4 ms·

Using Stockfish to identify ideal squares

83 points by akkartik 2y ago

notarealllama 2y ago
Nice, long time lichess player including blitz arenas but had no idea they had articles. Some interesting analysis here and how to approach the problem domain (pawn structure as a concern) but that first example, there is just no way to get that knight there in any reasonable way. Talking about unrealistic moves.
- 29jm 2y ago
  The first example is from an actual game, Larsen-Korchnoi, where White does manage to put the Knight on d5. The article includes that part of the game, it's a pretty funky manoeuvre.
- sevg 2y ago
  That knight manoeuvre is actually what is played by the GMs in that (very real) game. Which means it's as realistic as it can get. Even at GM level (perhaps especially at GM level) it's not uncommon to see 3-4 moves spent rerouting a knight to a better square.
- adleyjulian 2y ago
  It's a closed position. You have time to reroute the knight and it's one of the few clear plans besides doubling the rooks on the only half open file. The knight on that outpost was a monster and it ended up winning the game for him.
- Scarblac 2y ago
  It's only four moves, Nf3-d2-b1-c3-d5. (I don't know if that was the actual route taken in the game)
- kthejoker2 2y ago
  The board in the article is interactive you can actually play out Larsen's moves with the arrows underneath
imrehg 2y ago
Very interesting, especially when compared to shogi (Japanese chess), where captured pieces can be dropped in anywhere on the board. So for shogi players this "ideal square" calculation can be even more natural and more flexible as well: besides the "getting existing pieces from A to B", the "drop on B" is a lot simpler. No wonder that piece exchanges (so there is something in the hand to drop) are basic feature of the gameplay. (Source: being a fan of shogi but very very very early in my learning journey, so experts would likely describe this differently.)
- ycombinete 2y ago
  That extra dimension makes Shogi such a brain burner. It also forces something of a permanent middlegame.
- BoardsOfCanada 2y ago
  There is a chess variant like this - Crazy house.
  - nordsieck 2y ago
    > There is a chess variant like this - Crazy house. Interesting, I've always heard it referred to as "Bughouse" or "Siamese chess". https://en.wikipedia.org/wiki/Bughouse_chess https://en.wikipedia.org/wiki/Bughouse_chess
    - mtlmtlmtlmtl 2y ago
      Bughouse is the 4 player version played on 2 boards. A lot easier to play with conventional physical boards.
    - Scarblac 2y ago
      Bughouse is the 4 player, 2 boards variant. You capture your opponent's pieces and give them to your team mate who can then drop them on their board. That way there is no problem with the piece being the wrong color to drop. Shogi solves that by having the pieces be flat with the two colors on different sides, afaik. Crazyhouse is online only, so it has no problem switching the color of the captured piece.
rhdunn 2y ago
The unrealistic squares issue could be resolved by using legal moves for the piece. You could then evaluate the best for 1 move, 2 moves, up to e.g. 4 moves. You could also eliminate moves that visit a square already reached by a previous square to avoid duplicates. It may be interesting to apply weights to reduce the score of squares many moves away. This would need to handle advantageous positions like check or checkmate.
- FeepingCreature 2y ago
  Evaluate the best, following valid moves, while ignoring your own pieces? Modelling your ability to create opportunities.
greymalik 2y ago
> This exercise involves looking at a piece and imagining which square it would be best placed on, without initially worrying about how to get it there. Without using Stockfish, how does one do this in their head? What rubric can you use to evaluate whether one square is better than another? (I’m a very low level player, obviously,)
- arh68 2y ago
  Knights like to sit fairly forward, reinforced by our pawns, on squares their pawns can't attack. Bishops like long diagonals, and to complement the pawns' attacking squares (Ps & Bs tend to cover only one color). Also good for "countering" a knight 3 squares away (Nf6 & Bf3, for example, though it seems to prefer Bg2). Rooks like open files, doubling up, 7th rank plays, and (in this case) the rare opportunity to line it up w/ their king (Rg3). Queens like not being trapped or sidelined (by either side's pieces), good diagonals, good files, or both (like some natural mix of rook & bishop wants).

