Not all Chess960 positions are equally complex(arxiv.org)
57 pointsby MaysonL6 days ago8 comments
  • __s2 days ago
    > indicates a slight tendency for White to face harder opening decisions

    supporting the quip "the hardest game is to win is a won game"

    Not surprised at end re classical position not being the most even configuration. In that configuration bishops & knights practically start aimed at controlling center, so there's little awkward properties to dampen White's initiative. One of the rooks even get to castle out of the corner

    Chess960 would be better if they just got rid of castling in it, tho wouldn't be surprised if that makes for certain positions getting even worse for Black

    See also https://news.ycombinator.com/item?id=26066844 for thought on game theory of strategy when playing perfect is computationally infeasible

    • NickC252 days ago
      Chess960 would also be better if both sides were asymmetrical and there were novel positions for both players in every game.

      I go to a chess event 2-3 times a month in the city where I live, and there are a few of us that are big into variants and play a lot of Bughouse, Crazyhouse, Racing Kings, etc. 960 is a bunch of fun but asymmetrical 960 is a blast, and asymmetrical Bughouse 960 / Crazyhouse 960 is the most fun and hard version of chess I've ever played. There is no theory, just pure tactics and reaction.

      • nilslindemann2 days ago
        > Chess960 would also be better if both sides were asymmetrical and there were novel positions for both players in every game.

        Yes! I never understood why people are so much into Fischer Random when there is also e.g. Benko's Pre Chess, where the players just place their pieces on the first and eight rank at the start of the game. Every player can decide to break the symmetry or not. They can even set up the normal chess position if they desire to do so. But for some reason today only Fischer Random is played, probably because Fischer was more famous than Benko. But Benko's version is more elegant, the players have full control and there are more start positions.

        https://www.chess.com/forum/view/chess960-chess-variants/pal...

        • nilslindemann2 days ago
          Also interesting: Start with an empty board and let the players place their pieces and pawns however they want in their half of the board, as long as the piece does not attack an enemy piece.
      • indoordin0saur2 days ago
        Probably many asymmetrical combinations are unfair to black. Maybe running through combinations and simulated games with a chess engine could identify ones that are fair, asymmetric and fun? Then a database could be built up of these combinations and it could be randomly selected to start your game.
        • mr_wiglaf2 days ago
          From my understanding that is exactly what this group are doing: https://chess960v2.com/en
        • billforsternz2 days ago
          Maybe there are asymmetrical combinations that actually give Black the advantage? Because Black's setup is nicely harmonious and White's is clumsy. Or maybe not I'm entirely unsure.
        • __s2 days ago
          Similar idea is randomized openings. Checkers does this already. TCEC does chess AI tournaments using sharp preselected openings (matchups playing 2 games, one of each color)
    • copper4eva2 days ago
      I can't say for 960 specifically, but for standard chess getting rid of castling usually results in the players just manually castling their kings. I believe that is why the move was introduced in the first place. So it really doesn't accomplish much except make the opening a bit more limited, since they have to leave themselves a way to manually run the king over one of the rooks. Usually to the short side, since that's quicker. Basically makes queen side much less viable to leave the king at. And queen side castling was already the rarer of the two options. I imagine it would be a similar story for a lot of 960 positions. I'm not sure how getting rid of castling would benefit anything. In 960 you already get a lot of super crazy aggressive positions with exposed kings even with castling.
      • kibwen2 days ago
        > I can't say for 960 specifically, but for standard chess getting rid of castling usually results in the players just manually castling their kings.

        The entire design of 960 is backwards when it comes to castling, because it was deliberately designed to facilitate castling. This is the whole reason there are "only" 960 positions, as opposed to 2880 positions if our only restriction is that bishops are on the opposite color (and that both sides are symmetric). By reifying castling as something that must exist rather than a gross and unfortunate hack to paper over the flaws of the standard chess position, the ruleset puts the cart before horse.

  • ummonk2 days ago
    They’ve defined memorization complexity as having to memorize the best out of almost equally good moves (as opposed to being able to play the best move without memorization because it is so obvious.

    In reality it’s almost the other way around. Because white usually has several good moves at every point, they can just memorize one of them, while black needs to memorize how they’ll respond to every good move white could make.

  • firasd2 days ago
    I've been analyzing classic "romantic" games using Stockfish with multipv (showing the top 4-5 lines rather than just the best move)

    1. Morphy vs. Duke of Brunswick (The Opera Game)

    https://lichess.org/study/xAo78qLb/truC6WoM

    16. Qb8+.

    This is viewed as Morphy doing a stylish Queen sacrifice

    But if you look at the MultiPV:

    Qb8+* leads to forced mate.

    Qc3 or Qb7 drops the advantage significantly.

    Qb5 actually allows equality

    If he had played anything else, he would have been imprecise. It wasn't a gamble

    2. D. Byrne vs. Fischer (Game of the Century)

    https://lichess.org/study/UZlSqSLA/Ku9M59je

    Fischer plays 17... Be6, leaving his Queen hanging.

