OK, this brings us to the point of Fischer-Random.
The tendency that positions and game lines get repeated in many games - that would be a minus for a game. It is certainly a big minus for FIDE chess.
A system to classify chess variants
Sort:
Is this supposed to be about classifying chess variants or about determining which characteristics make a good chess game? If it's the latter, I would suggest:
A carefully chosen selection of pieces that complement each other well
Minimal redundancy in piece movement capabilities
Checkmate as the primary goal
A well tested initial setup
Within the rules, the players have concrete means to make process towards a goal
We should visualize the criteria. For this one I choose:
musketeerchess2017 wrote:
Concerning the pieces, do you think that adding for example an Archbishop = Bishop + Knight is redundant ? The chess pieces are complimentary when we put the pieces in a 5x5 checkered board with alternate pieces on the central square of that 5x5 Board the pieces are complimentary.
Yes, that's the sort of thing I was thinking of. It's not a big thing, but I think it's better to have pieces that have unique abilities than to have pieces that just duplicate the powers of others. Of course there are exceptions to this. Maybe I'm overthinking it.
It is defenitely good for a chess variant if every piece has a weak spot, which makes it vulnerable for some other piece. It is funny that there seems to be an intuitive liking amongst chess players for super-poweful pieces that can 'do everything', like the Amazon (Q+N) in a FIDE environment. But in practice such pieces do not make very good games.
Many variants do make an attempt to have a complementarity between the pieces. E.g. in Omega Chess the Knight, Champion and Wizard together cover the entire 5x5 area (plus the Camel squares around it), with not a single move in common. In Falcon Chess the Falcon was intentionally given all the moves to squares in the surrounding 7x7 area that are not reachable by FIDE pieces (i.e. the Camel and Zebra squares), but the moves were made blockable rather than direct leaps to not make the piece too strong.
In a game that typically leads to static Pawn structures (like with FIDE Pawns) it can be very important to have a certain type of move at your disposal that can undermine those. Even if only one of your pieces has it, you can usually afford the time to bring it in position; the Pawn chain is a quasi-permanent structure, and won't go away. Chess would not nearly be such a good game without the Knights.
I guess this explains the tendency in western chess variants to have pieces with many different kinds of oblique leaps, while in the large shogi variants (which often have 10 times as many piece types) such leaps are almost completely absent: in Shogi Pawns capture straight ahead, and do not form the impassible barriers they do in Chess.
Another remark: we should not discard opening theory too easily. There are actually many chess players that like opening theory. Preparing an opening line and surprising your opponent with it is a fun game, and a relatively easy way to improve your performance. But in orthodox Chess it just has gone too far. I participate in a correspondence tournament for Tenjiku Shogi now, which is a game where the opening phase is crucial, but opening theory was almost completely absent. It is really fun to devise your own opening lines from scratch. If you can just look them up in a book? Not so much...
Randomizing a game as in Chess960 so that you cannot do any perparation at all thus kills part of the fun for many. Playing from a fixed but novel setup can be better, as long as it is understood that such games have an 'expiration date', and that when too much gets known, you have to move on to something else.
I think some things that make a good chess variant are that the variant is too complex to calculate to the win, and that who wins should not be strongly tied to who goes first, or who is playing which army.
I think a good thing about fairy pieces is that you may need to try to calculate what you think the value of the piece is based on its movement as it tends to be harder to find information on their values.
I think variants that have different win conditions from standard chess can also be interesting as you have to think about a strategy for something other than trapping a royal piece in order to force a win.
I think variants that involve elements of chance or hidden information could also be interesting as you then must then take probabilities into account when making calculations, or try and figure out where the hidden pieces might be. Also synchronous chess involves both players move at the same time instead of taking turns like in standard chess, meaning that some positions may be like a complicated version of rock paper scissors, in which every legal move from both players has moves that it wins against and moves that it looses to.
In Judo Chess you move pieces continuously so being able to move faster than your opponent would be an advantage even if the game wasn't timed.
Interesting points from acgusta2. In case of any randomness, incomplete information or Judo-like rules, one can ask if you can still consider that chess variants. They are all not to my taste.
