The site has moved to a new server, and there are now some issues to fix. Please report anything needing fixing with a comment to the homepage.

The Chess Variant Pages

[ Help | Earliest Comments | Latest Comments ]
[ List All Subjects of Discussion | Create New Subject of Discussion ]
[ List Latest Comments Only For Pages | Games | Rated Pages | Rated Games | Subjects of Discussion ]

Comments/Ratings for a Single Item

LatestLater Reverse Order EarlierEarliest
Castling in Chess 960. New castling rules for Fischer Random Chess. (8x8, Cells: 64) [All Comments] [Add Comment or Rating]
Gary Gifford wrote on 2006-04-21 UTC
So, what were we to think when Kramnik played the 'dead' Berlin Defense
against Garry Kasparov's Ruy Lopez, and won?

You have convinced me of nothing really, other than you seem to dislike
backgammon and want to say goodbye to chess...  which is fine.  Just play
the games you want. 

And us non-masters (and World Champions like Kramnik) can continue to play
'dead' openings.  

P.S. Funny that my use of a 'Dead' Grand Prix attack got me a draw with
a Master last fall (in an over-the-board tournament).

M Winther wrote on 2006-04-21 UTC
Gary, thanks for helping me make my point. You mention the Grand Prix attack and Morra gambit. But these are completely dead openings on grandmaster level. Even on an average tournament level you can hardly play them anymore, because the average amateur knows the remedy. It is very frustrating for the white player. More and more openings become dead like this.

M Winther wrote on 2006-04-21 UTC
Michael, I don't write off grandmaster's view for so little, but I tried to make a point. A technically proficient person will take a liking to games where this technical skill comes to expression. Have a look at the Omega Chess site where there are some Javascripts of Omega chess games. As I predicted, this game is quite technical. There is a lot of hopping around with the light pieces. This suits the grandmasters very well. But how fun is it for amateurs that will notoriously commit blunders within 20 moves? It is a fun and interesting chess variant, but I don't believe it's 'The Next Evolution in Chess' as they claim.

Evidently, a grandmaster will like the game because he can master the complicated tactics. But an amateur would like to be able to survive the middlegame. I contend that the next evolution in chess must be a variant with a better balance between strategy and tactics.

I think this discussion could bring us somewhere as long as we tolerate each other's views.

Gary Gifford wrote on 2006-04-21 UTC
Mats:  You wrote, in part 'Truth is that hardly anybody knows what Omega
Chess is.'   
Response: Okay.  I can help remedy that.  Those who want to read more
about it can go to this informative site:

Mats also wrote, in part:  'And this is what we are trying to remedy here
by discussing which type of chess variant could appeal to the average
Response: Chess already appeals to the average player.

Mats wrote: 'Fide-chess is approaching a crisis. The game is becoming too
Response: I disagree.  Sure, if we take the sum of man's knowledge it is
well researched.  But when we play the great game of chess we are dealing
with a single human mind.  It cannot possibly retain all that
well-researched material.  One player may know the Halloween Gambit... it
is likely that most players will not.  Even when someone plays a Sicilian
Defense... will the Smith-Morra Gambit take him by surprize?  What about
the Grand-Prix Attack?   Because we have human mind vs. human mind, chess
is very exciting.  Strategic and Tactical abilities outweight opening
knowledge.  Opening knowledge won't help you in an endgame.  It won't
help you find a clever mate-in-three.  There are over a billion possible
positions after just the first few moves in a game of chess... it is
hardly played out.  I've been reading some long lost chess documentation
from the year 1590... guess what, it is exciting stuff that can be used
again today over 400 years later.  People that want to play something
besides chess can.  I see no need to cure the non-existent illness. 

Mats wrote: In my country 50% of the players have been lost in 20 years,
for a number of reasons. And they don't turn to Makruk or Omega Chess.
My response: That is their perogative.  If they want to abandon chess,
fine by me.  But I'll keep playing it, and Shogi, and Xianqi, and Navia
Dratp, and Kamikaze Mortal Shogi, etc.  I don't need the rest of the
world to join in.  Just one opponent per game is fine.

M Winther wrote on 2006-04-21 UTC
I have tested Omega Chess. As usual the board graphics is bad, so I created a new grapics which can be downloaded here (zipped).

