These are chat archives for ujh/iomrascalai

23rd
Apr 2015
Urban Hafner
@ujh
Apr 23 2015 06:12
Yes, it's amazing how good this is when it's correctly implemented. ;) I also ran my usual tests and it wins 65% now (up from 45% for my first buggy implementation).
Urban Hafner
@ujh
Apr 23 2015 11:32
And expanding the leaves of the tree later doesn’t make any difference. It still wins ~ 65% against GnuGo.
Urban Hafner
@ujh
Apr 23 2015 12:07
@iopq I think you need to update the strenum crate to work with the latest compiler.
Urban Hafner
@ujh
Apr 23 2015 12:47
I had a quick look, but I’m not sure how to get it to compile again.
Also, which code is Imrscl-022-e exactly? It does perform quite well, I think.
iopq
@iopq
Apr 23 2015 17:41
--use-empty-area-prior true
Urban Hafner
@ujh
Apr 23 2015 17:50
I'll have to give that a try too once everything compiles again. As well as turning on your ladder reading in the playouts.
iopq
@iopq
Apr 23 2015 17:50
having it default to true I think it's being used
Urban Hafner
@ujh
Apr 23 2015 17:51
Nope, see my pull request. The code in main() turned it off by default.
iopq
@iopq
Apr 23 2015 17:53
oh, I see, it was switched
iopq
@iopq
Apr 23 2015 21:31
I need to wait for the num crate to get fixed