Opening Theory according to Computers

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

Peter Berger
Posts: 789
Joined: Thu Mar 09, 2006 2:56 pm

Re: Opening Theory according to Computers

Post by Peter Berger »

I remember that lc0 was the first engine to like this line for white on my computer, as I looked up the ancient Steinitz-Zukertort game where it was first played in high-level chess after seeing that it liked it even after a long thought. I don't know about current lc0 versions.

There is some relevance chess-wise. If black can reach this position "by force" after 4.O-O against the mighty engines, then 4.O-O is unlikely to be the best move for white in the first place. There is 4. d3 after all that should at least keep the little advantage of having the first move for a little longer.

People are right that my post was not scientific and uses rather vague terms. I still thought that looking at this line is interesting as it is so obviously drawn, even for players who are not that strong, and Stockfish usually is extremely good at coming up with opening lines that keep an ever so tiny edge for white.
jefk
Posts: 1085
Joined: Sun Jul 25, 2010 10:07 pm
Location: the Netherlands
Full name: Jef Kaan

Re: Opening Theory according to Computers

Post by jefk »

Peter Berger
Posts: 789
Joined: Thu Mar 09, 2006 2:56 pm

Re: Opening Theory according to Computers

Post by Peter Berger »

jefk wrote: Thu Dec 30, 2021 1:16 pm 15.Qe2! is better
:D
jef
https://www.lulu.com/en/en/shop/jef-kaa ... 8r8qz.html
I think you have a point here. Of course black will still equalize eventually after 15. ... c6 followed by Nd6, but it is not quite as trivial as I had thought before - thanks for pointing this out. Is this line covered in your book?
jefk
Posts: 1085
Joined: Sun Jul 25, 2010 10:07 pm
Location: the Netherlands
Full name: Jef Kaan

Re: Opening Theory according to Computers

Post by jefk »

Peter B,
the line with 5.Re1! is mentioned in a foot note in the book, (while
having the 'old' line with 5.d4 leading to queens exchange as main),
depending on (opponent ) level people can make a choice, obviously.

But anyway, the book contains a [i]practical[/i] repertoire (for
human players) up to intermediate level; and the repertoire still is
sound, even with the new(er) Nnue evals.
(for human play, i don't think anyway engine evals have
much value (nnue or not), also not at higher human levels, there
usually are more factors to consider (as Towforce pointed out)
than only the result of alfa/beta minimax (eg sharpness,
the possibility for the opponent to make mistakes, etc)

For such reasons, when playing as human, i prefer 1.Nf3 or 1.d4
btw, although some tactical players still may prefer 1.e4.

PS my book contains quite an extensive repertoire for 1.d4
for White at intermediate level (and i don't even know it
by head yet myself, although obviously i often come
through the opening phase without difficulties, probably of
having absorbed some knowledge (without active learning
during the years of analyzing/writing).