You can’t tell everything just by looking at a game position on a traditional physical chessboard. If you see:
You can’t necessarily tell:
- the board orientation
- whose turn it is to move
- what castling rights exist
- which pawn, if any, can be captured en passant
There are also lots of other things you can’t tell, such as whether certain types of draws can be claimed, or even whether the game being played is actually normal chess, or Chess960, or some other game. But I’m only going to worry about the items listed above. In common situations, that’s the extra information you need, in order to know what moves are legal.
With an actual physical chessboard, there are practical reasons that players don’t, for example, place special tokens on the board to help remember whether they have castling rights. It would be an annoyance, and it’s expected to be easy enough for the players, and anyone following the game, to keep this information in their heads.
But with computer chessboard displays, there is no such drawback. And we do, sometimes, get to see some of this traditionally hidden information. For one thing, many chessboard diagrams have coordinates, so we know which side of the board is which.
(If that’s too obtrusive for you, it suffices to mark just the a1 square in some way.)
Sometimes the player to move is indicated by some sort of mark near the board, whose color usually indicates the player to move next. For example:
The mark might be a circle, or a square, or a triangle. Unfortunately, there seems to be no standard for this, and some such marks are ambiguous unless you know what system is being used.
Also, a lot of chessboard images include an indication of the previous move, either by highlighting the “from” and “to” squares, or by drawing an arrow.
This tells us the player to move next, and also suffices to tell whether en passant is possible. This type of annotation is common during actual games, or in screenshots from games. But it’s not really possible to use it with things like chess problems, or diagram generators, where there may not be a meaningful previous move. So it’s not a replacement for a dedicated “whose turn is it” mark.
Clearly, none of these annotations were intended to provide complete information about the board position, since none of them cover castling rights. Castling in chess is not obscure; this cannot have been accidentally overlooked. It’s a deliberate decision not to show this information.
I guess it would just be sacrilegious to do something like this:
But I’d like to see it happen anyway. It may well be that some chess software has an optional feature to display such annotations. But if it were a common thing, I think I would have seen some evidence of it by now.
Of course, you’d want graphics that are less ugly, and more subtle, than mine. And you don’t need both yes-you-can-castle badges and no-you-can’t-castle badges. For standard chess, you’d probably use only the no-you-can’t-castle badge. The lack of a badge (in a position where castling rights could exist) would mean that castling rights do exist. The main reason I offer a yes-you-can-castle badge is that Chess960 has some crazy castling rules, and I wonder if using the yes badge would be preferable in that game.
Some details. A badge should only be shown if the position of the pieces makes it relevant. For example, a white rook would only potentially have a badge if it is on a1 or h1, and the white king is on its home square of e1. In the lion’s share of chess games, you might never see a badge.
The castling badge (or the lack of a no-castling badge in a suitable board position) is an indication of abstract castling rights. It’s not about whether the rook can castle on the next move, or even whether it will ever be possible for it to castle. It could be that all the legal next moves will cause castling rights to be forfeited. But for now, it still has castling rights, and the badge should indicate that.
The en passant badge marks the pawn that could be captured, not the pawn that could do the capturing. This is simpler, because there could be two pawns that can do the capturing.
It can happen that an en passant capture is not legal, because it would leave that player’s king in check. If there are no legal en passant captures, as in the following position, then I say the pawn should not get the EP badge.
It is unfortunate that this rule is somewhat complex to implement, and that it is inconsistent with how the castling badge works. But I think it is the right rule.
Now that we have this annotation system, I would like to use it to clarify the rules of chess. A way to decide whether two positions are “the same” is needed by some rules regarding when a draw can (or must) be declared. According to me, the criteria should be:
Two positions are “the same” if and only if they have the same kinds and colors of pieces on the same squares, and they have the same next-player-to-move, castling-rights annotations, and en passant annotations.
Whether this is equivalent to the actual rules of chess, I simply don’t know.