Comments by JorgKnappen

Ralph Betza gauges his armies based on play between humans, this is an important point. Human chess masters play the FIDEs much better than the different armies, therefore there FIDEs have a bonus, and the unusual pieces have a malus. Thus Ralphs estimates are still valid for human players, unless we have grandmasters of Chess with different armies.

As far as I can see, the name 'vicount' (without an s) is indeed Peter Aronson's invention. The piece itself is absent from Derzhanski's list ( http://www.chessvariants.org/piececlopedia.dir/whos-who-on-8x8.html ) and from Töws' Generic chess piiece creation system ( http://www.chessvariants.org/unequal.dir/genericchesssystem.html ). I recommend a look into the latter for some other names for rarely used pieces like Boxer (Commoner + Beaver) or Foursquare (FNLJG in Betza notation, L is Camel, J is Antelope, and G is Tripper).

Derzhanski's list also gives sources to games, Töws obviously draws from similar sources, but does not give them.

Hi, Charles, I think you are supposed to follow the link and undergo the 'purchase' procedure, as there is currently a special offer at $0.00 for this nice chess variant.

There is far too much in it to summarise it here in a comment. There are 6 different armies (one of them the FIDEs), and each of the new armies is equipped with 'mutators' (as George terms it) that still feel fresh and unusual. 

Besides the new armies, the general rules are extended by new ways of winning the game, bringing down the number of draws, and duelling (introducing a poker-style element). So Chess 2 is different even with FIDE vs. FIDE armies.

This is a double excellent for this game.

The first excellent goes to Hans for digging this game out and posting it here.

The second goes to the game design: The inventor has really thought hard of the initial position, and it makes a fine game. There is a lot of tactical tension in the setup, but there are no obvious exploits. Great game!

Finally I found a work-around that allows me to update the information here, editing in the missing user name to submission form. The forward Ferz move is now added to the Colonel, as it should be.

The piece I inadvertedly created, the Charging Chancellor, may also be of the right strength to tone the Nutty Nights down. An army with srunken Nights and Charging Chancellor together will probably already being on the weak side, but inside the error margin of my estimates.

Grandmasters of the Nutty Nights may prefer this very weak version.

The lore apprentice (Zauberlehrling) of Lines of Relay is an offshot of the querquisite, I conceived it by misunderstanding its moves. See

http://www.chessvariants.org/index/msdisplay.php?itemid=MSlinesofrelay%28l

Hi Charles,

I skimmed through my sources again and could not locate a reference to 'gazelle' as a synonym for the okapi (Knight-Zebra-compound). Where did you get it from?

I share your feelings with respect to the name 'prince': it is quite unspecific and used for a whole bunch of pieces, most notably the commoner. The aim of my comment was to give a summary of this important source.  Note that Jelliss gives prince for Knight+Ferz (not Knight+Wazir) following the male/female logic from the King/Queen pair. And problemists call the Janus/Palladin piece princess.

The name hospital(l)er is  worth mentioning for the knight-alfil compound, kangaroo as a synonym ist certainly worth mentioning, too. Outback Chess was a contest winning variant, at least.

Going through my sources I found a synonym for the newt: it is called counsellor in Quang Trung Chess and this name propagated in Töws' piece creation system and in Ivan Derzhanski's list.

I have an answer to your 2nd question:

The R3 is named Kongo (Guardian of Gods) in Taikyoku Shogi.

There's also a B3, named Rikishi (Wrestler) in Taikyoku Shogi.

Note that the Taikyoku Shogi pieces move differently then pieces with the
same name in other larger Shogi variants.

Thanks, George

The only measure to get solid information about the strength of a certain piece is to playtest it (in human play and/or computer play). As far as I know, there are few divergent pieces really tested.

Mike Nelson's Seperate Realms Rook comes out on a value of about 75% Rook plus 25% Dabbabarider, suggesting that capture constitues about 3/4 of the piece value, while moving alone constitutes about 25%. But this particular piece has rather similar move and capture patterns; the more the patterns diverge, the more playtesting is needed, and there is probably no simple formula describing the results.

Note also that the results of playtesting depend on the testbed: What board is choosen, what is the oppenent's army, what pieces are in the same army (sociability effects).

David, that's a aice X-mas present for all of us!

There is one minor poiint I want to correct (it is not your fault; already Jeliss has this particular sentence wrong):

>null piece (Exotic): n. Piece that, combined with any other piece leaves it unaltered = Dummy = (0,0) leaper. Not the same as zero. Link(s): All the King's Men. [index] Source(s): [ATKM] Unique ID: [CVT-10825]

A dummy is a piece without any move; as a leaper it is an \emptyset-leaper, not a (0,0)-leaper. A (0,0)-leaper has one move to its starting square, essentially the piece-bound licence to pass a move.

This is always confounding to see that zero is something different from the emptyset.

