Website founded by
Milan Velimirović
in 2006
18:09 UTC
| |
MatPlus.Net  Forum Internet and Computing Gustav heuristic vs. Brute Force |
| |
|
|
|
|
You can only view this page!
| | | (1) Posted by Sarah Hornecker [Sunday, Nov 28, 2021 23:41] | Gustav heuristic vs. Brute Force
Are there any known cases where heuristic mode failed to catch a cook or solution on the solving software "Gustav" by Olaf Jenkner, assuming standard settings except looking for all move numbers (2 flight squares, all captures and promotions allowed, 0 threat depth, unknown move number)?
James had me test a selfmate and wanted to know about it. :-)
| | | (2) Posted by Olaf Jenkner [Tuesday, Nov 30, 2021 17:10] |
See
https://pdb.dieschwalbe.de/P1127241
If you give 4 flights, Gustav only finds authors solution.
The cook needs 5 flights.
| | | (3) Posted by Sarah Hornecker [Tuesday, Nov 30, 2021 18:18] |
Yes, examples with other settings are also welcome, obviously. Many thanks! So Brute Force indeed is necessary for testing 100 percent.
| | | (4) Posted by Olaf Jenkner [Tuesday, Nov 30, 2021 19:40] |
An example:
https://pdb.dieschwalbe.de/P1238025
Gustav finds the solution using automatically adjusted parameters within 3 seconds.
After each white move black has only one move. Otherwise the white move will not be considered.
The cook has 2.Sxd6! and black has a plenty of moves. It takes a lot of time to find (compute) such cooks, but they are quite often.
| |
No more posts |
MatPlus.Net Forum Internet and Computing Gustav heuristic vs. Brute Force |
|
|
|
ISC 2024