83 points by akkartik 2y ago

Nice, long time lichess player including blitz arenas but had no idea they had articles. Some interesting analysis here and how to approach the problem domain (pawn structure as a concern) but that first example, there is just no way to get that knight there in any reasonable way. Talking about unrealistic moves.

29jm 2y ago
The first example is from an actual game, Larsen-Korchnoi, where White does manage to put the Knight on d5. The article includes that part of the game, it's a pretty funky manoeuvre.
sevg 2y ago
That knight manoeuvre is actually what is played by the GMs in that (very real) game. Which means it's as realistic as it can get. Even at GM level (perhaps especially at GM level) it's not uncommon to see 3-4 moves spent rerouting a knight to a better square.
adleyjulian 2y ago
It's a closed position. You have time to reroute the knight and it's one of the few clear plans besides doubling the rooks on the only half open file. The knight on that outpost was a monster and it ended up winning the game for him.
Scarblac 2y ago
It's only four moves, Nf3-d2-b1-c3-d5. (I don't know if that was the actual route taken in the game)
kthejoker2 2y ago
The board in the article is interactive you can actually play out Larsen's moves with the arrows underneath

imrehg 2y ago

Very interesting, especially when compared to shogi (Japanese chess), where captured pieces can be dropped in anywhere on the board. So for shogi players this "ideal square" calculation can be even more natural and more flexible as well: besides the "getting existing pieces from A to B", the "drop on B" is a lot simpler. No wonder that piece exchanges (so there is something in the hand to drop) are basic feature of the gameplay. (Source: being a fan of shogi but very very very early in my learning journey, so experts would likely describe this differently.)

ycombinete 2y ago
That extra dimension makes Shogi such a brain burner. It also forces something of a permanent middlegame.
BoardsOfCanada 2y ago
There is a chess variant like this - Crazy house.
- nordsieck 2y ago
  > There is a chess variant like this - Crazy house. Interesting, I've always heard it referred to as "Bughouse" or "Siamese chess". https://en.wikipedia.org/wiki/Bughouse_chess https://en.wikipedia.org/wiki/Bughouse_chess
  - mtlmtlmtlmtl 2y ago
    Bughouse is the 4 player version played on 2 boards. A lot easier to play with conventional physical boards.
  - Scarblac 2y ago
    Bughouse is the 4 player, 2 boards variant. You capture your opponent's pieces and give them to your team mate who can then drop them on their board. That way there is no problem with the piece being the wrong color to drop. Shogi solves that by having the pieces be flat with the two colors on different sides, afaik. Crazyhouse is online only, so it has no problem switching the color of the captured piece.

rhdunn 2y ago

The unrealistic squares issue could be resolved by using legal moves for the piece. You could then evaluate the best for 1 move, 2 moves, up to e.g. 4 moves. You could also eliminate moves that visit a square already reached by a previous square to avoid duplicates. It may be interesting to apply weights to reduce the score of squares many moves away. This would need to handle advantageous positions like check or checkmate.

FeepingCreature 2y ago
Evaluate the best, following valid moves, while ignoring your own pieces? Modelling your ability to create opportunities.

greymalik 2y ago

> This exercise involves looking at a piece and imagining which square it would be best placed on, without initially worrying about how to get it there. Without using Stockfish, how does one do this in their head? What rubric can you use to evaluate whether one square is better than another? (I’m a very low level player, obviously,)

arh68 2y ago
Knights like to sit fairly forward, reinforced by our pawns, on squares their pawns can't attack. Bishops like long diagonals, and to complement the pawns' attacking squares (Ps & Bs tend to cover only one color). Also good for "countering" a knight 3 squares away (Nf6 & Bf3, for example, though it seems to prefer Bg2). Rooks like open files, doubling up, 7th rank plays, and (in this case) the rare opportunity to line it up w/ their king (Rg3). Queens like not being trapped or sidelined (by either side's pieces), good diagonals, good files, or both (like some natural mix of rook & bishop wants).