This page is concerned with what many people will normally think of when the subject of a modified form of Chess is mentioned; variations in which a limited number of new pieces are added to the array, making for an enlarged square, or in some cases rectangular, board. Versions that start from the regular Chess array and versions that start from the Shatranj array are both discussed. More imaginative alternate forms of Chess, like the inventions of Christiaan Freeling, R. Wayne Schmittberger, or V. R. Parton, three-dimensional and hexagonal Chess, or even the Duke of Rutland's Chess, where not only is the board greatly enlarged, but each player has four Bishops and three Knights in addition to one Empress and two pieces with the move of both the Rook and the Fers, are avoided. By dealing with what might perhaps be considered the most humdrum of variant forms of Chess, I can both limit the size of this page, and observe some connections between the different forms in this range.
On a previous page, I discussed a number of forms of enlarged Chess from a number of cultures that were not modern inventions. On that page, I only discussed two enlarged forms of Western chess, Timur's Chess and Courier Chess. Courier Chess was particularly important in the transition from the old moves of the chess pieces to the new moves, and Timur's Chess is considered to be the most playable of the traditional enlarged chess variants.
I would have preferred to avoid Timur's Chess, because I needed to make new symbols for the Pawns for it, and because of a general dislike for enlarging the board with citadels, special squares into which the King can escape on the side, perhaps to force a draw. But on reviewing the various traditional forms of enlarged Chess in the references available to me, or even the more recent proposals to enliven Chess by the addition of a few new pieces, I could find no other variant with an array of pieces of comparable interest. On this page, I will quickly review many of the other choices I had considered.
While I have tried to divide these pages into sections, there is considerable overlap between them. Thus, in the page of a version of four-handed Chess of my own invention, I also describe several pre-existing ones; in the page about types of Hexagonal Chess, while I mainly describe existing modern inventions, I added a couple varieties of my own as well. Here, I will be discussing both traditional forms of enlarged Chess and modern inventions.
The diagram at the left shows the standard array of Chess pieces.
To facilitate the discussion of the older versions of enlarged Chess in what will follow, I will again note that the moves of the Bishop and the Queen were different in the older form of Chess, which can be referred to by the distinctive name of Shatranj. The Queen had been the Fers, which moved one space diagonally, and the Bishop had been the Alfil, which moved two spaces diagonally, jumping over any intervening piece without affecting it.
The standard name for a piece that moves two spaces orthogonally, jumping over an intervening piece without affecting it, is the Dabbaba, taken from Timur's Chess; the Camel which has an extended version of the Knight's move, also comes from there. The Giraffe, a piece with a doubly extended form of the Knight's move, comes from the Spanish Grande Acedrex, and the Zebra makes a move whose displacement is that of 2 spaces in one direction and 3 directions in a perpendicular direction, thus extended by one diagonal space from a Knight's move.
Three other pieces that are common in versions of enlarged Chess are the Princess, which has the moves of both the Bishop and Knight, the Empress, which has the moves of both the Rook and Knight, and the Amazon, which has the moves of the Queen and the Knight.
The names "Princess" and "Empress" which I have chosen to use here as the standard names for these types of piece are favored by composers of Fairy Chess problems, Chess problems where there are new pieces on the board or the rules are changed in other ways; they were not originally taken by the problemists from any Chess variant. On the other hand, it is possible that the name "Amazon" was used in a mediaeval Chess variant, although it could also be that it simply comes from a modern name for that variant.
Another piece found in several of the older forms of enlarged Chess is the one called the Griffin. It moves first one space diagonally, then changes course, and moves like the Rook. Slightly different forms of this piece exist in different games. In the Spanish Grande Acedrex, the piece is called the Aanca, and must move at least one space like the Rook, so it can move as Knight, Camel, Giraffe, or beyond. In Timur's Chess, this piece, called a Giraffe, must move at least two spaces orthogonally, and so it cannot move as the Knight. In another version of Great Chess, it is called the Great Fers, and must move at least three spaces orthogonally, and so cannot move as the Knight or as the Camel. To make this distinction, I propose that one refer to the Aanca of Grande Acedrex as a Griffin-1, the Giraffe of Timur's Chess as a Griffin-2, and the Great Fers of this other Great Chess as a Griffin-3.
Another common piece is the Man, which has the same move as the King, but which is only an ordinary piece.
