Fischer said

In chess so much depends on opening theory, so the champions before the last century did not know as much as I do and other players do about opening theory. So if you just brought them back from the dead they wouldn’t do well. They’d get bad openings. You cannot compare the playing strength, you can only talk about natural ability. Memorization is enormously powerful. Some kid of fourteen today, or even younger, could get an opening advantage against Capablanca, and especially against the players of the previous century, like Morphy and Steinitz. Maybe they would still be able to outplay the young kid of today. Or maybe not, because nowadays when you get the opening advantage not only do you get the opening advantage, you know how to play, they have so many examples of what to do from this position. It is really deadly, and that is why I don’t like chess any more.


[Capablanca] wanted to change the rules already, back in the twenties, because he said chess was getting played out. He was right. Now chess is completely dead. It is all just memorization and prearrangement. It’s a terrible game now. Very uncreative.

and (emphasis mine)

I love chess, and I didn't invent Fischerandom chess to destroy chess. I invented Fischerandom chess to keep chess going. Because I consider the old chess is dying, or really it's dead. A lot of people have come up with other rules of chess-type games, with 10x8 boards, new pieces, and all kinds of things. I'm really not interested in that. I want to keep the old chess flavour. I want to keep the old chess game. But just making a change so the starting positions are mixed, so it's not degenerated down to memorization and prearrangement like it is today.

2 Questions: Does shogi, xiangqi or other co-hyponyms suffer any of the same problems that chess does, according to Fischer, which led him to invent chess960? Are there any analogues of chess960 in shogi, xiangqi or other co-hyponyms?

Notes to explain what I understand and to hopefully further clarify this question:

Note 1: I mean I really don't see any difference. They're abstract strategy mathematical board games that have pieces and stuff already on the board (unlike go) and there's always the same starting setup, soooo...

ETA Note 1.1: There seems to be an answer about how xiangqi is indeed like standard chess and yet another an answer about how shogi isn't. Most of what I know about shogi is from Shion no Ou, but I really do not see how standard chess is unique in this or at least how shogi isn't the same this regard. Again, they are all mathematical and abstract strategy games that involve pieces on a board and have the same starting position of the pieces. (I believe it's the same with draughts/chequers, but of course in this case all the pieces at the start move the same). In each game, opening theory is I believe not only extremely viable to study but will dominate a very large portion of what a player ought to study.

ETA Note 2: Actually, I think the same problem will befall, for example the chess variants closer to the standard (in terms of the pieces used), such as 4 player chess, bughouse, crazyhouse, torpedo chess (the torpedo pawn thing), '5D Chess with Multiverse Time Travel': The initial setup is always the same, so it seems like, eventually, it's just gonna be about openings. Of course, each of these chess variants can be fixed in this regard by doing chess960 type shuffles or something.


For shogi and xiangqi, the answer is no. There is no analogue for western Chess960 in those chesses. Disclosure: I'm a decently strong player in all three.

The problems Fischer identified in those quotes you gave are that western chess favours the player with more opening preparation, and that there is no room for creativity. (Bear in mind that this is his own view as an insanely strong player who had reached the top level of the game back then.)

To answer the question from a broader context, I will give a little bit of history and cultural context to show why shogi and xiangqi, despite also suffering from the "dogma of opening theory", never arrived at randomisation of the start position as a solution - indeed, the opening theory is not even perceived as a severe problem.

(Also, I will keep the context to professional/master players; at the amateur level, really anything goes for openings in all of the chesses.)


Opening preparation is indeed important at the top levels. Historically the game has progressed from slower, cautious buildups favoured in the 17th-19th centuries, towards the more modern, faster game of the mid-20th century to today.

There are very sharp openings where one/both sides start attacking early and both players must walk tightropes (e.g. New Rapid Ishida; yokofudori/Side Pawn Picker; some lines of Gokigen Central Rook).

There are openings that have been examined to great depths (60 ply, even 80 or more ply) and involve many move order subtleties (Double Yagura games; Bishop Exchange; One-turn-loss Bishop Exchange).

There are openings which have gone even further, to survive in the ecosystem around it. (Fujii System was analysed literally all the way to mate in some lines by its inventory, Fujii Takeshii 9-dan.)

But yet, in the middle of all this, there is still room for creativity and variety. By the 1990s, of the two main styles, Static Rook was seen as the superior choice over Ranging Rook. The great player Habu Yoshiharu sparked something of a renaissance by playing Ranging Rook as an upstart young challenger in a title match - and winning. Since then different Ranging Rook styles, and even double Ranging Rook games, have never truly fallen out of favour at the professional level, even when engines overtook humans and universally preferred Static Rook.

There is more than enough to analyse, and still more creativity to be had (especially in double Ranging Rook games) - there was not a need for randomised shogi. This is of course not even mentioning arguments from tradition - there is no way that one can redo national culture by randomising the start position.