This has been a very interesting discussion. What will eventually replace classical chess is a difficult question, although it will probably be something only incrementally different from classical chess itself. Chess 960, although in many ways a very ugly solution to the problem of computer assisted opening preparation does have the appeal of not being at all different from classical chess, except for the lack of a standard starting position. It certainly should have the inside track although as HG Muller pointed out, many do like opening study and preparation, although why is beyond me. Still my younger self from the 1970s or 80s might have disagreed with the "long of tooth" veteran chess player. I certainly remember having a little disdain for such things as Transcendental chess as I pored over chess informants and other opening study guides for correspondence chess.
Simplicity, similar rules and something that has the feel of classical chess, with its pawn structures, quick development and ability to affect your partner's play from the very beginning of opening play would be important for me in a variant. Also interesting endgame possibilities, the ability to make unbalancing exchanges as well as to be able to play either in a position or tactical manner seem necessary as well.
Like or dislike in chess variants certainly is a matter of taste. Personally, I would not like to play on anything bigger than a 10x10 board. Anything beyond that probably should enter the realm of military board gaming where multiple moves are used for your pieces. Think "Feudal" (basically multiple move chess on a 25x25 field) for those of you old enough to remember that game. I have mostly played variants similar to chess. I do like Capablanca style games on a 10x8 board. They go by different names, but basically are the same thing. Again in these variants perhaps it is of some importance to have the minor pieces closer to the center for instance as in Janus or Carrera's original thinking. Schoolbook chess with the Archbishop on e1 and the Queen and Chancellor out on the flanks perhaps is one of the more interesting solutions to solving the problem of keeping all the pawns protected while keeping the minor pieces close enough to the center to effect opening play. Centaur Chess is another spin on the Capablanca world worth more play. King and Knight moving Centaurs are used in an arrangement similar to Janus Chess.
I also really enjoy playing Seirawan and Musketeer Chess. The latter is especially interesting in my view. While the version formerly played online offered 10 unique pieces of which two could be used in any given game, Musketeer chess has the potential to introduce about any type of variant piece that has been thought up. The ten pieces actually used were all more powerful than a rook, but there is nothing that would forbid the introduction of less powerful pieces say for instance like the Wizard or Champion from Omega chess, the Chinese Cannon, the Camel or other "weaker" pieces if players wished to use them in a game. Musketeer chess also has the advantage of mostly using standard equipment and boards. To play the game all you need is two other uniquely styled pieces you can call some other piece, use them for the gated pieces and still play the game.
I have a copy of Omega chess. It could certainly have benefited from using an arrangement like that used in Grand Chess, dropping the corner squares and deploying most pieces on the 2nd and 3rd ranks. As marketed, the pieces were too far away from one another and perhaps too weak for a 10x10 board. Also with the weird corner squares, not being able to mate a rook up with K + R vs K was rather strange!
HGMuller wrote:
Another remark: we should not discard opening theory too easily. There are actually many chess players that like opening theory. Preparing an opening line and surprising your opponent with it is a fun game, and a relatively easy way to improve your performance. But in orthodox Chess it just has gone too far. I participate in a correspondence tournament for Tenjiku Shogi now, which is a game where the opening phase is crucial, but opening theory was almost completely absent. It is really fun to devise your own opening lines from scratch. If you can just look them up in a book? Not so much...
Randomizing a game as in Chess960 so that you cannot do any perparation at all thus kills part of the fun for many. Playing from a fixed but novel setup can be better, as long as it is understood that such games have an 'expiration date', and that when too much gets known, you have to move on to something else.
To me, the obvious solution to over-analyzed openings is a bigger board and more pieces. Why is it that chess players tend to stick with the smallest games available?
About pieces, is a game better for having multiple pieces with similar strength, like bishops and knights?
Problem with more pieces is that it tends to make games longer. Try for instance Chu Shogi (12x12 board, 46 pieces per player).
Multiple pieces with similar strength is good, because that invites 1:1 trades between different types, leading to a larger variety of material combinations to occur often.
HGMuller wrote:
In a game that typically leads to static Pawn structures (like with FIDE Pawns) it can be very important to have a certain type of move at your disposal that can undermine those. Even if only one of your pieces has it, you can usually afford the time to bring it in position; the Pawn chain is a quasi-permanent structure, and won't go away. Chess would not nearly be such a good game without the Knights.
I can't help mentioning that the Witch is a very nice piece to have in a closed position with a closed pawn-structure.
what is happening here
His evaluation of Chu Shogi is entirely based upon his total inexperience with the game - as he admits himself. This brings us to a difficulty in evaluating 'playability'. A longer learning curve will repell more potential players, but this doesn't make a game 'inplayable' or bad in itself.
