Chess Variants
Game or Chess Variant?
What features must a game have in order to qualify as a Chess Variant?
From the discussion in the comments it seems the following might be the answer.
The single defining quality of "Chess" is that
the winning condition is predicated on one (the royal) of two (royal and non-royal) classes of pieces
If a game exhibits this quality it is a Chess Variant, if it doesn't it isn't.
page revision: 7, last edited: 13 Sep 2007 14:19
Recent comments at CV.org have led me to look again at the introduction of chance into a CV. (for example ChessWar.)
I had thought such an introduction would render the resulting game too far from FIDE chess (assuming that's the standard, but why not take XiangQi as the standard?) to be called a CV. But now I'm not so sure. Perhapss there are grades of CV? (CVs of the third-kind, anyone?)
A chess game must have at least one royal piece, such that the unavoidable loss of all a player's royal pieces causes that player to lose the game. The opposing player who caused that unavoidable loss of the [last] royal piece is generally, but not necessarily, the winner. Is a one piece army enough? I would argue no, and rule out The Maharajah and the Sepoys as a chess variant. The concept of "royal" brings with it the concept of "non-royal". A player may lose all the non-royal pieces without necessarily losing the game. In the Maharajah and the Sepoys, it is irrelevant if the singleton Amazon is royal or not. A true chess variant "must" have royal and non-royal pieces on each and every side.
So Extinction Chess wouldn't qualify as a true chess variant since there are no "non-royal" pieces: if a player loses all pieces of any particluar type, he loses. Is that right?
Actually, there are at least 10 non-royal pieces per side: 7 pawns, 1 knight, 1 bishop, and 1 rook, the way I see it, making Extinction a [very good] chess variant. The concept of royal piece becomes tricky in some variants, and Extinction is one of the best examples. Because, if you look at it, you can consider the royal piece to change from game to game, depending entirely on what singleton piece is "mated" in a particular game. It can change within a game, if a player "checks" a singleton rook, then later "mates" a singleton bishop instead because the first player got out of "check" by promoting a pawn to a rook, thereby removing the original rook's royal status.
The way I see it, extinction chess will qualify as a chess variant as chess is just a game of boards, pieces and about anything would qualify as a chess variant
I strongly disagree with Joe Joyce's statement, "A chess game must have at least one royal piece, such that the unavoidable loss of all a player's royal pieces causes that player to lose the game." I simply ask the question, "Why must this rule exist?" Joe Joyce asks, "Is a one piece army enough?" And goes on to state that he would argue "no, and rule out The Maharajah and the Sepoys as a chess variant[s]." But, I have played Maharajah and in that variant, the one piece for black is enough… enough for a fun game. Pieces are, in essence, how they move. The idea of Royal or Non-Royal simply implies their status… but a player with no-concept of Royal and non-royal could play a variant, as long as he [or she] knew the rules… understood the laws of physics, so to speak, of that chess world. So, to the statement, "A true chess variant "must" have royal and non-royal pieces on each and every side." I say, it simply is not true. And Maharajah is an excellent example.
I agree with you Gary. The defining quality as set out on this page says nothing about the distribution of the two classes, only that a CV must have both and that the winning condition is dependent in some way on just one of those classes. On that basis Maharajah is certainly a CV.
Once again, I am stuck arguing a minority position. Fortunately, this is nothing new for me. Unfortunately, at the CVPages, I already conceded a similar argument. But, like Galileo, I will mutter under my breath: "It still moves". Because I am at risk of not offending enough people here, I will move my argument to the topic: the "Chess" family of games", as there's another post there I wish to take wild and inaccurate swings at. Here, all I will do is admit this is only my opinion, and those of you who wish to be wrong may feel free to dispute it. Enjoy.
What is the nature of the game of chess, and what makes a game a chess variant? Here is my opinion:
"Chess", in the widest sense (ie. in the sense of a family of games), has certain properties:
1. It is played on a "board" with "pieces". Each piece occupies a position on the board, and each position on the board can be either occupied by a single piece, or can be empty.
2. It is a competitive game (ie. a game where a player attempts to win and tries to avoid losing) played between two or more players.
3. The players alternate making moves in a "round-robin" sequence. A move is the changing of position or type of one or more pieces on (or off) the board.
4. When a piece is moved on the board, that movement may result in a "capture", which is the removal from play (ie. from the board) of one or more pieces.
5. There are at least two different piece colors. The color of a piece determines the rules for which player may move that piece.
6. There are at least two different piece types. The type of a piece determines the rules for how that piece may change position on the board.