(Oh, and yes, randomised shogi would heavily disrupt existing opening theory, at least with regard to concrete moves.)


Even before I talk history, the simple logistics of it is that xiangqi pieces don't randomise well. The kings, advisors and elephants are bound to the few points they can leap to, and are almost worthless if randomly shuffled away; the cannon movement means that they also don't randomise well, since on the first move already captures may be possible.

Now the history. Again, opening preparation is important at the top level. This should be no surprise again, as top players do like to seek any advantage they can get.

In ancient times (think 800 years ago), xiangqi was thought to be a win for red, the first player - classical manuals do some analysis on opening traps and games for red to win with central cannon versus screen horses, or opposing cannons (either one).

Fast forwarding to more modern times, there was the same sort of status quo from chess and shogi at one point: it was thought that black had trouble neutralising red's initiative, and the only defence against a central cannon was screen horses. The theory debates in this flowed from the 1930s onwards in national tournaments in China, with multiple red and black setups tested in battle.

In the 1960s, Grandmaster Hu Ronghua decided to revisit a "busted" opening. Was the Fan Gong Ma defence for black really losing? With tons of effort into research and with sheer playing strength, GM Hu Ronghua managed to bring the dead opening back to life, winning big tournament games with it and drawing other masters to try it. Again, this brought forth another period of experimentation that has not quite died down; even today, masters are exploring more than ever other red openings than a central cannon, with more fertile ground waiting to be explored.

  • thanks Remellion. 'if randomly shuffled away' --> well it doesn't have to be purely random? even chess960 isn't
    – BCLC
    Jan 22 at 8:12
  • 1
    'This is of course not even mentioning arguments from tradition - there is no way that one can redo national culture by randomising the start position.' --> wait what are you talking about? no one's replacing anything. it's additional. chess960 doesn't replace standard chess
    – BCLC
    Jan 22 at 8:27
  • 1
    @BCLC For the cannons, imagine the starting position includes a pair on one of the files where they normally are. That's already probably a good initiative if not material win. Even randomising just the 6 big pieces can create this sort of situation. As for the shogi traditions - "960", or indeed most new shogi variants won't really take off because the history and culture of hon shogi (the "standard" shogi) is just too big. Eastern cultures value tradition a lot, and so newfangled things may be automatically rejected.
    – Remellion
    Jan 22 at 16:29
  • 1
    I bow to your superior knowledge of Shogi, but Xiangqi does indeed have variants: see my own answer
    – Laska
    Feb 3 at 8:43
  • 2
    @BCLC I think it's fine to introduce both shogi and xiangqi (and might include Indian, Thai, etc too). You will have to unpack your 2nd question for me. What do you mean "mathematically speaking" and "bound to end up like". In Shogi, all the pieces get recycled like Crazyhouse, so that makes the game have a fundamentally different feel. In Chinese chess, the kings are stuck in their palaces so never really roll-up their sleeves and get involved in the endgame. Also no notion of stalemate being a draw. So these factors make the games very different. Like Magic the Gathering vs Hearthstone
    – Laska
    Feb 3 at 10:00

Randomizing variants of Xiangqi

Xiangqi has numerous variants: here are three which involve randomization.

Lucky Xiangqi

When I lived in Hong Kong, I saw numerous old guys playing Xiangqi in the villages - while their wives played Mah Jongg back home, I conjectured. About half the time, they would play a randomized version of the game, where all the pieces of a side except for the king start upside down, but on the 15 usual starting squares. Neither player knew which was which. Unlike Chess-960, there is no correlation between the two players' arrangements, so one might call it Xiangqi-115967627816040000, or if on friendly terms with it, Xiangqi-1.16E+17 for short.

Detailed rules of Lucky Xiangqi

According to a friend (ex-Hong Kong Xiangqi champion), a piece makes its first move according to its starting square and then flips to discover what it will be for the rest of the game. So a cannon on a horse square makes its first move as if it's a horse, and then behaves as a cannon for the rest of the game. The boundaries for elephants and advisers no longer exist. Pawns still may move sideways only when they are on the far side of the river, and the king cannot leave the palace.

There is now a random element, but this plays into the strategy. For example, if I have a chance to trade an unknown piece for a known piece, what are the relative values? This also removes both parity & boundary issues associated with elephants and advisers, which had always struck me as rather dull. As the game advances, more information is revealed and the game converges to regular Xiangqi, but in kinds of positions which are not normally reached orthodoxically.

The random element would not stop this version being classified in the Wikipedia list of Abstract strategy games that are not combinatorial. This category is defined as "including hidden information or set up, random elements (e.g. rolling dice or drawing cards or tiles) or simultaneous movement". Other hugely successful such games with a random element include Backgammon, Dominoes, Mah Jongg, Royal Game of Ur & Stratego.