Capablanca's Chess is a well-known modern variation on Chess. It was devised by the famous Cuban world Chess champion José Raoul Capablanca sometime around 1921. Its array is:
R Kt Pr B Q K B Em Kt R
but the Princess is called an Archbishop, and the Empress is called a Chancellor. The board was originally 10 squares wide, but 8 squares deep, and it was noted that it can also be played on a 10 x 10 board, in which case the pawns can also move three spaces on their first move. But I've seen recent accounts of Capablanca's Chess that only mention the 10 x 10 version, despite the fact that after Capablanca played some games of this variant with Edward Lasker, they agreed that the version on the 8 x 10 board was superior.
The name used for the Empress in that game came from Chancellor Chess, invented in 1887 by Benjamin R. Foster. In that game, only the Empress is added. The array for that game is:
R N B Q K Em N B R
Since only one piece is added, the Knight and Bishop on the Chancellor-side are reversed so that Bishops will be on opposite colors.
A variation with the same complement of pieces as Capablanca's Chess was devised by H. E. Bird for the 8 x 10 board, but with the arrangement:
R Kt B Em Q K Pr B Kt R
In his version of the game, the Princess is called a Guard, and the Empress is called an Equerry. This version was first described on page 111 of Volume I of the City of London Chess Magazine, in what I take to be the June 1874 issue.
The book Chess Eccentricities credits Carrera with designing, in 1617, and again on an 8 x 10 board rather than on a 10 x 10 board, a version with the layout:
R Pr Kt B Q K B Kt Em R
with the Princess called a Centaur, and the Empress called a Campione. This early extended version of Chess is also mentioned in B. R. Foster's 1889 book about Chancellor Chess.
Personally, I strongly prefer Carrera's arrangement from among those that have been tried adding the Princess and Empress (for which I use A. S. M. Dickins' nomenclature) since the central part of the array remains the same, with the Bishop and Knight both covering the Q2 and K2 squares. The fact that the Princess and Empress both have the Knight move as one of their moves, of course, also means that a solid wall of Pawns in front of the position after Castling remains possible, even with the new pieces in the outer part of the array.
However, Capablanca's Chess has the advantage of having the powerful pieces, the Queen, the Rooks, and the Princess and Empress, evenly spaced. Incidentally, Capablanca's first version of his Chess variant was almost the same as that of H. E. Bird, except that the Princess was beside the Queen, and the Empress was beside the King.
Yet another variant of Chess with this same complement of pieces is noted as having been called Universal Chess, with all the pieces being renamed with military-themed names, by Dr. Bruno Violet in 1928; two arrangements of the board were provided, one of which was identical to the 10x10 version of Capablanca's Chess. A photograph of a chess set for this game appears in A History of Chess by Jerzy Giźycki, but without the name of the inventor being given; this book is perhaps the first place where I heard of Gliński's hexagonal chess.
No listing of variants of Chess in which the additional pieces are the Princess (Knight+Bishop) and the Empress (Knight+Rook) would be complete, however, without taking note of what is perhaps the most recent major variant of this type: Seirawan Chess. Actually, that is the name the Chess world gave that variant by default when it was first introduced without a name, but now its creator has given it the official name of Sharper Chess.
Designed by noted grandmaster Yasser Seirawan, this variant is played on the normal 8 by 8 Chessboard, with the normal initial array of pieces.
I had mistakenly thought it was similar to one of the few variants of Chess that has often been played by real Chess players for entertainment, Pocket Knight Chess. In Pocket Knight Chess, each player has a third Knight which he can enter on the board at some point during the game instead of making a move. Surprisingly, at least to me, it is permitted for that Kight, as a result of its initial placement, to give check or even inflict checkmate.
This game is said to be particularly popular in the Netherlands, and it is also noted that a few tournaments of this variant have been played.
Unlike Chess, Go has not yet reached the point where computers are able to play the game better than almost everyone. One important reason for this is that an immense number of moves is possible at each step, since a stone can be played to any vacant location on the board.
This is an important advantage that is also shared to a lesser extent by Pocket Knight Chess, since until the pocket Knight is entered on the board, all the possible moves of entering it on the board also remain to be considered.
In Seirawan Chess, each player has both a Princess and an Empress off the board, the Princess being called a Hawk and the Empress being called an Elephant. But they aren't put on the board in the fashion that the Pocket Knight is, or a captured piece in Shogi.