Sharper chess is fine, the limited introduction of the new pieces needs to be change tho. For me it would come next to Fischer Random.
Chu shogi is the best-balanced game I know of, despite being on such a large board. If I spent 500 years trying to find a better setup, I doubt I would make any progress. Very playable. Very harmonious. If someone can't appreciate the indirect shuffling aspect of the opening phase, then chu shogi isn't for them. I think it's beautiful.
Players engage each other when there is a clear path. Many rush too quickly to engage the opponent. Patience is required to play well. Misunderstanding of key principles is what I believe hinders those who are put off by the game from truly appreciating it. Chu shogi principles are a far cry from chess principles. Because of this, it can take a very long time to "unlearn" one's chess habits.
I like the delayed interaction. It allows players to fake each other out with bluffs and what-not.
Just because there are many pieces doesn't make the game overly complex. Pieces are clearly divided into groups, with each piece in a group serving a slightly different purpose. I do admit, it takes a long time to develop the intuition as regards which slow piece to bring out in specific structural melees: gold, silver, bronze, or leopard.
Thank you Zied for the Kaufman link. He writes there about Grand chess:
„As chu is to shogi, Grand chess is to chess. The larger board and
extra pieces (bishop + knight and rook + knight) add a whole new dimension
to the game. I'll have to give it a zero for history and tradition (a few
games by Capablanca don't qualify it here). I believe the draw percentage
would be very low among masters (I haven't had one yet), and the advantage
of first move small enough. Memorized theory doesn't exist, though it could
become a bit of a problem if the game became popular, so I'll give it 1/2
here. Variety of play is good, more than chess without reaching the point
of overkill as with Chu, but perhaps still a bit less than I would like, so
I'll give it 3/4. Game length seems about right to me, a bit more than
chess but nothing like Chu. Early interaction is the same as in chess, and
the strategical principles should be similar. So Grand chess, despite its
meager following, scores an amazing 6 1/4 out of 8 on my criteria, by far
the best so far. It really is an excellent game and deserves a bigger
following.“
Grand chess was the first chess variant from which I bought a physical board and it inspired me for a very long time, because it allowed me to play OTB with my friends for the first time. Meanwhile I have a huge collection of different boards and all chess pieces from Musketeer, Superchess and many other variations. "The appetite comes with eating".... 
I agree that Grand Chess is a very well designed chess variant. I minor criticism is that the spectrum of piece powers is a bit top-heavy, with 3 Queen-class pieces but no other additions to the FIDE army (other than two Pawns). As is common in variants with the Capablanca army, this sort of condemns the super-pieces to being traded for each other. But since the three are all different and close in value, this still causes sufficient inhomogeneous trading to still cause reasonable variety of material, in games. And the 'leveling effect' encourages trading of a super-piece for Rook + minor.
I actually designed a somewhat similar chess variant ('Elven Chess'), intended as a stepping stone for normal Chess players to Chu Shogi. This is also played on 10x10 in a setup reminiscent of Grand Chess (Rooks in corners, most pieces on 2nd rank, Pawns on 3rd), and augments the FIDE army with 5 pieces of 4 new types (plus two Pawns). But I did distribute the piece strength over the entire spectrum, from minor (two non-royal Kings) to above Queen (Lion), with Dragon Horse as one extra Rook-class piece, and Dragon King as an intermediate between Rook and Queen. The use of 'crowned' rather than 'mounted' pieces gives it a bit more Shogi character (and limits the strength a bit, as these get only 4 extra moves rather than 8), and furthermore makes them 'Lion proof'.
No one plays it, of course...
I did participate in the only recorded game of Elven Shogi on Chess Variants. It was a pleasant game. I prefer playing with shogi pawns to FIDE pawns.
My first analysis was that it would be unwise playing as black (second) to develop many pawns at all. Thus, white can get a forced considerable development advantage. I still won with black, though, so it can't be too bad.
My favorite variant is "replacement chess". Instead of capturing pieces, they must be relocated to other squares. This emphasizes position over material, and capturing a piece can be used to block a check, with the only restriction being bishops have to be placed on the same color as the square they were originally on. They should also create a supervariant. 3 dimensional, bughouse, progressive chess with 5 boards!