My contact says Lucky Xiangqi has been developed for fun only, not because the opening theory of the orthodox game had been exhausted.

Here's a link which confirms that this is indeed a thing, but gives no more details: Lucky Xiangqi.

Formation Xiangqi

According to this site a version similar to Chess-960 does exist.

Detailed rules of Formation Xiangqi

One player's pieces are jumbled up, then placed randomly on one side of the river, except for the generals and advisors, which must be at their usual positions, and the elephants, which must start at two of the seven points they can normally reach. The other player's pieces are set up to mirror the first's. All other rules are the same.

There are 873,180 different starting positions. I have never seen this played. It retains the dull limitations on elephant and advisor movement of orthodox Xiangqi. I might offer a novel sub-version of Formation Xiangxi, "Reformation Xiangqi"? under which one can shuffle all 15 non-royal units, and ignore palace and river (except for king and pawns). But Lucky Xiangqi is much more engaging, as its popularity shows.


For completeness, I should also mention Banqi (aka: Half Chess, Blind Chess) although this is completely different game, that just happens to use a Xiangqi set and board.

It is played on the 8x4 squares of a half-board, rather than on the 10x9 vertices of a full board. Movement is just to an orthogonally adjacent square, but capturing depends on unit type (so: similar to Jungle, or to a lesser extent Stratego). Since every square is occupied at the beginning of the game, it is very random & chaotic like ferrets in a sack. But it works very well. I've actually played this one in Taiwan, back in the day, but I've never seen in it in Hong Kong.

  • 1
    @BCLC Thanks for your questions. One was "Are there any analogues of chess960 in shogi, xiangqi or other co-hyponyms?" Yes: I have described one. "Does xiangqi suffer any of the same problems that chess does, according to Fischer, which led him to invent chess960?" Probably, as random xiangqi exists. But I can clarify further if my friend Teddy gets back to me. I don't get your point about abstract strategy. Your use of "err...I guess" suggests you're sure I will get it, and you are being timid out of care for my feelings. Honestly I don't get it: so please explain in detail! Thanks!
    – Laska
    Feb 2 at 14:05
  • 1
    Oh wow. Irrelevant to the topic, but this answer immediately reminds me of “Blind (Chinese) Chess” where you put all pieces face-down in a 4x9 grid (say, using the squares, not intersections, of half a Xiangqi board), then you start by flip over pieces. The first piece flipped determines the colour of the first player. All pieces can only move one square in a cardinal direction (no diagonal). Un-revealed pieces cannot be taken. Stronger pieces can take weaker ones; it goes King-Rook-Horse-Cannon-Advisor-Elephant-Pawn, except Pawn can take King! The player to lose all pieces lose the game. Feb 2 at 19:32
  • 1
    i wouldn't quite consider what you described as like chess960 because chess960 is still abstract strategy (conditional on the setup) ? i'm looking for abstract strategy shuffles not just any particular variant, especially things with hidden information or luck/randomness (except the initial shuffling). the idea is to have the same flavour but reduce power of opening theory. maybe i don't quite understand what you mean. all the pieces are face down and then you flip them up and move right?
    – BCLC
    Feb 3 at 3:19
  • 1
    @BCLC please see my expanded answer. According to wikipedia, an abstract strategy game can still involve random elements including hidden information, but it is no longer "combinatorial". Lucky Xiangqi certainly has the same flavour as regular Xiangqi, but removes the power of opening theory, and the random element enriches strategic considerations. Very elegant
    – Laska
    Feb 3 at 3:47
  • 1
    Laska, ayt thanks. i'll have to read more into your answer later on
    – BCLC
    Feb 3 at 9:50

I don't really know either game that well, but I have played both.

Xianqqi has lots of weirdly place specific rules. For example, the kings are restricted to a 3x3 square in the middle of the board and aren't allowed to move out of it. This makes it hard to do a chess960 type setup for it. On the other hand, with only 5 pawns for 9 rows, the restriction on king movement, and the cannons, it's a very tactical game compared to chess.

Shogi on the other hand is extremely positional at the start. The king is the third most powerful piece on the board. Most pieces can only move one square, and only in restricted directions. There is a good deal of opening theory, but I'm not sure rearranging the pieces on the back rank would really change the opening theory all that much, because a lot of the opening theory is about taking fifteen moves to rearrange your king and some of your pieces into a very defensible position, and changing where the pieces start would just mean a different fifteen moves to put your pieces into the same very defensible position. With only two long range pieces, your opponent can't do much to disrupt your opening.

  • thanks. 'the kings are restricted to a 3x3 square in the middle of the board and aren't allowed to move out of it. ' --> so why don't you just not randomise this part? 960 doesn't randomise everything
    – BCLC
    Jan 22 at 8:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.