12

As we all know, in algebraic notation, we usually suppress the notation for pawns. That is, unless they are capturing something. Then, for some reason, we must notate the file of that pawn. For all other moves, you only need to notate the file (or rank, or both) if multiple pieces of the same kind could move to the field in question, so that this extra information is necessary to avoid ambiguities.

For instance, the Scandinavian defense usually goes something like 1. e4 d5 2. exd5. Why not just 2. xd5?

Gaussler
  • 279
  • 1
  • 8
  • 1
    These are already shortcuts for `1. e2-e4 d7-d5 2. e4xd5` It is just a matter of convention or taste how much *redundance* you are willing to live with – Hagen von Eitzen Dec 14 '22 at 13:25
  • 2
    There is already a well-established abbreviation for that move: `ed`. And it's shorter than `xd5`! – TonyK Dec 14 '22 at 13:40
  • Technically, you could describe that particular move as `2. x` and it's still unambiguous... but most shorthands have limitations. In this case if there were any capture in the position for White besides e4xd5, this would not work. `2. ed` works so long as no other White pawn on the e-file may capture into the d-file. '2. xd5' works only so long as no White c-pawn can take on d5. '2. exd5`, however, is always sufficient to describe e4xd5. – user45266 Dec 16 '22 at 02:24
  • Some of the current chess notation seems to be truly redundant, though - chess could be transcribed just fine without the x for captures, the + symbol for checks, 0-0 and 0-0-0 for castling, any of the various en passant notations, or even the # for checkmates (agreed draws, timeouts and resignations still need to be recorded though). – user45266 Dec 16 '22 at 02:24

6 Answers6

21

You seem to want the notation to be as efficient as possible, disregarding ease of use. Let's take this to extremes:

  • If White only has a single move or capture that ends on c4, then instead of Bc4 or Bxc4 should we just write c4?
  • If White only has a single legal Rook move, then instead of Re5 should we just write R?
  • If White only has a single legal move, then...what? Just notate it as "" ?

The conventions in use are a compromise between efficiency and ease of use. And they work very well.

TonyK
  • 817
  • 4
  • 15
  • It seems to me more that OP is looking for consistency than efficiency. The normal situation is that you indicate the type of piece moved (by omission, if it's a pawn), whether or not there is a capture, and the destination square, with information about the starting square if necessary to disambiguate. Except for pawn captures. – Especially Lime Dec 14 '22 at 08:10
  • 1
    The difference between your examples and the situation of pawn capture is that `xd5` contains more or less the same information as `exd5`, except that you may have to look at the board to see where the pawn was. But even without looking at the board, you kinda know what's going on. In your other examples, you don't. We notate captures, checks, mates, en passants, etc. because this is all helpful for getting an overview of the game using only the notation. But with `xe5`, you wouldn't really lose much of this information. – Gaussler Dec 14 '22 at 08:23
  • 1
    @Gaussler: Situations where two pawns would both be capable of performing a capture are pretty common. Indeed, if Black moves a pawn from d7 to d5 when White has pawns on c5, c6, e5, and e6, there could be *four* white pawns poised to capture the black pawn on the next move (two of them capturing directly on d5, and two capturing in passing on d6), – supercat Dec 15 '22 at 21:13
  • 2
    Some friends and I actually at some point had fun in trying to write down games in this kind of notation. Mathematically you could define it as something like "a shortest substring of the move's algebraic notation that isn't a substring of any other legal move's algebraic notation". Not useful but was certainly a fun exercise. The best part of it was being able to often write the game-ending move as just a #. – JiK Dec 15 '22 at 22:45
  • Although now that I think of it, my definition must be off because 1.c3 would not be possible to write down with that rule. Perhaps we had another definition. – JiK Dec 15 '22 at 22:47
  • @JiK Fool’s mate looks good in this notation: 1. 2. 3. 4. # – Peter LeFanu Lumsdaine Dec 16 '22 at 12:08
18

There is no real reason, it's not technically necessary.

Equally, you don't generally have to write the 'x'. Knowing where the piece is going is enough to know that it's a capture.

People must have thought this notation hit the sweet spot between maximum terseness and readability.