7. A piece may also be brought back into play (ie. moved from off the board, back onto the board) or may change type while on the board.
8. It is a game that involves no "hidden" information (except for the thoughts of the opponent(s)). All events that occur on the game board are determined by rules and information that are known prior to the start of the game, and known by all players.
9. It is a game that involves two classes of pieces, "royal" and "non-royal". The winning condition of the game is to capture or checkmate one or more of the royal pieces.
It is interesting to note, that the game of checkers, has the same properties, except for the last. I believe it is this last property that differentiates chess from other non-chess games (such as checkers).
If you change the winning condition of, let's say FIDE chess, such that the winner is the player that captures all of the opponent's pieces, then, in my opinion, that game becomes much more like checkers than chess. Similarly, if you changed the winning condition to be the player who is first able to move their King to the last rank, I think the game will play more like parchesi, or one of the other "race" games.
That is not to say that one cannot change the winning condition and claim a new chess variant. Certainly in the case of Extinction Chess, that has been done. And certainly any of the properties above could be changed to produce a new chess variant. It just seems to me that the last property is the most crucial to making chess unique from other board games.
David, that is a truly excellent post. You have certainly given an extremely comprehensive framework on which to hang this discussion, if you haven't ended it by being so thorough. And in saying this is your opinion, you have demonstrated just what "expert opinion" means, not to mention expert writing. Most impressive. Thank you.
It seems, much to my delight and surprise, that David has come up with a single simple defining quality, which I have expressed in the wiki page itself.
"Similarly, if you changed the winning condition to be the player who is first able to move their King to the last rank, I think the game will play more like parchesi, or one of the other "race" games."
- David, I love your list and your synthetic, taxonomic compilations. I think my "Ï'm a Ferz, Get Me Into There" has a lot more in common with Chess than Parcheesi. (Btw, several members of my family have gotten angry at me for my extensive use of blocks in Parcheesi; it's a tactic they don't appreciate).
A king race is not unfamiliar to devotees of FIDE chess. Often endgames are decided based on who can get their king to shepherd the pawn to the last rank.
They say the exception proves the rule. I think this "Ï'm a Ferz, Get Me into There" racing chess game might qualify as such an exception. In order to get to the goal or prevent the opposing king from reaching its square, players have to gain complete control of the neighboring square. That is similar to a checkmate condition. Btw, I hope to produce a much more sophisticated version of this game soon. Joe Joyce is giving me some excellent discussion on ideas for how to improve it. Other goal oriented games like soccer chess variants might also have this similarity. They bring the same sort of familiar focus on squares directly surrounding the king.
The defining quality as stated on this page allows (quite correctly in my opinion) king-race type games to be classified as CVs. The definition only requires that the winning condition be based on a royal piece or pieces and is not concerned with the form of that condition. "King reaching square x" is as good as "King checkmated" for qualifying as a CV. Although the form of the winning condition might be used to further classify games beyond the simple Is:Is-not divide.
I must disagree with many of the previous posters, and state a minority opinion. To me, chess must have a certain irreducible complexity. This includes a royal piece, the "capture" of which ends the game. This rules out race games. And chess must have a non-royal piece, something which may be freely captureable without ending the game. Thus, no game that has only one piece on a side can be chess, in my opinion. Further, if "bare king loses" rules are in effect, then there must be more than the royal piece [king] and one non-royal piece, because capturing the "non-royal" piece also ends the game, thus making that piece also royal. I also oppose making chess variants where there is only the "king" and one other kind of piece, as I feel it's probably too weak a game to be good. I'm also not sure about games that lack pawns. So, yes, I categorically rule out games that other posters include. What can I say besides I'm inherently conservative, and you're all wrong? Well, I enjoy most of the games I've ruled out of the chess category, they're good games [check out David Paulowich's version of the Maharajah and the Sepoys - with no queen, the game is much more even]. And I can say I also don't know where the upper boundary of chess is. But, as long as you don't treat me as pond scum, I'm happy to be a minority of one. And the games I design will only be labelled "chess variants" if they fit my definition. If you read the Notes in some of my more unusual variants, you'll see that I claim a game is only "chess-like", and not necessarily chess [eg: Chieftain Chess - because there are no pawns]. All the games I've ruled out are certainly chess-like. To me, Roberto Lavieri's Maxima is schizophrenic; sometimes it's chess and sometimes it isn't. This is because there are 2 different ways to win: checkmate of the king, which is chess; or occupying 2 "goal" squares, which makes it a race game. So here is a game that I can't say is or is not chess until it's over. But you guys are stuck with saying checkers and Axis And Allies are chess.