So how about Omega Chess? My first reaction is that it's somewhat slow. (But perhaps that good?) Secondly, there are four light piece-types (bishop, knight, champion, wizard) that have the same value and are worth less than the rook. How does this affect play? Won't the game be centered around these pieces, that is, won't the majority of the moves be made by those pieces?

Playing against the zrf I see that the engine hops around with these pieces, avoiding moving the pawns. Is there any idea to expand one's territory by advancing the pawns, or could one just as well go on hopping with the light pieces? Does anyone have any experience of this game? I don't trust grandmaster opinions, because they think that backgammon is a good game, despite the fact that it's immensely boring.

M Winther wrote on 2006-04-21 UTC
Gary, If this had been correct that chessplayers are enjoying all kinds of chess variants then everything would be ok. Truth is that hardly anybody knows what Omega Chess is. And this is what we are trying to remedy here by discussing which type of chess variant could appeal to the average player. Fide-chess is approaching a crisis. The game is becoming too well-researched. In my country 50% of the players have been lost in 20 years, for a number of reasons. And they don't turn to Makruk or Omega Chess.

Gary Gifford wrote on 2006-04-20 UTC
Mats, you overlooked the games I mentioned and brought up backgammon, not
sure why. But I did mention Omega Chess.  GM Michael Rohde (about 2695
rated)in regard to Omega Chess excitedly stated '. . . all the elements
of chess are preserved' and he stated that'. . .new tactical twists are
created by the extra pieces, larger board and extra corner squares.'  He
went on to say, 'The Wizard and the Champion complement very well and
quite entertainingly the different strengths of the Knight, Bishop, Rook
and Queen.'  He pointed out that Omega Chess groups have shown up in
Toronto, New York, Budapest, and on the internet. Susan Polgar (a high
level chess master) also likes Omega Chess.

Former World Chess Champion Kramnik played Makruk (Thailand Chess). He
played a match of Makruk against German journalist Dr. René Gralla (May
1st, 2004).  Kramnik likes Makruk.

But for now, there is really no reason to replace chess.  It is an
excellent game... there is no reason we can't continue to play and enjoy
chess along with other games like Shogi, Xianqi, and many of our
little-known inventions here.

Players will play what they like.... in time something may come along to
replace chess... but I think we'll both be pushing up daisies long before
that happens.

M Winther wrote on 2006-04-20 UTC
I already uploaded a new bugfixed version of my Barion. It plays curiously, but it's worthwhile to study the concept of potential pieces.

M Winther wrote on 2006-04-20 UTC
Gary, of course, grandmasters play other games, not the least backgammon. But this is not the issue. The question is whether we can find a powerful enough candidate to supersede chess. We sure don't wish it to be backgammon.

Another candidate with a potential for superseding Fide-chess is the type of CV:s that employ undefined or potential pieces, such as Bario, and my own variant Barion. I have improved the rules for Barion greatly and uploaded a new version today.

Potential pieces, or 'Quantums', or whatever you prefer to call them, is an area that could be researched more.


Gary Gifford wrote on 2006-04-20 UTC
We already see Chess Masters playing Omega Chess, Thailand Chess, Glinski Hexagonal, Shogi, Go, etc. So, while some chess masters may not want to play games other than chess, many are certainly willing to do so. Great Chess Masters are not simply memory banks of opening books, as some would have you think.

Mark Thompson wrote on 2006-04-20 UTC
'Chess Master/Grand Masters will never accept a new game that takes away
their book opening knowledge advantage.'

No, I wouldn't expect them to; they have too much invested in their study
of openings. But if I'm optimistic about the future of Chesslike games,
it's from hoping that the next generation, who haven't become Chess
experts, might be attracted to CV's.

John Lewis wrote on 2006-04-20 UTC
Certainly this article only attempts to remedy the flaws in FRC... and does
not propose to solve all the problems in standard 'Book Opening Chess'. 
The idea of creating a game where you may buy your pieces from a selection
of pieces is very appealing to many and deserves it's own article.  I have
for a long time studied the problem of allowing people to buy their armies
in Chess and I have various opinions I've come to based on
experimentation and dialogue with players.

