Nah...the move is sub-par and a bias of SF(controlling the d5 square).
Im still waiting for it to recognize that Bf4/Qd2/O-O-O is best. It does....after Bf4 is played,as black like depth 30. According to chesstempo, Bf4 is 51% human win rate for white, 19% for black. It does the same thing(missing Be3/Qd2/O-O-O early on) in the Petroff's Nc3 attack, but there is a way out of that one.
On the plus side, the engine has a new move for what to do if black doesnt take the pawn!
Nah...the move is sub-par and a bias of SF(controlling the d5 square).
Im still waiting for it to recognize that Bf4/Qd2/O-O-O is best. It does....after Bf4 is played,as black like depth 30. According to chesstempo, Bf4 is 51% human win rate for white, 19% for black. It does the same thing(missing Be3/Qd2/O-O-O early on) in the Petroff's Nc3 attack, but there is a way out of that one.
On the plus side, the engine has a new move for what to do if black doesnt take the pawn!
I am amazed that your plan is exactly identical to Leela plan 6. Be2 7. Qd2 8. O-O-O ( in low depth SF still prefer Bc4.)
This game was from CCCC Rapid Rumble , SF vs Crafty. From move 6 to 16, look at how stockfish beautifully set up king bishop fianchetto .(popular human set up)
Position at move 6.
Position after move 16.
At move 6, most engines or engines with bad hardware will choose 6. Be2 or Bc4 as it immediately give bishop mobility bonus. However SF Dev Version did not.
Stockfish progress since stockfish 9 and beyond is truly remarkable, especially in deep understanding of positional changes.