[Computer-go] cgos <--> kgs rating
3-Hirn-Verlag at gmx.de
Mon Jun 27 12:35:07 PDT 2016
2 cent from an outsider:
"Detlef Schmicker" <ds2 at physik.de>
> ... I am still fighting with resigning in the case of
> value-network and playouts disagree, ...
Might such positions be interesting starting points when
looking for solutions for the discrepancies?
In general, positions where programs of about equal strength
have very different evaluations should be interesting for
attempts to improve the positions.
PS. In 1999, Chess grandmaster Arthur Yussupov participated
an experiment where he good computer hints from two or more
different chess bots. Most he liked positions where the
programs had different opinions.
More information about the Computer-go