Always room for a minority, Joe, and I understand where you're coming from. I just like the simplicity of the winning condition test, and would probably see your stricter criteria as forming the basis for a CV sub-classification structure. Thus my CV-class-1 might be what you would call true CVs while CV-class-2 might only rate a Chess-like appellation in your book.
Just out of curiosity, how do you feel about the use of dice or other means of introducing probability? Would that stop a game from being a CV?
Joe J. wrote, "And the games I design will only be labelled "chess variants" if they fit my definition." I {g.g.} will stick to Shakespearean logic here, "A rose by any other name would smell as sweet." Joe continued, "If you read the Notes in some of my more unusual variants, you'll see that I claim a game is only "chess-like", and not necessarily chess." My {g.g.}response is that only "Chess" is "Chess" and that the CVs are simply chess-like, some more so than others…. Joe J. also wrote, "To me, Roberto Lavieri's Maxima is schizophrenic; sometimes it's chess and sometimes it isn't. This is because there are 2 different ways to win: checkmate of the king, which is chess; or occupying 2 "goal" squares, which makes it a race game. So here is a game that I can't say is or is not chess until it's over." My response to this is simple, Maxima is a "chess variant," but it is never "chess" [it simply relates to chess… because it is a chess-variant]. There is nothing wrong with a Goal win. Navia Dratp (an excellent chess variant) has 3 ways of winning… one of which is a goal line crossing and another which is based on economics. Joe J. also states that us players who disagree with him are "stuck with saying checkers and Axis And Allies are chess." Interesting, I never said that, and never heard these other CV players say that either. Yet I do recall you calling GO a chess variant (where are your Royal pieces and Checkmates in that game?) The last time I checked, your comments about GO being a Chess Variant can still be read at ChessVariants.com. Interesting here, was that I still do not consider GO to be a CV.
Grayhawke wrote that king-race type games could be classified as CVs. Yes, I agree with that. "King reaching square x" is as good as "King checkmated" for qualifying as a CV. Yes, excellent point.
And elsewhere here I read of Extinction Chess - a game with another interesting winning condition which is a far stone's throw from the checkmate concept. But it is still a chess variant. Because it is close to chess (much more so than most of my variants) and because it varies by a slight change in the rules.
A note in closing: If Joe is correct, I suppose we need to remove a lot of games from Chess Variants.com because they are not chess variants. How sad most of us haven't noticed.
"Mo-om, Gary poked me!" I think we stopped the car on that one!
Before some people take me, or this discussion, deadly seriously, let me state that Gary is absolutely right with his statement:
"[…Joe J. wrote, 'And the games I design will only be labelled 'chess variants' if they fit my definition.] 'I {g.g.} will stick to Shakespearean logic here, 'A rose by any other name would smell as sweet.'"
Quite true. When it comes right down to it, the definitions of chess, chess variants, and chess-like are made by majority opinion, and I've already admitted that I hold a minority opinion. So you guys win. Eventually. It doesn't mean I'm gonna go down without swinging, or easily. :-D
It sounds like Gary has a far narrower definition of chess that I do. To quote him:
" … [X] is a "chess variant," but it is never "chess" [it simply relates to chess… because it is a chess-variant]…"
My question is: which chess is chess? If there is one chess, and the rest are variants, and not really chess, then which is the real, true chess? Is it FIDE? XiangQi? That's played more than FIDE. How about shogi? Heck, that can't be chess, it has drops, and where did drops come from? Shogi's not FIDE or XiangQi. What is the real chess? Is it Shatranj? This is the oldest variant that we know the rules to. How about Chatarunga? That's supposed to be the original that shatranj grew from. But we don't know the rules, so we can't play it. My answer is that they are all chess, each and every one, 100% true, genuine chess. As are most of the variants posted at the CVPages. And your excellent games from Shanghai Palace [a game I think was seriously underrated when it first came out] to Shatranj of Troy, to your Six Fortresses variants, they are all chess. Chess, pure and simple. I don't see chess as "a game", I see it as "a type of game". That I define the type more narrowly than most just means that, if I do make a race game, I'll say it's "chess-like" in my notes on the rules page - a rose, by any other name…
As far as Go, I'll quote what I said a year ago - JJ 2006-04-30, Big Board CVs:
"Hey, Gary! Agreed Go is not a chess variant. It is at once much simpler and
more complex than chess. I was using it as an example of a 'large' board
game that has about the simplest, least powerful pieces possible. They
just exist, they don't even move. The game is played 1 stone at a time.
