I no longer use hyper-bullet games for testing, as I reported in Arasan's blog (https://arasanchess.org/blog.shtml). I have regularly had changes pass very fast time control testing (usually 0:08+0.1) and fail at longer time controls (typically blitz, 2:0+1). YMMV, but for me hyper-bullet is not reliable.
--Jon
About stockfish's optimizing for bullet
Moderators: hgm, Rebel, chrisw
-
jdart
- Posts: 4366
- Joined: Fri Mar 10, 2006 5:23 am
- Location: http://www.arasanchess.org