Pauli Chess | Hacker News Reader

Pauli Chess

blog.plover.com

75 pointsweinzierl33 comments

"The main idea is: two pieces can be on the same square... Pierre-Françoys says he wishes that more than two pieces could share a square... The name “Pauli Chess”, is inspired by the Pauli exclusion principle, which says that no more than two electrons can occupy the same atomic orbital."

Pierre-Françoys' variant would be Bose chess, I suppose?

Also, shouldn’t true Pauli chess allow an arbitrary number of pieces in the same square, but no two of the same kind and color?

Back in 1980 or so, my family invented a game that has an interesting mechanic. Sort of like chess, but also like battleship, with two boards, and each player can only see his own board. When a player moves, they announce the piece and which square the piece is departing from, but not which square the piece is landing on. You track your opponents pieces one move behind. Collision resolution happens the next time you move the piece, and find out you did or didn't make it. You're allowed to guess where you're opponent might be moving and attempt to intercept. There's quite a bit of luck in the game since you're largely playing blind, but it's the kind of luck that feels more natural than injecting randomness with dice.

Rules, 3D-printable parts, etc. are here: http://smcameron.github.io/dueling-masters-of-space-time/

Like Kriegspiel but with half the information and you're in space. https://en.m.wikipedia.org/wiki/Kriegspiel_(chess)

An improvement to your pieces design is to copy the idea behind the internationalized shogi where generals icon is describing its movement https://images.app.goo.gl/preLM7N1sb3taqor7 or the duke board game if feeling adventurous

Anyone have links to play online?

Some sites I can think of that support other variants:

https://lichess.org/ ( https://github.com/ornicar/lila/ )

And some crazier variants:

https://www.pychess.org/ ( https://github.com/gbtami/pychess-variants )

https://chessvariants.training/

I bet a superhero from lichess would be happy to bust this out in a weekend

Lichess's policy is generally not to implement tons of variants, even those that would play nicely with PGN (this wouldn't).

You could definitely hack this together on a fork of lichess, but don't expect to see it on the main site.

https://pippinbarr.github.io/chesses/

(doesn't have Pauli Chess, but does have some fun variants)

I threw together a quick-and-dirty implementation of this game. You can play it at http://paulichess.jsdavid.com/

The interface is _extremely_ janky and the rules are not fully implemented (I just started earlier today), but it should give you a basic idea of the play style.

I'm pretty sure the game is gonna turn out to be extremely unbalanced, but it'll be interesting to see how people who are much better at chess than I am can do at it.

Other chess variants beyond the common: 1) friendly fire is on. You can capture your own pieces if you want.

2) You can move yourself into check, but only if you checkmate your opponent simultaneously, in which case the game is a tie

One of my favorite variants is suicide chess. The goal is to be the first to lose all your pieces. All rules are the same with one addition: if you can take a piece, you must.

It's really fun! You can play it on Lichess under "Antichess."

> 5. Pieces of opposite colors sharing a square do not threaten one another.

This is going to be strange with kings. You could capture a piece in a square with a king and the king would require two moves to capture it, and if it was a queen the king couldn't move. Then just have to check the square for mate.

I don't think that could happen, if an opposing king shares a square with another piece, you would choose to take the king to win the game.

You don’t capture kings in chess, and I don’t think that changes in this variant (at least the article doesn’t suggest that). If you attack the square with the king on it, the king is in check. If there’s no legal move that puts the king out of check, that’s checkmate.

Perhaps interestingly, you can create a discovered check with two pieces on the same square, say two rooks. Like a double check in regular chess, you can’t remove the check by capturing, because you would only capture one of the rooks.

But then you'd be on the same square as the remaining rook, and pieces of opposite colours coexist: the remaining rook wouldn't threaten the king.

> Perhaps interestingly, you can create a discovered check with two pieces on the same square, say two rooks. Like a double check in regular chess, you can’t remove the check by capturing, because you would only capture one of the rooks.

That wouldn't work. The king would have been in check because of the first rook already, so either the king had to move away, the first rook would have been captured on the previous move, something would have been placed between the king and the rook, or it would have been checkmate.

Unless of course the two rooks would have been in place but you create the check by removing an obstacle, that would work.

There are not all that many ways this could happen, though. I think you can only enter the square of the opposing king with a pawn. With all other pieces, you’d have to threaten the square first, which would be check. You could promote a pawn into the king and create a queen on its square.

The other way would be if the king itself moves. If the king moves onto a square with an opposing rook and a queen, and decides to capture the rook, the king is trapped until the queen moves. In this scenario, the queen must have moved there in the last turn, putting the king in check. Maybe that’s a new way to get a stalemate.

Alphazero mastered chess by already-implemented self-play.

A slight change in the game rules can have a significant effect on the game and strategy, as anyone that delves into chess variants knows. Is the same true for their implementations?

What is the limitation for state of the art AI in adapting in such a way to rule changes?

Turns out DeepMind has already been working on exploring variants with Vladimir Kramnik!

https://www.chess.com/article/view/no-castling-chess-kramnik...

Currently they have to train from scratch, but adapting to new tasks is an active area of research for RL broadly.

No-castling chess didn't need to be trained from scratch. Starting with a trained or partly-trained AlphaZero would certainly have worked, since no-castling chess is a strict subset of chess. The only reason not to is if you are DeepMind and training resources are really cheap.

For more esoteric variants like this, it makes sense to start from scratch, though eventually you could imagine an AI that is good at adapting to all perfect-information two-player games, or even others.

It's a very simple change, since the fundamentals of the game is still the same. It should take no more than a day to change the code and re-train the model.

Just switch the knights with the bishops, and you're in a different universe.

I dreamt about playing chess last night because of this... very very strange.

Either two pieces can be on same square this would definitely mess with tactics. But strategically speaking is just like normal chess but with more squares

Not at all. As the page says, you can’t block pawns, which entirely changes the game. You also don’t need to move pawns to develop pieces.

Super leaky chess? Looking forward to an implementation so I can try and wrap my mind around how it works. What's to prevent both players from pushing all pawns through the leaky sieve and promoting everything to Queens right off?

Captures

You can still block pawns using a pawn plus another piece.

Can you? The rules are unclear. Interpreted literally they would mean that the pawn could move into the square (directly in front of it) and then take one of the opponents pieces in that square.

>All pieces move and capture the same as in standard chess

so pawns have to capture diagonally still... I guess if you want to see 4 as an exception to that then they can capture straight ahead, but I'm assuming it isn't.

"strategically speaking", it is a completely different game:

* you can only take pieces from fields occupied by two of the opponent's pieces.

* you cannot just move through the board because - you can only move field by field through adjacent fields occupied by one piece (of any! color) each

> you can only take pieces from fields occupied by two of the opponent's pieces.

Not true, you can choose whether to occupy the empty place or to capture

Reply on news.ycombinator.com