Two things is certain:

1. Chess Master/Grand Masters will never accept a new game that takes away
their book opening knowledge advantage.  Particularly if played

2. Chess960/FRC is a fine game, but improvements in it will not be
accepted by it's participants any more than standard chess Grand Masters.

Mark Thompson wrote on 2006-04-20 UTC
'But clearly, Fide chess is approaching a crisis. It could soon be renamed
'Opening Study Chess'. It's becoming ridiculous. I think there are two
ways of meeting this challenge. (1) Follow Capablanca's proposal and
increase the board size, or (2) introduce a form of drop-chess along
Burmesian lines, as my own proposal Swedish Chess.'

I think there is an option (3), or at least (2b), which is what I've
called 'Mercenary Chess'. Let us start a world CV organization that
maintains a catalog of pieces, perhaps a bit less inclusive than the
Piececlopedia, but with a price for each piece, measured in points. (The
organization should have some system for monitoring the empirical value of
different pieces based on their observed usefulness in tournament play, and
adjusting prices periodically based on what they learn.) Each player starts
with 1000 points, or perhaps it should be 100 points per file on the
rectangular board chosen, and the players start the game by alternately
purchasing their starting pieces and dropping them on the board. Such a
system would be amenable to handicapping, by giving one player a few more
points than the other. Equal players might decide to give Black a few more
points to compensate for moving second.

This idea has been proposed in various forms by several people. I think I
heard that Bob Betza was first, calling an idea very much like this one
'Generalized Chess.'

M Winther wrote on 2006-04-19 UTC
Michael, those grandmasters get paid for playing FRC. Those who aren't, the bulk of chessplayers, 99.9%, are not particularly interested in FRC. What matters is what amateurs think. But I would gladly play FRC, too, if I got paid.

M Winther wrote on 2006-04-19 UTC
Adrian, to unleash their creativity amateurs cannot go on studying openings using a chess database, because then creativity and fun is thwarted. I knew an ambitious amateur (around Elo 2170) who devoted years and years to perfecting his opening repertoire. But, unlike grandmasters, he lacked the capacity to creatively improve the variants, and, by this monotonous activity, he managed to deaden his natural passion for the game. And then it became obvious that he couldn't play those variants because his opponents would prepare against his variations using their own databases. An ambitious amateur cannot afford keeping alternative opening systems, it's too much work involved. On the other hand, amateurs can seldom play small openings systems like Réti, because they lack the capacity to make the most of those small positional advantages that can be utilized in the endgame.

In his upcoming series Kasparov will discuss the 'opening revolution', which had its beginnings in the seventies. For the ambitious amateur, the present development in chess, the advanced level of opening science, is injurious to creativity and phantasy. In Rubinstein's and Lasker's time grandmasters could still play the exchange variation in the French with a good deal of success. But what's the point in feeding 25 moves in the Sicilian Dragon, against another ambitious amateur, and then shake hands since a theoretical drawn position is reached? I see amateurs do this. In orthodox chess you are cooped up in opening lines which you don't really like. You are forced to play against your own nature because there are no good strategical alternatives. Everything else is drawish. This is due to the advancement of opening science. I don't think Fide-chess should vanish, it's just that it's high time that we think about alternatives.

That's why I think that such initiatives like the Circular Chess World Championship are praiseworthy, because it speaks to our phantasy and creativity.


Adrian Alvarez de la Campa wrote on 2006-04-19 UTC
Interesting commentary Mr. Winther. I agree with you about FRC, but why do you say that FIDE chess is approaching a crisis? I don't see how extensive opening study is leading to a crisis.

M Winther wrote on 2006-04-19 UTC
Fischer Random is a lost cause anyway. It's quite good for training games, but it won't supersede regular chess. Chess players aren't fond of it. I actually asked Seirawan once, and he rejected it. Chess players want to be in control, but in FR they are always faced with unknown territory. In some of the initial positions white cannot secure an advantage. And in certain cases black's position is not tenable. To be able to study a standard opening position is a necessary prerequisite for chessplayers. It contributes to the vitality and depth of a game if it can be studied in advance. It's more fun to play also, because you can adopt your favourite attacks and defenses. In this way Fischer Random goes against the instincts of chessplayers.

