Your network blocks the Lichess assets!

lichess.org
Donate
Giraffe

https://pixabay.com/photos/cat-animal-sleep-cats-pet-animals-800760/

When to break opening principles and steer away from popular practice

ChessAnalysisTacticsStrategyOpening
An interesting Neural Network Opening discovery in a line of the English Opening

Hi all

It seems humans are pretty principled sometimes and the principle of

"Generally capture pawns towards the center"

can sometimes have interesting exceptions where capturing away from the center is actually far more solid.

I was studying a key world championship game recently, and there is this critical moment:

Petrosian played Nxc6 and to my surprise, Botvinnik replied with Nxc6 which to me intuitively gives White a fantastic grip over the d5 square - an English Opening player's dream.

https://lichess.org/study/JIEm0Eqe/2KMB52qX#16

I looked at what is played in practice, and usually, it is bxc6 - this seems to have a principled human bias of capturing towards the center. It seems though that the Latest Stockfish Neural Network-powered engine likes dxc6.

dxc6 seems to solve the d5 square issue, and also doesn't seem as though Stockfish is impressed by the "counterplay" of bxc6 with plans of Rb8 and c5 later. Perhaps dxc6 is the way to go but it does invite an exchange of queens. However if one is playing black and wants a draw, this seems like an interesting opening innovation.

https://lichess.org/study/JIEm0Eqe/agfE354y#16

Anyway, Botvinnik's choice led Black to an awkward position where White's d5 outpost seemed like positional torture.

https://lichess.org/study/JIEm0Eqe/2KMB52qX#29

Petrosian neutralises e-file pressure from black later with Bf3 and looks forward to a powerful c5 break:

https://lichess.org/study/JIEm0Eqe/2KMB52qX#43

Final position

https://lichess.org/study/JIEm0Eqe/2KMB52qX#121

Takeaway points

  • The newer neural network engines do seem to have interesting candidate contributions to Opening theory
  • Practice has not caught up with modern dynamic neural network engines in many cases
  • Neural networks may cause opening books to have to be revised - including the more positional openings such as the English opening
  • It may be scary to step into unchartered opening theory, but the top neural network engines may be rewriting or revising Opening variations or at least setting up interesting experience for future chess practice

Hope you enjoyed this blog :). Any likes and follows are really appreciated. Also, I also have some interesting chess courses at https://kingscrusher.tv/chesscourses to check out.

Cheers, K