    Standard narrative: "Fischer offers his Queen for a mating attack!"

    Engine reality: 17... Be6 is the correct move. Trying to save the Queen actually loses the advantage.

    Byrne taking the Queen (18. Bxb6) was a massive blunder. The engine actually wants Byrne to ignore the Queen and trade off Fischer's Knight on c3. He ends up with a Queen stranded on a3, a total spectator

    • reassess_blind2 days ago
      Qb8+ is a fairly obvious mate in 2. I don’t think anyone views it as a gamble.
      • firasd2 days ago
        Right. So I guess that's my quibble with the term sacrifice (shared by Rudolf Spielmann)

        But what's interesting to me is the counterfactual like outside of these 3 queen moves he would have lost the entire advantage. So it was like a tactical shot like capturing the golden snitch in Harry Potter

        • reassess_blind2 days ago
          Sure, I get what you're saying. It's still a sacrifice, but the compensation is just mate in 2, so the there's no real "sacrifice" here.

          That being said, any sacrifice that doesn't guarantee a better (or at least equal) position isn't a sacrifice either, it's just "hope chess", aka a bad move. In Blitz or Bullet you can make the case for a "bad" sacrifice for positional complexity and putting time pressure on your opponent to make accurate defensive moves.

          In the Opera game, Black just played a poor game start to finish. Giving up the bishop for the knight, pushing the B pawn while the king wasn't castled.

        • le-mark2 days ago
          I had two “brilliant” moves in one chess.com game today. One was a bishop sacrifice that would have led to mate in three. The other was a queenside castle that the engine wanted me to do sooner. I suck at chess, although I did see the bishop sacrifice as the right move. The engine rated me at 1500 for the game.
    • Trufa2 days ago
      Most of the times they mean the amazement of just even considering that move a couple of moves ahead and not discarding that branch.

      But yes, a true gambit could be considered something that's objectively bad, but humanly makes sense.

    • TZubiri2 days ago
      This is a common theme, gambits are such depending on what your level and calculation depth is.

      The queen's gambit opening (almost inarguably a gambit as it is part of a well accepted name of a second move), really isn't a gambit in the sense that you can always recover the pawn, however it is a gambit in the sense that you temporarily give it up.

      If we were particularly short sighted, no doubt, responding to an early white bishop threat on g5 or b5 with a knight on f6 or c6 would look like a gambit, as we are sacrificing the knight, but lo and behold, we regain the minor piece afterwards with xf6 or xc6!

      The distinction would be whether the gambit or sacrifice is solid or refutable. But it is in both cases a sacrifice.

  • econ21 hours ago
    I have a special way of playing vs computers where every trade and every constraint is considered bad. The point is to keep enough paths open that it becomes a game of intuition and gambling. I never win of course but did have some games where it went in full retreat or spend 3-4 moves on putting pieces in their previous position or gradually changed its evaluation from a sizable advantage to disadvantage after thinking for a while.

    I do the same with players who are to good to beat at a normal game.

  • NickC252 days ago
    Ironically the starting position that I've found to be the most balanced for black and white is just swapping the starting square of the bishops and knights.

    It turns normally-sound opening moves like 1.d4 and 1.e4 into liabilities and emphasizes the knights as blocking material to occupy squares like d2 and e2 , but the tradeoff is that a early-developed bishop can get a lot more active centrally via an open wing.

    Such a layout makes for a very cautious opening phase where neither side really feels comfortable giving up much material. Really a fascinating setup.

  • thatfunkymunki2 days ago
    interesting that the most balanced one is extremely similar to the default, which is not so:

    > \#198 (\texttt{QNBRKBNR})is the most balanced, with both evaluation and asymmetry near zero... Remarkably, the classical starting position-despite centuries of cultural selection-lies far from the most balanced configuration.

    • copper4eva2 days ago
      That formation is pretty close to the standard position though. Just swaps a Queen and Rook. It puts the Queen in the corner, a less aggressive position with less options to develop. I've only played a little 960, but these queen in the corner positions seem to often lead into more closed positions.
      • NickC252 days ago
        Yeah, agree, but in the setup you mentioned, 1.b3 and 1. b4 are both strong moves, because it basically forces the game to develop kingside from the get-go.

        Seems the opening can get really sharp, or basically a race to bunker via 1.Nf3

  • FergusArgyll2 days ago
    I can't get myself to enjoy chess960. So much of my understanding of the game comes from pattern recognition and familiarity with concepts (e.g. This structure is basically a reverse carlsbad, black should play for the minority attack, white has to get a kingside attack rolling and the knight to e5)

    960 is disorienting

  • TZubiri2 days ago
    "Remarkably, the classical starting position-despite centuries of cultural selection-lies far from the most balanced configuration."

    Right, balance would be one of many parameters that make a game popular. Complexity is even a positive aspect of a game, a simple game with a simple optimal strategy would not be the most popular.

    Simplicity of rules is another parameter, the classical configuration is somewhat symmetrical on the Queen King axis, many candidates would provide a rather asymmetrical and hard to remember initial configuration.