lizzie
lizzie copied to clipboard
featurecat
Issues Analyzing Handicap Games
I'm trying to analyze the below handicap game using KataGo on Lizzie and two strange things are happening.
White seems to have a very high starting percentage even though komi is 0.5 and there are 3 handicaps. The engine doesn't seem to recognize that the handicap stones are there and keeps trying to play on the spots where the stones are.
I've asked in KataGo Github and they said that maybe it's an issue with the GUI sending the incorrect commands to the engine. https://github.com/lightvector/KataGo/issues/396#issuecomment-767637363
Any idea whether this is really an issue with Lizzie and how the issue can be fixed? Let me know if any more info is needed.
It was one of the bugs in Lizzie-v0.7.4 and was reported in #765 shortly after Lizzie-v0.7.4 was released.This bug also happens in my environment. It was soon fixed by #766 and merged into Lizzie's master. So I think the bug will probably be fixed in the next release, but the latest release hasn't been announced yet ... Compiling Lizzie yourself with #766 will solve this problem, but if that's difficult, I recommend "Lizzie-improved".
I borrow a lot of pull requests from Lizzie and maintain my own repository so that I can quickly deliver a more user-friendly Lizzie. https://github.com/hope366/Lizzie-improvements/releases/tag/v2.2 Of course, the bug you reported about the handicap stone has also been fixed 😄 (Note that komi is set to 6.5 by default)
The Lizzie-improved seems to be working fine! Thanks a lot!
Hope they can release an updated version of Lizzie soon~