For those of us who are not experts, there isn't even a clearly defined
end to the game. But it is an awesome game, and conceptually much simpler
than chess. On a big board. Consider it a point in game-space, that
nebulous conceptual area where all games reside, just outside a boundary
of chess." [This argument was about big board games.]
Graeme, in a later comment, I put wargames just outside a boundary of chess in conceptual space, on the other side of chess from Go. Again, to quote:
"I'll… suggest… that Go pieces are only a shift from wazirs and ferzes.
In conceptual space, Go is fairly close to one 'side' of chess, and
'Little Wars' or Axis and Allies are roughly on the other side of chess,
fairly close, along the complexity line. Tic-tac-toe is on the other side
of Go from chess and the other games along that complexity line."
So, "Chess with a Combat Results Table", that is, a game that uses all the other elements of chess with a somewhat random method of determining which piece wins in a capture attempt, is exceedingly chesslike, and is probably legitimately a chess variant by most definitions. even if everybody swears it isn't, me included.
Joe: You wrote a series of questions: I will answer each peceded by GG
JJ: My question is: which chess is chess?
GG: Chess is Chess. By that I mean Fide Chess
JJ: If there is one chess, and the rest are variants, and not really chess, then which is the real, true chess? Is it FIDE?
GG: Fide Chess is "Chess." But Xianqi and Shogi are just as valid - but they, by labeling, are not Chess (which refers to Western Chess). Indeed, Xianqi is the most widely played of the three. All three have a common origin… and are variants of that from which they have evolved.
JJ: XiangQi? That's played more than FIDE.
GG: Yes, it is played more often than is chess. But, Xianqi is not "Chess (Fide Chess)" It is Chinese Chess. It is Xianqi. Besides, how often a game is played has nothing to do with making it more chess-like.
JJ: How about shogi? Heck, that can't be chess, it has drops, and where did drops come from?
GG: You are correct. Shogi is not "Chess (Fide Chess)" It is Japanese Chess. It is Shogi. And the earlier Shogi that it came from did not have drops. So in Shogi we see a variant of Chu-Shogi. Note that Shogi, Chess, and Xianqi are all variants of Shatranj. The term Chess Variant could have easily been replaced with Shatranj Variant - but many people don't know what Shatranj is.
Fergus Duniho has suggested in the past, the term Caissa to refer to the family of games, or to the platonic ideal that all chess games are physical world substantiations of. Perhaps we could shorten it to Caiss, to differentiate it from the goddess Caissa.
John William Brown uses the term Meta-Chess, which is less poetic, but more descriptive.
Using the term Chess to refer to the family of chess games seems to me too open to confusion and ambiguity.
Somehow [I won't mention the roles my wife, son, daughter and mother-in-law played] the reply I started in the morning to Gary Gifford's post got delayed in posting until hours later, after another post by David Howe [the voice of reason]. He's right, and I suggest using "Meta-Chess" as opposed to "Caiss[a]", as it lets visitors know what we're talking about.
Following is the original text of my post, warts and all:
Okay, I think we've pretty much kil… um, covered this part of the conversation. We can all declare victory and go on. [Although I think David Howe has won the field.]
It seems chess is pretty much like science fiction. Everyone has a somewhat different definition. Or like another kind of printed matter, that is quite hard to adequately define, but as the legislator said: "I know it when I see it!"
We pretty much all agree on what it's not: it's not checkers [draughts], or Go, or Axis and Allies, and that should be enough to have a good working arrangement. We all work in pretty much the same area, and generally enjoy each other's stuff; what more do we need?
Finally, in reference to the "shatranj" comment at the end of Gary's last comment, half of what I do is shatranj variants. Does he mean none of them, even Modern Shatranj, is actually shatranj? And should I move them from the "chess" site??? ;-) Enjoy!
I feel like I'm in a verbal vortex here. But it seems to be slowing down and I regret having been sucked into it. But in Joe's last post [and please note that Joe is a friend of mine - we even had pizza together once] but, he has apparently misunderstood me. And here is his point of misunderstanding, and my attempt to clarify.