Instead, when one of the pieces (other than a Pawn) makes its first move from its position in the initial array, as part of that move, the Hawk or the Elephant can be entered on the board.
It is not allowed to place both the Hawk and the Elephant on the board by means of Castling; at most one piece can be entered each move.
This is an ingenious way to approximate having the two new pieces as part of the initial array without having to enlarge the board. It does create enough additional variety in the initial layout to reduce the importance of memorizing opening variations as well.
This mechanism of entering additional pieces to a board that retains the conventional 8x8 size has been further refined in what looks to me to be a very interesting new variant of Chess, the game of Musketeer Chess.
This game was invented in 2011 by Dr. Zied Haddad.
The pieces are set up in the conventional initial array for normal Chess. Behind each player's pieces, there is a row of spaces which are not part of the normal playing board.
There are ten additional special pieces in each player's set of pieces, two of which will be used in any given game.
White selects one of the two pieces to use, and then Black either selects the other piece to use, or rejects White's choice, in which case the game is played with a default pair from among the additional pieces.
Then the players take turns, starting with White, placing one of their two extra pieces in a particular space behind their men.
The significance of this is that the extra piece enters the game when the square in front of it is vacated.
It is not allowed, however, to place one piece behind the King, and the other behind a Rook, so that there is no chance of Castling causing two pieces to enter the game at once.
Also related to the variants in which the added pieces are the Princess and the Empress is Cagliostro's Chess, with the array:
R Kt B Pr Em Q K Am Pr B Kt R
In this game, the Princess is again called an Archbishop, the Empress is again called a Chancellor, and the Amazon is called a General. This game, played on a 12 by 8 board, includes quite a few powerful pieces.
It is named after its inventor, who lived in Florida, and invented it in or before the 1950s, not after the famous historical European reputed to dabble in sorcery.
My own Random Variant Chess started from a baseline where the Camel and Giraffe were added to the array. This kind of expansion to Chess has also been anticipated; three versions played on a 10 by 8 board with Camels added to the array are known.
In 1955, Stanislaw Hofmokl-Ostrowski invented a game called Mephisto, where the Camels were called Devils. It had the array:
R Kt B C Q K C B Kt R
D. B. Prichard's Encyclopedia of Chess Variants credits Gizycki as the source of his information; I had seen it mentioned in his book A History of Chess, but without any information as to the initial array.
In 1965, there was Mexican Chess, invented by Prince Joli Kansil; here, the Camels were called Conquistadors, and the initial array was:
R Kt C B Q K B C Kt R
A third version of Chess played on the 10x8 board with Camels added used the name Dreadnought for the Camels. It dates from 1929, the inventor was Max Rieck, and a mention of this variant in the British Chess Magazine from 1929 notes that it was being played in Cape Town, South Africa. The initial array was not given.
A game played in Northern India was Atranj, with the array:
P P P P U U P P P P R Kt B Q K Am Q B Kt R
Pawns promote to the piece of their file. The Amazon is called a Prince. Instead of Pawns, the King and the Prince/Amazon each have an Urdabeqin in front of them, which in this game has the power to capture by a move one space in any diagonal direction, and to move one space backwards or forwards. As this is to avoid the issue of a Pawn promoting to the too-powerful Amazon or to King, when a Pawn changes file due to a capture, the piece it can promote to does not change, so labelled Pawns are required as in Timur's Chess.
Incidentally, the move of the Urdabeqin in this game was given to all the Pawns in the Prussian National Game, on an 11 by 11 board; this game included four Bishops and three Knights for each side, and its description would belong on another page.
Another notable chessplayer who suggested an enlarged version of Chess was Giuseppe Ciccolini. In 1820, he suggested a version with the array:
R Kt Ge Z K Q Z Ge Kt R
played on a 10 by 10 board. The Zebra was called an Elephant. The other new piece, the General, was a combined Bishop and Dabbaba-Rider. Thus, it could be thought of as the Queen of squares of one color. Unlike Capablanca, it would seem very unlikely that his invention was anticipated.
But this new piece was used again in at least one other variant, Emperor Chess, designed by H. R. Lambert in 1954 for the 12x12 board. The array of that game was:
R Kt B Ep Q Am K Q Ep B Kt R
the General being called the Emperor, and the Amazon being called the Commander. However, the Emperor was different from the General in one way; it could be blocked along its Rook lines by pieces on squares of either color.
