Lyudmil Tsvetkov wrote:
On your mentioned, and as you said, recommended by noomen.ctg, 7.h4, black simply plays 7...h6, staying with advantage.
On 8.h5 g5, and white could draw at most with completely closed position.
Where is white advantage above? What move is best for white?
On 7.f4, as in the winning white game you posted, black plays, of course, not 7...0-0, which is weak, but 7...Nd7 (the place of the knigth is on e5, and not of the bishop):
Where is the white advantage? What move is best for white?
Any games in your database with this particular position?
8.f5 pushes things too far, on 8.Nf3, black could castle and try to push f7-f5 at some point, but black also always holds a secure draw by simply playing 8...f6, then h6 and g5, and position gets fully closed.
So, please tell me, where is the white advantage?
well i suggest you make a database yourself with GM moves,
to examine such positions; the average rating difference in those
six games i referred to was negligible btw.
The main reason -besides time (i might have a look later) that i don't want to look up -or analyze- the positions you mentioned above, is that they
are rather theoretical (therefore there also were not much games; in the variations you mentioned there will even be less games, and then by lower rated players i suspect) After 1.d4 c5?! 2.d5 d6 (or g6) White is not
always playing c4, but often Nc3 or 3.e4. Playing Nc3 saves a tempo
while getting hold of the e5 square, and a later 3..e5?! as in the
old Benoni would be weird (and bad).
PS we can discuss endlessly about how to evaluate positions; there
are various methods and not allways the GM's would agree; anyway I
do agree that engines as Stockfish and Komodo not always find the right moves. But how to find the really best moves is another topic; if you
know it so well i suggest you take up correspondence chess and see
how far you come. You could also find maybe an engine programmer
who is willing to try out the parameters you suggested in your
(interesting) paper which i found here:
https://chessprogramming.wikispaces.com ... endium.pdf
2012 edition, maybe you would have some updates now;
but then you still would have to see if it would work in practice
for such an engine, it's often a matter of delicate tuning i guess
(not even talking about speed optimizations); or an engine
programmer may have a look at making -some of these- parameters
an UCI option for human tuning/experimenting; volunteers?
Maybe someone in Bulgaria ?