JJ: … in reference to the "shatranj" comment at the end of Gary's last comment, half of what I do is shatranj variants. Does he mean none of them, even Modern Shatranj, is actually shatranj? And should I move them from the "chess" site??? ;-) …
GG: Your Shatranj-based games are not Shatranj. They are very good Shatranj Variants. "Variants." Because they "vary" from Shatranj. There is nothing wrong with that. What is wrong is to call them Chess, or to call them Shatranj. And you have the modifier of "Modern" for one, thus it is "Modern Shatranj" but not "Shatranj"… Shatranj has different rules. It is a much older game. Your "Shatranj Variants" are great games. You started a new family of chess [Shatranj] variants and should be proud of that and respected for that. I consider you the modern "father of short range games." Your games should certainly remain at ChessVariants and I never implied otherwise. I look forward to more of your "variants." Respectfully, Gary (sorry to have been misunderstood)
Gary, my deepest apologies. If you will send me the name and address of that pizza place*, I will attempt to make restitution. That last paragraph about shatranj was another example of my twisted sense of humor misfiring. Of course, it was exactly the same argument we'd just had, only substituting "shatranj" for "chess". Next time I do something like that, I'll try to smile twice at the end. :-(
I'd suggest a different concept for defining "chess variant". Rather than trying to find a single simple property that defines it, I'd start with a list of N properties (N=9 in your proposal) and say that a game has to meet at least K of them, for some to-be-determined value of K. Or I might assign a certain number of points to each property and say that a game has to reach a certain point-value to be considered a chess variant.
Hi, Mark:
What you suggest sounds like defining the conceptual space I mentioned in a previous post. I quote [because I'm too lazy to do it again specifically for this]:
"'In conceptual space, Go is fairly close to one 'side' of chess, and 'Little Wars' or Axis and Allies are roughly on the other side of chess, fairly close, along the complexity line. Tic-tac-toe is on the other side of Go from chess and the other games along that complexity line.' So, "Chess with a Combat Results Table", that is, a game that uses all the other elements of chess with a somewhat random method of determining which piece wins in a capture attempt, is exceedingly chesslike, and is probably legitimately a chess variant by most definitions, even if everybody swears it isn't, me included."
You've come up with excellent stuff before, so I'd love to see where you're taking this. I admit to figuring things "by guess, by God, and by eye". This is admittedly subjective. But I forsee squabbling over whether a particular feature gets 9 points or only 8. I'd suggest a scale with FIDE and/or other major/standard chess variants rated arbitrarily at 100, say, and then saying a 90 is very chesslike, and a 50 is a so-so variant, and a 32 is checkers or Stratego[tm] or somesuch. Luck!
Speaking purely for myself, I am no longer trying to find, but have found the simple quality that determines whether or not a game falls into the domain of meta-chess1.
I think your suggestion is a way that CVs (or should we now call them meta-chess1 games - MCGs?) might be further classified. Perhaps we could define lists of qualities - preferred, desirable, possible…etc - ordered according to importance, and then classify games by the highest list where all (or the majority?) of qualities were satisfied.
1. Again speaking personally, I'd prefer the term Para-Chess. Meta-Chess, to my mind, represents the "rules-behind-the-rules" realm and is essentially abstract, whereas Para-Chess suggests the individual real-world expressions of that abstract-realm in the form of CVs - or para-chess-games : PCGs
according to the article, and Joe Joyce, Suicide Chess is not a chess variant!
Hi, Abdul-Rahman, and welcome to the madness! Yes, you're right, I do not consider Suicide Chess a chess variant, because there is no royal piece. Missed your post when I made my last reply.
While I was typing this, I realized that my "10K" variant may well turn out to not be chess by my definition, because I have no clear idea now of how to get royal pieces into the game…
Yes, I also agree that Suicide Chess is not a CV since the winning condition is not based on a piece-class (referred to by custom as Royal)
But I must disagree with Joe when he says
as I don't see where their winning condition is based on a single class of playing piece.
In fact I think we could class Suicide Chess as a Checkers-Variant played with chess-pieces.
Sorry, Graeme; that was just meant as a potshot [aka: gratuitous insult] directed at Gary. You are, of course, right. I will try to control my twisted sense of humor better, but when we [Gary and I] get into these things, often something surprising comes out, that usually involves a checkered board… and I think he's easily one of the greatest active variantists. I'm glad to have him as a friend, and glad he's here. This won't stop either of us from poking the other to see what'll happen. If we were siblings, our parents would be stopping the car all the time.
I have to disagree. I played a lot of Suicide Chess games in FICS, and it can be easily observed that the tactics and strategies are similar to chess (not checkers.) It's a very tactical game to the beginners, but among the experts it's mainly a game of strategy.
