Frank Quisinsky wrote: ↑Sun Nov 08, 2020 4:05 pm
I saw the contributors ...
And I read from persons often ... we don't need the classical chess programs by chess programmers.
What we need is LcO.
LcO have not only strong contributors, LcO have a lot fan boys.
I'm not in Lc0 development in detail but there are probably some very(!) talented developers in the core team. Of course developing a framework for learning and using NN is different from traditional engine development but it deserves same respect imo.
And you can't blame the developers for their fan boys. There are a lot of SF fan boys out there too... but not a single Rubi fan boy I'm aware of
Latest Lc0 cpu package is easy to setup and comes with a network file so you don't have to search for the best in the endless lists.
Regards, Andreas
PS. Never thought I would become a supporter of Lc0 in this life
I know that (strong developers) but LcO alone goes in the wrong direction.
NNUE is a clearly more interesting way for all of the programmers!
LcO is more or less ...
A means to an end, one important way for further developments in computer chess only!
Very important with great ideas and strong developers.
... and what I forgot ... an own program!
"Traditional engines" I dislike a lot.
Please _without_ traditional.
Such combination of words fan boys like to read!
After magic bitboards nobody speak from "Traditional engines" (for an example only).
RubiChess is just great.
I like a lot that you working on own NN files and that you able to create fantasy for make Rubichess stronger.
Computer chess is more as a puzzle of good known sources.
Each own idea is important and not ... I am good in copy and paste.
RubiChess have an own style (nice to see in Excel stats from the still running FCP Tourney-2020).