But clearly, Fide chess is approaching a crisis. It could soon be renamed 'Opening Study Chess'. It's becoming ridiculous. I think there are two ways of meeting this challenge. (1) Follow Capablanca's proposal and increase the board size, or (2) introduce a form of drop-chess along Burmesian lines, as my own proposal Swedish Chess. It looks promising, but I don't don't know whether it speaks to the instincts of chess players. Nor do I know if it's complex enough, i.e., so that a multitude of deep strategies are possible.

Chess commentators have been ridiculing Capablanca for his big board proposal, and even argued that it was an expression of megalomania. But Capa was right. It's only that his proposal came too early, prior to 20th century's immense development in opening play. But now chess is approaching a crisis again. The big board alternative is underestimated because people think that its complexity is beyond human capability. I think this is a misunderstanding. This form of complexity is something which a human brain is very apt to handle. In regular chess a 'simple' rook ending could be immensely complicated, putting great demands on both understanding and calculus, whereas a complex middle game position could be much more easy to handle. This is a paradox in chess. In complicated positions with many pieces the vast majority of continuations are simply impracticable. A rook ending contain much fewer possibilities, but a much greater percentage of practicable continuations. In middle games there are many uncritical positions, which means that there are many playable alternatives. You could almost choose a move by random. This paradox, that it is often easier to play complex positions, makes the big board alternatives wholly playable, at least some of them. To many orthodox chessplayers such a game as Renniassance Chess, for instance, is unthinkable. It must be a joke. Nobody can control the complexity of this game. But thanks to this paradox of complex positions certain big board variants are probably quite manageable.

I don't think it's worth it to waste more time and energy on Fischer Random, as such. A better alternative is simply to pick good and interesting positions from the array of positions in FR and adopt them as new additional opening setups, like I did in my own proposal Chess-B. To chessplayers, this is much more fun , because then one can study and discuss this particular FR position.


John Lewis wrote on 2006-04-18 UTC
I was thinking of some of the Chess960 starting positions and I realized that some have pretty funny castling in them. For example:

rknnbqrb/pppppppp/8/8/8/8/PPPPPPPP/RKNNBQRB w KQkq - 0 1

I picked this from random selection. (Position #747)

In Chess960, you need at least five moves to castle to the right from this position:

rknnbqrb/pppppppp/8/8/8/3NNPB1/PPPPPQPP/RK4RB w KQkq - 0 1

The resulting castle would have the King move 5 spaces and the rook 2. I suppose that's okay. I see Kings make these kinds of moves all the time...wait I mean I see Rooks make these kinds of moves all the time. That's odd. Kings shouldn't be moving so far should they?

Rooks are meant to slide long distances. It's what they do. In Standard Chess Castling provides a unique extra move for the King. In Chess960 that extra movement can result in remarkable rearrangement of pieces, and rather than looking like the King is jumping over a Rook which just made a 'normal' move next to it, it looks like the Rook is jumping out of the way of a speeding King.

rknnbqrb/pppppppp/8/8/8/3NNPB1/PPPPPQPP/R4RKB w kq - 0 1

Ok, so what if you happen to be castling the other direction?

rknnbqrb/pppppppp/8/8/8/3NN3/PPPPPPPP/RK2BQRB w KQkq - 0 1

In this case you only need to move two pieces to give you the space to castle. But what's this? The King moves one space to his right and the Rook jumps three spaces. Why in the world would a rook be jumping three spaces? And why is the King moving right to castle left??

rknnbqrb/pppppppp/8/8/8/3NN3/PPPPPPPP/2KRBQRB w kq - 0 1

Granted, castling is the only time you can move two pieces simultaneously, so it kind of odd anyway, but all this Rook jumping business is very strange! In Standard Chess the only pieces that can jump are Knights and Kings... and Kings only get to do it on their first move as part of a castle.

It really seems more like these pieces aren't Castling so much as teleporting to pre-defined spaces. Which, of course, is exactly what they are doing.

In a bid to keep the game backwards compatible with Standard Chess, castling was restricted to the castling squares found in that game. It wasn't the only way to keep the game backwards compatible, but it was the one chosen by Bobby Fischer.

