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

Later Reverse Order Earlier
Witch's Chess A game information page
. Two or three player hexagonal variant.[All Comments] [Add Comment or Rating]
Ralf Hansmann wrote on 2013-12-12 UTC
Ok, I have now changed the description of the castling rules on: to: "Analogously to the conventional chess, there exists for the raven (king) the possibility of castling, i.e. to exchange the position with one of the towers through one single move. Like in the conventional chess castling is only possible, if both figures (tower and raven) have not previously been moved, and the field inbetween them is not threatened by the other party (respectively the two other parties in the three player version), and the field inbetween is not occupied by any figure." Is this formulated well? (I am not a native english writer)

Ralf Hansmann wrote on 2013-12-11 UTC
Hi there, thanks for your comments. I have now just looked at the (my/our) rules for castling as formulated at Actually, I also find that these need improvement... What also surprised me, is that they are not consistent with the castling (Rochade) rules we formulated in German language on These are formulated in closest possible analogy to the real chess, which means that castling is only possible if tower and raven (king) have not been moved previously; no field between them is threatened by the other party (respectively the two other parties in the three player version), and if no field inbetween is occupied by any figure ("Es besteht für den Raben die Möglichkeit zur Rochade, d.h. zum Stellungstausch mit dem Turm innerhalb eines Zuges -- analog zum herkömmlichen Schach. Wie dort ist diese hier nur möglich, wenn der Rabe im Moment des Zuges nicht bedroht ist, wenn weder der beteiligte Turm noch der Rabe selbst sich zuvor während des Spiels bewegt haben und wenn das Feld zwischen Turm und Rabe weder bedroht noch besetzt ist.") I think it is best and I will adapt the english rules to a correct translation of my German website version of the rules and hence, to closest anology to the real chess. It is some years ago now that the rules were put online in english, and I was not alone in developing the game. So this inconsistency between the english and german formulation of rules may be due to a discussion on whether to allow castling in the three version game also in case of a threatended field inbetween tower and king to give some flights possibilities. However, I now think that the rules should be formulated in closest possible analogy to the conventional chess rules in both the two and three player version.

(zzo38) A. Black wrote on 2013-12-09 UTC

This looks good to me, however I do not quite understand the rules for castling. Does the raven and tower simply exchange positions (unlike FIDE)? Does the space in between need to be vacant? What about castling through or out of check? Is this allowed in both 2-players and 3-players game, or is only possible with 3-players game?

If you like the game where colorbound pieces are available for all color of cells, you may play the variant:

  1. Replace the witch in the starting position by a third broom.
  2. Once per game you can change one of your own brooms into a witch, either as a move or immediately after moving that broom (whether or not it is a capturing move).
  3. Even by promotion you cannot have more than 1 witch, 3 brooms, 3 towers, and 3 bats, on the board at one time.

Ben Reiniger wrote on 2013-12-07 UTC
Thank you, it has been fixed (both in the page and in our index).

Aquarius1970 wrote on 2013-12-07 UTC
This game is described as to be played on a board with 81 cells on you website, but actually this board contains only 61 cells. Could this mistake be corrected by the ones, it may concern, please? Otherwise it might be not found by those who search it. Thank you!

Ralf Hansmann wrote on 2004-09-16 UTC
That's a lot of questions, Charles. However, I am very pleased to answer them, as I gladly learned from you - to my own surprise - that the pun of the hexagon leading to a witch's game does somehow work in the English language, too. Well, the original German 'Hexenschach' names of the figures are Rabe, Hexe, Besen, Fledermaus, Turm and Kröte. You can find these names together with the German names of the conventional chess figures on the game's original webpage <a href=''></a>. Concerning the two besoms I want to refer you to the FIDE chess rules, which explicitly state (see Article 2.2: 'At the beginning of the game ...') that each party has two bishops. We wanted to make as few changes as possible to the FIDE rules, which do not raise the issue of whether the bishops cover the fields of all colours. Moreover, we did not want to have too many figures because early experiments with really large boards and many figures led to everlasting games. For reasons of symmetrie, we left the centre field uncovered. However, to tell you the truth, that's not all. I mean, the proof of the Pudding is in the eating. Thus, I have just added on the <a href=''>Chess Witch's Homepage</a> further information as to what happened to the third broom.

