Why Pawns are the Worst 5D Chess Pieces

Pawns fuck everything up. They are my most favorite and most hated piece on the board. On one hand, they’re complicated, which is pretty interesting. On the other hand, they’re fucking complicated, which is infuriating. Let me explain why. Point one: Pawns are the only piece that make a distinction between “forward“ and “backward.“ When extending to higher dimensional space, the addition of extra directions begs the question, what does forward mean? Well, it depends on the layout of the board. Take this for example

We see that the x, z, and v axes do not bring us any closer to the opponents side, yet the y and w directions do. So should be allow a pawn to be able to move two “forward” directions (y-w) at once to capture? Obviously not! It breaks the balance of the game. Should we consider that restriction for all pieces? This is a fierce debate between me, and literally everyone I’ve every explained this to. I think we should. Point two: Pawns are a combination of 3 different pieces, each with unique conditions to their moves that almost no other piece has.

When actually coding an engine, you need to add all sorts of conditions that take this into account. The way I’ve done it is make the pawn a composite piece, consisting of a non-capture forward motion, a capturing forward-diagonal motion, or a starting double forward non-capturing motion. If we take this approach though, both queens, shogi pieces, and a check detector moveset can be defined as composites of other simpler pieces, with shogi pieces also utilizing “forward“ directions.

So complicated, but useful. Fine pawns, you win this one. Its interesting and actually not too hard, once you figure out all the fucked up conditions... for the most part...

Then you realize they can En Passant AND they’re the only piece to promote. Great, the promotion condition is made easy by defining extra-dimensional forwards, but the actual moves are worse. Not only is En passant only valid for one turn, but it also requires another piece be in a different location than the target motion, and captures a piece that it doesnt directly land on. No other piece capture works like this. Castling works fairly similar, and we can draw parallels to the two. You can view En Passant as first moving the pawn horizontally to capture, then moving it forward in a second command. Promoting is moving the pawn forward, and then moving a friendly queen on top of it. This also neatly applies to shogi drops. Castling is just moving the king and rook accordingly on the same turn, but like en passant requires another piece to be in a specific space other than where the king lands, and has non-checking conditions. Not that I want to think about Castling on a 5D Chess board, god forbid anyone fucking suggests it again. But what we learn from this is you need to break moves up into something smaller, that I call commands. A simple piece from ,piece to, coordinate from, coordinate to. By making moves out of these, they are also completely reversible. So great, moves are made of commands, we’re done here right? Not yet because this problem is made worst by the parallel universes.

See once we start introducing an actual multiverse, we start to need to perform multiple moves within a single turn, then submit our turn when all those moves are complete. Its not too bad if you’ve understood the move break down, but its still slightly annoying. Then comes the network. God fucking help me with the networking. Currently I’m writing a serializer that converts a turn into a series of integers and the nested structure is not conductive to that. This wouldnt be a fucking problem if it weren’t for the god damn pawns though. Note some of these are actually reverse pawns as well, so they can go the opposite direction around the torus.

So, it has been a very, very long time since the birth of this project. It has to be something like 3 years ago I made my first 4D Chess board in VRChat, and even before then I had made a shogi board before scripting was even implemented. Months later, I made the visualizer for the 4D, and we joked about taking things one step forward. 5 Dimensional Chess! How crazy would that be? Surely thats too much! Down the line I eventually made 7D Chess with multiverse time travel, to which I failed multiple times to make desktop ports of it. Attempts to rebuild from the ground up failed a few times as well, until we eventually got here. nDimensional Time Travel Chess. And furthermore, after all this time, I have finally, officially hit a first that I genuinely never intended to do. I have denied this request from several friends, over and over again, but I would like to announce that it has happened. "Blake... Regular Fucking Chess Board"