BPP against NP
Posted: Mon Jul 23, 2012 12:37 am
A nice endgame position by John Daniel Bryant. This is probably very hard for engines because the longest winning line has almost 50 moves without capture:
[D]2n5/7k/5B1p/2K4P/6P1/8/8/8 w - - 0 1
White wins
Playing only DTM-optimal moves after 36(!) moves the position will look like this:
[D]2n5/7k/5B1p/2K4P/6P1/8/8/8 b - - 0 36
Not a big change, isn't? We have only WTM -> BTM!
If there is no other winning way than reaching again this position (but with the other side to move) then this could be the deepest "Vital Zugzwang" discovered so far.
See the author's analysis here:
http://www.youtube.com/watch?v=2h_b0puS8Vk
[D]2n5/7k/5B1p/2K4P/6P1/8/8/8 w - - 0 1
White wins
Playing only DTM-optimal moves after 36(!) moves the position will look like this:
[D]2n5/7k/5B1p/2K4P/6P1/8/8/8 b - - 0 36
Not a big change, isn't? We have only WTM -> BTM!
If there is no other winning way than reaching again this position (but with the other side to move) then this could be the deepest "Vital Zugzwang" discovered so far.
See the author's analysis here:
http://www.youtube.com/watch?v=2h_b0puS8Vk