RemcoGerlich
  • 27,663
  • 78
  • 120
  • 15
    You’re right. Technically, you don’t have to play chess either. It’s not the most cost-beneficial way to satisfy your need for nutrition and survive. But we do it anyway. – Gaussler Dec 13 '22 at 11:52
  • 1
    It seems common to just say "e-d" or "e takes". Just not to write it. – Edward Dec 14 '22 at 00:54
  • 2
    It might also help people check to see if they are still synced correctly with the game they are trying to read from the record. I know nothing about chess notation, however. – Mark Rogers Dec 14 '22 at 15:14
8

Imagine a situation where there is also a white pawn on c4, for example 1. e4 d5 2. c4 e6 3.xd5. You would have no way of knowing which pawn took on d5, was it the one on e4 or the one on c4.

dzi
  • 243
  • 1
  • 18
    True, but that same issue occurs when any other pair of pieces of the same kind can capture the same piece. In that case, and only then, we notate the file, or rank, or both. But with pawns, we do it no matter what. Why? – Gaussler Dec 13 '22 at 11:05
  • 3
    @Gaussler Your comment here should be part of the question to clarify things for other people! :D – justhalf Dec 14 '22 at 10:14
8

I'd suggest a different reason: to make it sound natural.

How would you pronounce Nxe5? I would say "knight takes e5". In that case, the putative pawn move "takes d5" would not sound natural, because it has a verb with no subject.

Once you decide to add a "subject", you may as well do so in a way that gives extra information, since this may be useful. So rather than "pawn takes d5", it makes sense to say e.g. "e takes d5".

There is no grammatical reason to add this information to other captures (which already have a subject) or normal moves (which have no verb).

Especially Lime
  • 998
  • 7
  • 7
3

Why not just 2. xd5?

Because we always indicate the piece being moved. The notation 2. xd5 doesn't tell us which piece is doing the capturing. We have to work that out for ourselves. We don't need to write 1. e2-e4, although we can, because 1. e4 tells us the e pawn is the one being moved.

For instance, in Petrov's Defence we don't write 1. e4 e5 2. Nf3 Nf6 3. xe5. Instead we indicate the piece being moved by 3. Nxe5.

Ola Ström
  • 269
  • 1
  • 3
  • 8
Brian Towers
  • 92,895
  • 10
  • 227
  • 372
  • 10
    True, but the usual convention is that “nothing” means “pawn”. If I move the knight to `d5`, I write `Nd5`. But if I move a pawn there, I write simply `d5`. Applying the same logic to capturing, `Nxe4` would mean that the knight captures whatever piece is at `e4`, whereas `xe4` would mean that a pawn does the job. – Gaussler Dec 13 '22 at 11:17
  • 2
    @Gaussler: There is never more than one pawn that can move to a particular location without capturing. If one knew nothing about what was on the board, the notation "e4" could mean either "move a pawn directly from e2 to e4", "move the first pawn in rank e from e3 to e4", or "move the second pawn in rank e (which had previously been in rank d or f) to e4", but the board position would make it obvious. – supercat Dec 13 '22 at 20:24
  • 1
    This answer does make sense. The notation always contains a letter which describes _something_ about the piece being moved—either the piece type or the file, sometimes both. For a simple pawn move like "e4," the "e" effectively specifies both the starting file and the ending file, since pawns stay on the same file when they're not capturing. – Tanner Swett Dec 14 '22 at 11:39
  • What happens is that there are many pawns but only a pair of the other pieces. So, other pieces can be named by their type, which is usually enough to identify them, but the pawns are better named by their file. "d5" means "pawn d moves to row 5". "exd5" says "pawn e captures d5". – Florian F Dec 14 '22 at 22:43
2

Just realized one other good reason: When using this notation, it’s easier to get an overview of the pawn skeleton of the field solely from looking at the algebraic notation. For instance, to find out where the e-pawn is, you can look for the move starting with an e. That would be e4 in this case. Next, you find the next one, which is exd5. Then you make a mental note that the e-pawn has moved to the f-file. Now you can keep following it as it advances along this file, until it maybe changes file again. And so on.

It doesn’t offer much help in figuring out which pawns have been captured, but at least it helps.

Gaussler
  • 279
  • 1
  • 8