If one removes the Emperors, one gets Power Chess on the 10 by 10 board, a commercial product from 1953, designed by one D. Trouillon, and in which the Amazon is again called a Commander.
In 1840, a game on a 10x10 board very similar to that for Capablanca's Chess was devised by L. Tressau, but the complement of pieces was different. One unusual attribute of that game was that Kingside of the board is on White's left, presumably so that the Queen can continue to take her own color while having a black square at the corner of the board to each player's left. This was The Emperor's Game, invented by L. Tressau. Its array was:
R Kt B Am K Q Pr B Kt R
where the Amazon was called a General, and the Princess was called an Adjutant. In this game, Pawns could move three squares on their first move, and the King could move either three or four squares when Castling.
He also invented, in the same year, The Sultan's Game. This one is played on an 11 by 11 board, and adds the Empress, which is called a Marshal. Since the number of squares is odd, all four corners are white, but the King remains on the left of the Queen, with the array:
R B Kt Em A K Q Pr B Kt R
Note that the Queenside Bishop and Knight are interchanged so that both Bishops remain on different colors, exactly as was done later in Chancellor Chess. The Amazon is again called the General and the Princess the Adjutant.
Given that the Rook, Bishop, and Knight represent the three basic types of movement present in the modern game of chess, it is not to be wondered at that other combinations of those moves, besides the combination of the Rook and the Bishop that is our Queen in Chess have been obvious pieces to add. A criticism of these variants would be that they add too many very powerful pieces to the board, making the game too quick, thus impoverishing rather than enriching it.
Perhaps the thing to do to produce the next form that Chess might take is instead to remove the Queen from the board, and replace it with the less powerful Princess, but add another whole row of Pawns to the forces on each side. This would ensure that even a very slight material superiority would be likely to translate into a surviving Pawn that could be promoted, probably to a Rook, so that draws would be less likely, but drawn-out and subtle play would still be required.
One Lieutenant Obert, in the 1870s, designed a variation of Chess where the new piece was the Guardian, which was a Fers-mover Wazir-capturer; an inverse Berolina Pawn, if one wishes, with the array:
R Gu Kt B Q K B Kt Gu R
Another game in which the new piece was not so overwhelmingly powerful as to unbalance the game, but was more powerful than the rather weak Guardian in the previous game, was Greater Chess, designed by one W. Day in 1942, with the array:
R Kt B Dk Q K Dk B Kt R
the new pieces being Dukes, which could move like a Queen, except only one or two spaces in each direction. This piece is known as the Mammoth or Mastodon in some recent Chess variants.
Of course, if adding a very powerful piece to the board shortens the game, at least it adds excitement, which brings some interest. The variations that add only pieces of limited strength, even if they are worthwhile, do not have that in their favor. As the Duke would seem to be stronger than a Knight, however, that game seems particularly well-balanced.
A more recent commercial variant called Dukes Chess also had a new piece called a Duke, but it moved one to three squares in each of the directions that a Queen could move. But it could only capture pieces which were two or three squares distant, not pieces on an adjacent square. When moving or capturing, it could jump over intervening pieces, but it still captured by displacement, not by jumping.
The game was played on a 9 x 8 board, as only one Duke was added to each side; the initial array had an unusual arrangement which was not symmetrical:
R Kt B Dk K Q Kt B R P P P P P P P P P - - - - - - - - - - - - - - - - - - P P P P P P P P P R Kt B Q K Dk Kt B R
The Bishop and the Knight are switched on the side of the board that is to the right of the White player, what would be the Kingside in conventional Chess, so that each side has two Bishops on opposite color squares, as was the case for Chancellor Chess.
But the Dukes and the Queens do not face each other, but instead the Duke is always to the right of the King from each player's viewpoint. Thus the array is neither centrally symmetric nor symmetric under vertical reflection.
D. B. Prichard's Encyclopedia of Chess Variants only gives the White initial array for this game; I had been able to see the complete array from photographs where copies of this game were being offered for sale online.
D. B. Pritchard's Encyclopedia of Chess Variants brought to my attention an extended form of Chess from 1696, called Enlarged and Improved Chess. It had the array:
- - P P - - P P P P En En P P P P R Kt B Gu K Q Gu B Kt R
The two added pieces were the Ensign, a Bishop-mover/Rook-capturer, and the Guard, a Rook-mover/Bishop-capturer.
