 33.) As far as I know, their study included the full analysis of 34 single moves, and they found no unique proof-game. But some quite nice by-products resulted from these analyses.

Hans Gruber
(Read Only)pid=896

No more posts

MatPlus.Net Forum Internet and Computing Proof Games