Shatranj was played on the same board, and with basically the same initial array of pieces, as conventional Chess:
The Bishop was instead the Alfil, or the Ship, and it moved two spaces diagonally, ignoring intervening pieces. The Queen was instead the Fers, or Counsellor, and it moved one space diagonally.
The Alfil in particular, because of the limited number of squares it could reach (one-eighth of the board), but also the Fers, seem to be so weak that it's hard to see that they were of much use.
This apparent flaw in the earliest version of two-player Chess (of course, Chaturanga preceded it) seems, at least, to put paid to any theories that Chess was too wonderful to be a human invention, and thus was given to us by friendly space aliens.
Could one change Shatranj to correct this in a less radical way than was done when the modern Bishop and Knight were introduced in the later Middle Ages?
It seemed obvious to me that the Alfil could be modified while retaining most of its original move in this way:
When it moves without capturing, it moves two spaces diagonally, but can be blocked by intervening pieces.
To capture, it moves one space in any direction, like the King - and it gives check to a King one space in any direction from it as well.
This seemed reasonable; since the piece is supposed to represent a ship, that it can stop at only a limited number of destinations is reasonable, but it should still be able to project power where it goes. So, while it can still travel, until it captures, to one-eighth of the board, now it can capture pieces on three-quarters of the board at any given time between captures.
(It is the Sanskrit word Roka, the source of the name of the Rook, that means ship, while the Sanskrit word Hasti, which named the Bishop in Shatranj, that means elephant. Thus, calling the Alfil a Ship is something of an anachronism; the names of the two pieces were switched when Chaturanga gave way to Shatranj. The shape of the Rook, as a turret from a castle, came from the howdah on an elephant, and so the pieces retained a tie to their old identity in Chaturanga even as their names were changed in Shatranj.)
How to improve the Fers wasn't as obvious. My first few ideas seemed to either make it too powerful or not powerful enough.
Here is what I finally settled on:
The Fers is primarily intended to protect the King, so it should continue to have a short-range move. I decided that adding an orthogonal move for capturing only wasn't worth doing.
But given that Pawns in Shatranj only promote to a Fers, something had to be done to make the game behave like normal Chess in respect of the importance of Pawn promotion.
So, what I decided was this:
• (1) The Fers moves and captures one space diagonally, but it gives check as if it were a Rook - any distance along an orthogonal line.
Thus, as far as contributing to checkmate goes, promoting a Pawn to a Fers is almost as deadly as promoting it to a Queen or at least a Rook.
This turns Shatranj into a game like ordinary Chess while still retaining much of the original moves of the Alfil and Fers.
As in normal Shatranj, there is no double-step initial move for the Pawns, and hence no en passant capture, and no Castling. Also, the King does not get a Knight's move to escape once from mate (including stalemate), as there seems no reason to make checkmate a best-two-out-of-three thing; people know how to play Chess properly these days (the thanks or the blame for that going to Steinitz and company).
Incidentally, to distinguish the Alfil and Fers of this game with their modified moves, I propose to give the pieces in this game the following distinctive names: the Alfil-mover King-capturer would be called a Frigate, to signify its status as a warship, and the Fers with the power of checking along Rook lines (as modified in the third alternate rule noted below, however, as I think this is likely to be the preferred rule) will be called the Minister (that this name is used in Chinese chess should cause little confusion).
Treating checks to the King in a different manner than the ability to capture, however, is unusual, and therefore I preferred to avoid this by means of a nearly equivalent rule:
• (2) Give the Fers the nearly equivalent power of making a capture, once and only once, as if it were a Rook. As long as it had not used that power, it could give check along Rook lines.
Keeping track of whether each individual Fers had used that power, however, would require an expedient such as using checkers, which could be flipped from the King side to the plain side, for each Fers. I did not like that as a characteristic of something which was to represent, albeit with tongue in cheek, the "perfect original form" of Chess, that the space people gave us. If the Fers was a piece that could be flipped over, that would be hard to forget.
But I realized that a slight modification to this rule would do the job.
• (3) Give each player the privilege of making only one capture along Rook lines with a Fers - any Fers, whether on the board, or yet to be obtained by Pawn promotion - only once during the game.
As long as that privilege has not been used, then every Fers present would still give check along Rook lines, since one only has to be able to capture the King once, with any Fers.
This third version of the rule adds an extra stratagem to the game: one could give check to the King in such a way that the only way to escape from the check would be to capture the checking piece by means of a Rook move by a Fers. Since this would now severely reduce the potential effectiveness of that player's forces, it would usually be well worth sacrificing a piece to do so.
Of course, this might make it too easy to escape checkmate, and thus this third version of the rule for the Minister may be objected to on this basis. Given that stalemate and bare king were historical victories in Shatranj, perhaps the way to address this would be to use the point schedule proposed by Emmanuel Lasker for regular Chess:
Checkmate 10 points - 0 points Stalemate 8 points - 2 points Exposed King 6 points - 4 points Draw 5 points - 5 points
While, in my opinion, for normal Chess, anything more than a 6 point/4 point split as credit for stalemate is too much, for Shatranj, this would be an appropriate schedule, and thus it might also be appropriate for Reformed Shatranj with the third variant of the Minister rule, if experience shows it is too easy to, effectively, turn that game into regular Shatranj by forcing the opponent to use up his Fers capture privilege.
However, on further reflection, forcing the opponent to use up the privilege of an orthogonal capture by a Fers in this fashion is sufficiently close to checkmating the opponent that it is inaccurate to characterize doing so as being "too easy", and thus checkmate can still, even in this third alternative, be treated as having the same level of difficulty as in ordinary Chess.