Another variant I have only seen mentioned in print in that large collection which has an unusually modern and creative type of added piece is known as Spanish Chess, dating from 1739. It has the array:
R Kt B Ar Q K Ar B Kt R
where the new piece is the Archer, which moves along the Rook line pointing straight forwards, or the two Bishop lines pointing backwards. It is played on a board with only 8 ranks to its 10 files, like Capablanca's Chess and its relatives.
A game on a 12 by 12 board with the array:
R Kt Pr B Co K Q M B Pr Kt R
is credited to Alfonso X, although it is not from the famous manuscript that gave us Grande Acedrex. The Princess is called the Unicorn; the Counsellor has the moves of the King and the Knight combined, and the Man, called a Fool in this game, the move only of the King, but is only an ordinary piece. This game, which I learned of from the book Chess Eccentricities, does add the Princess, but accompanies it with less powerful new pieces, making for a more balanced game.
The game of Grande Acedrex, its invention credited to India by Alfonso X, is second to Timur's Chess in the variety of pieces it offered; it had the array:
R Li U Gi B K G1 B Gi U Li R
and was played on a 12 by 12 board, but with two empty rows of squares between the pieces on the back ranks and the Pawns. The Bishop was called a Crocodile, and the Griffin-1 was called an Aanca. The other new pieces were the Lion, which moved three spaces in either orthogonal direction, and the Giraffe, which has a twice-elongated Knight's move, as opposed to the once-elongated Knight's move of the Camel in Timur's Chess. Pawns promote to the piece behind them on the file, except that the King's Pawn also promotes to an Aanca.
The move of the Unicorn was originally given in some sources, deriving from H. J. R. Murray's A History of Chess as that of the Bishop, except that its first move had to be a non-capturing Knight's move. More recently, closer study of the original Spanish text of the manuscript which described this game has led to the conclusion that it was a more powerful piece, somewhat analogous to the Aanca, that moved as follows: the first element of its move was a Knight's move in any direction, and it could stop at that, or it could continue one or more steps as a Bishop, but forwards only.
A variety of Turkish Chess noted in Chess Eccentricities, with credit being given to a previous account by Dr. Van der Linde, was played on a 13 by 13 board also had two empty rows between the Pawns and the pieces on the back rank. The array was:
R Kt B Pr Ca Q K Gi Ca Pr B Kt R
with the Princess called a Rhinoceros, the Camel called a Stag (or Gazelle), and the Giraffe called a Great Queen.
Another account of this game, likely more accurate than my interpretation of what I saw in Chess Eccentricities, has the Great Queen having the move of a Griffin-3. And instead of Bishops, the board has, as expected for a game derived from Shatranj, Alfils - and, instead of a Queen, of course, there is a Fers.
This is an interesting assortment of pieces, but the fact that both Bishops and both Camels are on squares of the same color must be accounted as a defect.
The enlarged powers of the pieces in modern Chess have meant that it seems that adding only one or two new pieces would be enough to add as much zest to the game as one might want. Back when Chess was Shatranj, with the Fers and Alfil instead of the Queen and Bishop, adding the modern Bishop to the game was already a major innovation, and other much weaker pieces, such as the Dabbaba and the Man, were more proportionate to the existing array.
Two very old versions of Chess on a 10 by 10 board are noted in H. J. R. Murray's A History of Chess. One has the array:
R Kt Al M K Fe M Al Kt R
(the front row of Pawns being assumed, rather than shown, as it is the same for most of these games, a row of Pawns extending as far as the row of pieces behind them) and the other, at least according to John Gollon, has the array
R Kt Al D K Fe D Al Kt R
although if one took the description in H. J. R. Murray's work very literally, the array might be:
R Kt Al K Fe D D Al Kt R
In the first one noted, the Man was called a Dabbaba, and in the second one, the Dabbaba was called a Camel.
These two enlarged versions of Chess were mentioned in works by the Arab historian al-Masudi, and, therefore, they were in existence in the Tenth Century, and, therefore, they predate Courier Chess. Courier Chess, dating from 1202, is the oldest form of Chess known in which there was a piece having the move of the modern Bishop.
On a previous page, we've seen Timur's Chess:
P P P P P P P P P P P R Kt Ta G2 Fe K Wa G2 Ta Kt R B Ca Da Da Ca B
The Taliah was a Bishop that could move two or more spaces, but not just one space. As the Pawn promotion rules are complex, I will refer the reader to that earlier page for its full description.