It also had the effect of putting the Castled pieces in familiar places to Standard Chess players. On the surface this might seem like a very good thing. Standard Chess players would be familiar with how such positions are defended and how to attack them. But I think this is counter to Bobby's own intentions. He was looking for something unfamiliar. Something where you're previous (extensive) knowledge of Standard Chess wouldn't give you an advantage. A game to put those who haven't studied endlessly in books a better chance of playing tactically with those that have.

So much for that, I guess.

In the end, a game can only be as good as it's rules.

Pete wrote on 2006-04-18 UTC
I am glad to see a simpler presentation of the castling rules for Chess960.
 The length of the initial rules presentation I had seen was rather
offputting, and what John has proposed is, in my opinion, a step in the
right direction.  

My question involves other requirements for castling.  In ordinary chess,
there is the requirement (3.8.ii.2.b) that the squares between the king
and the rook in question must be empty.  This rule ensures, among other
things, that both the king and the rook have open destination squares.  I
don't see such a requirement explicitly in John's post, and I suggest
that this be added or, if the rule is not necessary, that an explanation
of castling in a situation like BRKNNQ.. (with the spot the king would
occupy after castling already being occupied by another piece) be


Reinhard Scharnagl wrote on 2006-03-26 UTC
Hi Matthew,

Chess960 preserves the genuine asymmetrie of Chess, thus its castling
rules make sense, overmore thus becoming also a superset to traditional

But there is Chess480 with 'modern' castling (King makes two steps if
possible). But Chess480 with 'symmetric' castling, as you proposed, is
not yet a common variant.

Nevertheless you could play all those three variants at the SMIRF chess
engine and GUI. This is also true for the 10x8 board geometry.

Matthew Petersman wrote on 2006-03-25 UTC
I enjoy the Chess960 variant.  I think it is good for Chess and I believe
some form of it will be the dominant form of Chess in the future.  

I would like to propose a slight change to this variant.  For now, I'm
going to call it Chess960bg or Chess480bg.  'bg' referring to what file
the king can castle to.  It's exactly like Chess960 (FRC), but the
castling rule is slightly modified...

(1')The king moves to the b-file and the a-side rook moves to the
or the king moves to the g-file and the h-side rook moves to the f-file. 

(The FIDE castling rules 2-4 you state below still apply of course.)

I believe this could result in more opposite side castling, since the
would be the same distance from the center, regardless of which side he
castles to.  This could lead to less draws.  I also believe this would
make it simpler to teach castling in Chess960 to people not familiar with
Chess since the end result of castling is symmetrical.  

I'm curious what others think.

Thomas McElmurry wrote on 2006-02-27 UTC
The Chess960/FRC castling rule is certainly not 'overly complicated'. When it takes up the majority of an account of the rules, that is because it is explained in such a ridiculously complicated and confusing way. There is a problem here, which has probably turned some players away from the game, but the problem is in the presentation, not in the rule.

The rule itself is very simple:

The king moves to the c-file and the a-side rook moves to the d-file, or the king moves to the g-file and the h-side rook moves to the f-file.

That's it. One sentence (not including the restrictions on when it is permissible to castle, which are identical in all the rules discussed on this page).

The Chess480 rule, even though it was introduced as 'an appeal for simplicity', is no simpler, and arguably more complicated than the FRC rule.

Of course these are not the only possible rules. If I had been asked, before learning about FRC, how the castling rule should be generalized for random starting posiitions, I probably would have said that the king moves half the distance (rounded up) toward the rook, and the rook moves to the other side of the king. This rule is left-right symmetric and matches the Chess480 rule in 11/16 of the possible positions. But without the need for awkward special cases, it is in my opinion simpler.

I am predisposed to like symmetry, and it wouldn't have occurred to me to choose an asymmetric rule like the one in FRC. Yet there is something appealing about the asymmetry, particularly in this context where it produces twice as many actually distinct positions. For this reason I'm still inclined to prefer the FRC rule.

John Lewis wrote on 2006-02-27 UTC

I find Mr. Reinhard Scharnagl's preferences interesting, and I'll address each one in turn. Remember, I am biased for the 480Chess method of castling.

a) The Chess960 castling rule is consistent, in Chess480 there are small variations, when the king is near to the borders: then he will move castling one step 'shorter'.