I think the distinctive line between chess and checkers is 'variety'. In checkers, you have only one line of movement (three in Dama, but still,) one type of pieces, one type of promotees. In Chess (and Suicide Chess) you can go back and forth, maneuver, attack different points in the army [in checkers, the only weakness is a structure weakness. in chess, pawn structure weaknesses and exposure weaknesses for the different pieces, and all kind of weaknesses !].
I am not sure I made myself understood, but i think i made my point.
I will respectfully disagree with you, but you have made your point, I believe. I do think this comes down more to personal opinion than "reality", if that word can be used with abstract games. Suicide Chess has many of the features of chess, including the maneuver of different units in different ways, so the majority of people will consider it a chess variant. That I don't basically means I won't create a chess variant of that nature, most likely; or, if I do, will not call it a chess variant. I honestly believe this discussion says as much about each of us as it does about chess and chess variants. I will never actually believe that Suicide is really a chess variant, but it's close enough that I have no intention of arguing the point. In fact, I will here officially concede it is a chess variant, even though it's not. ;-)
[Yes, I want to have my cake and eat it too. But I'm on a diet… :-( ]
It seems like this discussion has been dead for many years. That is too bad! There are a lot of different insights here that deserve to be synthesized and discussed.
I personally agree with the concept of conceptual space and classifying games depending on how close they are to the rules of another game. Most variants introduce a few modifications to the standard rules of FIDE chess. Many variants turn things completely upside-down, but are still recognizable as a chess-like game. Some variants just don't look like chess at all except in some minor respect such as piece movement (I confess I've done things like that!). The point at which a chess-like game stop being chess-like, I assume, is when the distance to chess is greater than other games' distance to chess. How that would be measured, of course, is hard to say, but it seems fair to look at what makes chess chess and to tick down how many of these points are still there.
A consideration of the idea of conceptual space suggests that there is structure in chess space that can be considered somewhat similar to astronomical structures, with games, clusters of similar games, clusters of clusters, and superclusters formed around "prolific" games like shatranj. There's the beginning of a conversation on this topic under "Chess Clusters" in the Comments at chessvariants.org. However, this is an excellent place to kick ideas around. What did you have in mind, Francois? I've just begun to consider the question, and don't often have time to try to peer into all the odd little dimensions of chess' conceptual space. I'd love to hear someone else's ideas.
Consider that FIDE, Capablanca's [Carrera's], and Grand Chess as pretty much different aspects of the same game, expressed on 8x8, 8x10, and 10x10 boards. [Actually, I started with designing a series of shatranj variants, from which I got the basic ideas.] All 3 games use the same rules, the same basic pieces, and the same idea of combination of basic pieces to make stronger ones. Amazon Chess also fits in here, as do games which use different combinations of the basic pieces, for example the centaur, the knight-man, or the bishop-man, or rook-man, man being the non-royal king. These pieces are far less commonly used in games, but those games also fit in with this cluster of games.
Shogi starts its own cluster. All those monks playing those elaborate games left good-enough records to put shogi at the center of a cluster of games. Xiangqi is also at the center of a cluster, but it is much smaller, with only a few members, like Korean Chess. And shatranj is at the center of the FIDE, shogi, and Xiangqi clusters. Well, in brief, that's what I've been thinking about, and am now looking at what the coordinate axes are. Some are number of pieces, number of piece types, specific pieces, length of edge, length of side, number of dimensions, specific rules/mutators…
I ran across an interesting article with a different approach to what makes a chess variant, but this author is considering the origins of chess and only historical games. From the bottom of page 5 to the top of page 6, we get the following, by Gerhard Josten:
[http://www.wishop.com/games/Chess/Chess%20a%20Living%20Fossil.pdf]
Fig. 2: Living and extinct chess variants
But just what is the outstanding feature of all of these chess games? All chess games differ
from all other old games by one characteristic feature, namely that they have three different
types of pieces. These are, listed in order of significance for the game:
• Type 1: A relatively immobile central piece, the goal of the game being to paralyse this
piece. The way in which it can be paralysed varies. This rule does not apply in the game of
four-handed chess.
• Type 2: A number of pieces which can make various longer moves and – with some exceptions
– can move in all directions without restriction.
• Type 3: A number of pieces which can only move forwards, and if needs be sideways too.
All other characteristics of the game of chess such as the board, the symmetrical initial array,
the capture of opponent's pieces, the central piece or the moves in alternation can also be
found in other old Asian games. However, this concentration of three different types of pieces
is unique in chess and will form the basis of my supplementary thesis.
It certainly points out the differences between the historical forms of chess and some of the new variants.