R Kt Al B K Fe B Al Kt R
with the simple addition of the Bishop (here called a Dabbaba) to the old Chess array was used in Citadel Chess. This was played on a board with 10 files but 8 ranks (or, by some accounts, a 10x10 board), and with citadels added adjoining the four corners of the board; a King could draw the game by moving into a citadel on the opposite side of the board. As in Shatranj, the Pawn could only promote to a Fers, and there was no castling. Presumably, the Pawn also could not move two squares on its first move.
As Timur was born in 1336, and became known to history starting in 1370, this game, while it predates modern Chess (from 1475 or later), is later than Courier Chess. However, almost certainly the Taliah is an independent invention, as, needless to say, news of Courier Chess is extremely unlikely to have travelled from Ströbeck to the Mongol Empire.
Among the recent variants above, we saw a game called Atranj, which included a piece called the Urdabeqin. A much older variant, said to be from Turkey, also fit that description. It was played on a 10 by 10 board.
Here, the layout was like this:
R Kt B Pr K Am Em B Kt R P P P P U U P P P P -- -- Kt Kt -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- Kt Kt -- -- P P P P U U P P P P R Kt B Em Am K Pr B Kt R
Note that the symmetry differs from that of Chess and most variants; corresponding pieces are opposed on a line through the center of the board instead of a vertical line.
In this version, the Urdabeqin moves one square in whichever direction the opponent's King lies. I found that description of its move to be too vague for me. A more extensive discussion of this question is now provided on this page.
Another game with a very similar layout has pieces with slightly different powers; again, the symmetry is central instead of vertical, and the layout is:
R Kt B Q K Am Pr B Kt R P P P P Em Em P P P P -- -- P P -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- P P -- -- P P P P Em Em P P P P R Kt B Pr Am K Q B Kt R
This game is described in an Indian manuscript from the 18th century, and is said to be a Turkish form of Great Chess. I have now added a diagram of this game, and a description of it, to the same page as the discussion of Atranj.
Both of these variants are among the most interesting historical Great Chess versions. They may represent the development of the same game over time, although my first thought is that they were both descriptions of the same game, with one or both descriptions containing errors, since the information on some older chess variants is sometimes incomplete.
This game is played with the conventional board and initial array, but it requires a second set of pieces.
Described in the book Chess Eccentricities, like the Japanese chess game of Shogi, it allows captured pieces to be re-entered on the board. However, the manner in which this is allowed is considerably different.
When an enemy piece or Pawn is captured, as part of the same move, a corresponding piece from the alternate chess set of the player's own color may be entered on the board, on one of the ordinary starting squares for that type of piece, if a vacant square of that type is available. If no such square is vacant, then the captured piece can only be removed from the board. Either such square can be used; and thus, when capturing a Bishop that moves on the light squares, the Bishop entered corresponding to it can be entered on the starting square of the Bishop that moves on the dark squares instead if desired.
In the case of the Queen being captured, it may be entered on the King's starting square as well.
The description of the game claims that it has a tendency to continue for longer than ordinary Chess does, and placing a limitation on the re-entry of Pawns is recommended as a countermeasure. (The specific limit noted was not to allow a Pawn to be re-entered by a player who already has more than five Pawns in play.)
In the case of Shogi, re-entry of pieces, as it adds to the advantage of the winning player, accelerates that player's victory; one would expect the same here, even though re-entered pieces are entered at their starting points, instead of being placed aggressively as in that game.
However, what would tend to prolong the game is that even exchanges of pieces - or Pawns - would not simplify the game, so one way in which captures could shorten it is lost. And with careful play, even exchanges (and sound sacrifices, which would now require a greater positional gain to be sound) are more common than the simple gain of a piece - which still takes place, but later in the game, once one player has built up an advantage.
So, if the players are evenly matched, and play well and carefully, the one thing that doesn't happen is that a series of exchanges leads to a draw by insufficient material.
Given this, I should like to propose a variant of this game that addresses the specific issue.
When a piece is captured, instead of the corresponding piece being entered immediately, it is entered at the start of the capturing player's next move, before a piece is moved, if a vacant square is available.
With the additional exception that if the opponent, on the turn immediately following the capture, also captured a piece of equal value (Knights being considered to be equal to Bishops; otherwise, only the same kind of piece is of equal value) then neither piece may be re-entered.
In this way, material advantages accumulate more quickly, but exchanges can still simplify the game.