Charles Gilman wrote on 2004-09-15 UTC
Thanks for pointing out the fact - that I should have thought of given the
surname - that the names originated in a different language. It has JUST
saved the variant a rating of poor from me. Out of curiosity, what are the
original names? In English hex means a spell, so the pun does survive
translation. You have still not responded to the point about about having
only two pieces of the colourbound type, but I notice that the colour not
covered by them is that of the centre cell. Is this deliberate?
	While agreeing that an inventor should not rate their own variants I can
think of circumstances where multiple ratings are valid, which I have used
myself: (1) cancelling out a previous rating after a change of mind or
reading the rules more carefully; (2) uprating a Good to Excellent after a
change of mind; (3) adding a second Excellent as a tribute for inspiring a
further variant by the commentator. Does it really take as many as 9
Excellents to cancel out 10 Poors? I had assumed that it would take only

Michael Nelson wrote on 2004-09-14 UTC
Ralph, <p>It is an unwritten law that a game's author does not give it a rating. M. Howe's second 'Poor' rating was not a rating of the author or even of the game itself, but was intended to cancel the 'Excellent' that it should not have been given. <p>It is perfectly proper for the author to comment about the excellence of his game, but the author should always give it a rating of 'None' <p>It is also unwritten law that a given commentator does not rate the same game more than once (M. Howe is not violating that as his second rating was only to cancel yours--similarly if a person who disliked you gave your game 10 Poors, you would not be out of line giving it 9 Excellents to cancel the excess).

Ralf Hansmann wrote on 2004-09-14 UTC
Well, I just stated my true, subjective opinion, without any exaggerations. But anyway, <b>the game</b> should be evaluated by the ratings, and <b>not</b> its inventor. I must admit that the naming of the figures is not as creative, as it might appear on first sight. It is simply a word play. Hexe is the german word for witch, and therefore, as the board is hexagonal, we have named the figures like the figures surrounding a Hexe. Word plays of course loose their charm when being translated to a foreign language, but the renaming of the pieces was not simply to something cutesy. I can see a lot of creativity here.

Anonymous wrote on 2004-09-13 UTCExcellent ★★★★★
Well, the game was designed without knowledge of the previously existing games. The similarity of the result of the design process with those other games (which we obviously recognized at some point in time) for us seemed to be a proof for the quality of our considerations. We started with a very big board. The small board was chosen for the final project as the lack of space creates conflict. The three person chess game was in the foreground, and here the defensive strategy not to attack is obviously helpful. Thus we wanted to maximize the potential for conflict and the motivation for the participants to attack. The rule that when capturing the king (raven) of a party, the figures of that party can be gained (controlled) also adds to the motivation to attack. Thus, I think it is an excellent hexagonal three person chess version, because it is playable. Of course the honour to be the first variant is reserved for Glinski's version.

Charles Gilman wrote on 2004-09-13 UTC
On the whole I am sympathetic to Michael Howe's view. It is not
necessarily the case that Ralf Hansmann based this variant on, or even
knew about, Glinsky's game; he might easily have devised it
independently. However, the pont is that Glinsky's variant already exists
and this is not an improvement on it and adds little.
	Regarding the different piece names, they at least emphasise the fact
that they are not quite the same as FIDE pieces. Indeed I might even have
adopted some of them for my own hex terminology had this page been here
when I began contributing. Toad might be worth making standard for so
distinctive a piece, if it not used for any other. Does anyone know if it
is, or prefer it to my suggestion of Migrant?
	There are however problems even with that good points. Some of the names
are used differently elsewhere, Bat and Broom have the same initial with
no obvious rule as to which (no pun intended) alternative letter should be
used for one of them, and the point that the 'Broom' is not a FIDE
Bishop is somewhat neglected by the use of only ytwo of them when three
are needed to cover the board.

Larry Smith wrote on 2004-09-12 UTC
Whether or not a game contains pattern-bound pieces for each of its
patterns, being the diagonal in this case, does not negate the value of
the game.

In this case, the game begins with only two 'diagonal'-bound pieces for
each player on a field with a possibility of three.  In both the
two-player and three-player games, each side has these pieces on similar

If a game was composed of pattern pieces and each player had sole command
of particular patterns, it might be argued that there was an increase in
the potential of 'draw'.  But this also does not negate the value of the
game.  It would definitely impact the play of the game. 

In fact, a player can take advantage of a pattern in the game if there are
fewer piece which operate in it.  In this game, it could be seen that the
pattern-bound pieces initially 'push' the goal piece into the remaining

The un-qualified requirement for all pattern-bound pieces to be
represented for each and every pattern is subjective.  Even in this game,
there is the potential of obtaining a pattern-bound piece for the
un-occupied pattern with promotion.  Granted that this is very un-likely
since a player would most likely opt for a more powerful piece, preferably
one that is not restricted to a pattern of the field.

12 comments displayed

Later Reverse Order Earlier

Permalink to the exact comments currently displayed.