Consistency is in the eye of the beholder. It's true that in Chess480, the king on the 'b' or 'g' file can't make a two square leap to the near side. However I find this highly preferable to similar situations in Chess960 where the King doesn't move at all. Castling without the King moving seems rediculous to me. In Chess480 the King consistently moves and normally two spaces, unlike Chess960 where the King might travel anywhere from 0 to 5 spaces. From my perspective it's Chess960 that's inconsistent.

b) As reflected in the name, Chess960 preserves the natural asymmetry of the chess game supporting 960 different starting arrays. In Chess480 mirrored positions lead to equivalent situations (thus SMIRF proposes only such randomized positions for Chess480, where the white Kings is on white Queen's right side).

It's true that there are only 480 starting positions in Chess480 (hence the name). While you can still play the game from all 960 starting positions of Chess960 for variety, strategically there are only 480 that matter.

Having said that, the stated goal of both Chess960 and Chess480 is to open the game of Chess from it's years of studied opening play. Both games do this. I personally find the asymmetry of Chess960 to be a hinderence to introduction of the game to new players. (Which leads to your final point.)

c) After castling Chess960 positions are looking more similar to traditional chess games after the opening stage. Maybe that is the reason, why the masters will stay with Chess960.

If the masters gain some advantage from the board looking similar to traditional chess after the opening stage, then I think something has been lost from the inspiration of the game. Chess960 was intended to remove the advantage that is enjoyed by those very masters who have studied endless openings. Quoting David Wheeler 'Fischer's goal was to create a chess variant in which chess creativity and talent would be more important than memorization and analysis of opening moves.'. Liberating Chess from the book openings might also require removing from possible opening transpositions to similar openings. Chess960 is much more likely to have this happen as you've noted.

I understand why some who have invested so much time in Chess960 might view Chess480 is some kind of threat. The variant is easier to understand for novices and has many advantages over Chess960. Even hardcore 960 fans admit that the castling rules are overly complicated (they take up the majority of the rules themselves)... but none of them want to leave the rich history of games played in Chess960. I can't blame them. I'm sure players of regular chess didn't want to lose their history either.

Reinhard Scharnagl wrote on 2006-02-27 UTC
There have been some quarrels on what currently is named Chess480 and had
first been intended to propose a modification to Chess960. I welcome, that
now it has lead to an own, different variant proposal. Chess480 also has
been implemented e.g. in the multivariant GUI + engine SMIRF. Nevertheless
I personally prefer Chess960 because of following reasons:

a) The Chess960 castling rule is consistent, in Chess480 there are small
variations, when the king is near to the borders: then he will move
castling one step 'shorter'.

b) As reflected in the name, Chess960 preserves the natural asymmetry of
the chess game supporting 960 different starting arrays. In Chess480
mirrored positions lead to equivalent situations (thus SMIRF proposes only
such randomized positions for Chess480, where the white Kings is on white
Queen's right side).

c) After castling Chess960 positions are looking more similar to
traditional chess games after the opening stage. Maybe that is the reason,
why the masters will stay with Chess960.

Best regards, Reinhard.

Gene Milener wrote on 2006-02-22 UTC
[1]  The historical origin of the chess castling rule is interesting, but
it is irrelevant to chess960/FRC.  Chess960 is about engineering a better
implementation of the basic game of chess.  If historical adherence is
goal, then chess960 is ruled out entirely.

[2]  The problem with the proposed rule, that castling kings move only 2
squares, is that in chess960 it would lead to an even higher draw rate
than we suffer in traditional 'chess1'.

Opposite wing castling is far too infrequent in chess1 (8% of all games).

Yet games with opposite wing castling have a lower draw rate.  Chess960
should be engineered to increase the rate of opposite wing castling, not
reduce it.

In my new book 'Play Strong Chess by Examining Chess960', I discuss all
this, plus a modification to the chess960 castling rule that would
slightly increase the rate of opposite wing castling.  (Search
or etc for 'chess960', or visit

Thank you.

25 comments displayed

LatestLater Reverse Order EarlierEarliest

Permalink to the exact comments currently displayed.