Ah, and than there is another german word, related but not too close,
'patzig' meaning stroppy, snotty, bolshie. It describes the typical mood
of teenagers very well.

A maybe funny question:

What happens to a pawn when it is converted into an undead piece? Say, a white undead bishop attacks a black pawn: Does the black pawn turn into an undead white pawn or does it turn into an undead black pawn under white's control? The difference between the two is the inherited directionality: By becoming an undead white pawn the black pawn turns 180 degrees and marches against his old camp; while an undead black pawn under white's control continues to march against the white camp (and may block white pawns in their progress).

Another question: 

Can an undead piece attack a life piece of its own camp? Because this action may strengthen the piece by taking its life it might be a favourable move.

I suggest that you take a chess board, two kings and some piece representing the rhino (a knight works well for me) and try it out.

I must admit, that I didn't analyse the specific position nnz gave here, the placement of the supporting King is a bit awkward here. Both of your proposals to repair the situation work out fine.

For similar pieces: A gnu and a king can drive the king to the edge, but cannot give mate; the same holds for the quintessence (of quintessential chess).

Charles, after all, you are the game designer and it is your decision (after, I hope, some playtesting) what design you want to implement.

With the King being able to cross the channel, additional rules may be not necessary, but it is worth to think about game endings and how to judge them in any game design.

Adding new methods of capture (like overtaking [2/3] and approach capture [4/5]) to the dragons is certainly a good idea, you may also consider igui capture (killing a piece on a adjacent square without moving).

This will correct the balance of pieces protecting another piece over the channel on the ferry from 1:3 (Queen vs. Queen and 2 Rooks) to 3:5 (Queen and 2 Dragons vs. Queen, 2 Dragons, and 2 Rooks)---a balance I'd consider still very unbalanced, since Wales can try to sharpen it by exchanging Queens and Dragons.

I am updating the rating to excellent, because seeing the elegance of the knave and debtor pieces is obviously a non-trivial task. In fact, Abdul-Rahman Sibahi and Joe Joyce discussed the pieces (without naming them) here briefly

http://www.chessvariants.org/index/listcomments.php?itemid=MPcomplementarit

but they didn't see that they are exceptional. Perhaps it needs some hexagonal thinking to see it.

Nice to see you back again!

@Christine: Unfortunately, I do not even own an operating system necessary to run Zillions of Games; therefore I cannot make and test zrf files. 

@Charles: Yes, replacing Knights and Alfils with Crabs and Alibabas should make another army for Shatranj with different armies. Now waiting for more radical approaches touching the rook and the wazir, too.

In fact, a bishop limited to 3 steps can be found in Taikyoku Shogi (The
monster version) under the name Rikishi (Wrestler). On the other hand, a
piece named Rikishi or Wrestler with different moves occurs in Tai Shogi
and Maka Dai Dai Shogi: Here it is a limited bishop who can move also one
square horizontally. In Ko Shogi, the Rikishi or Wrestler is lion mover. So
sticking to Bishop's Dog for the limited bishop is probably the best
choice.

This also gives me the opportunity to advertise Derzhanski's list again:

 http://www.chessvariants.org/piececlopedia.dir/whos-who-on-8x8.html

The huge advantage of this list that you can actually find pieces when you
know how they move; either
by searching the move notation or by estimating their strength and scanning
the appropriate strength region.

Great work! You should submit your package also to CTAN; 

http://www.ctan.org

there are some more chess-related TeX and LaTeX packages (and as a bonus, a Go package), but they don't handle chess variants AFAIK. (Chinese Chess is there; Shogi is missing on CTAN).

I vote for the cat, alltho the other images are also good.

Fergus,

you underestimate the strength of the fairy problemist's community very much. Fairy chess problem do not only occur in the mentioned journals (note that I posted recently a list of fairy pieces and fairy conditions hosted by Die Schwalbe); they are even included in FIDE albums (where you can earn points for becoming a chess problem grandmaster).

And there are lots of active composers for fairy problems around, they are creative with new pieces (like the Zabel pawn), and they are younger than Dawson :-)

If you are really interested in chess problems, the language of the journal is a minor issue: The diagrams speak for themselves and the conditions or special pieces use some standardised notation. The journals or problem books often have bilingual or trilingual glossaries.

I don't think we should see problemists and chess mathematicians as a kind of enemy camp. And: As long as chess variant inventors are ignorant to prior art, why should anyone else care. A striking example is Seirawan/Harper chess: Allthough there are established and sometimes even well-known names for the Rook-Night compund and the Bishop-Knight compound, they came up with new names. Even worse, their new names are established for pieces with other moves.

As I say: With their (peer reviewed!) journals problemists have a far better infrastructure than chess variantists can even dream of. We have this web site and wikipedia. Neither of the two media is peer-reviewed.

We tacitly assume that strength can be measured by one number and that the numbers can be compared using a transitive relation like 'greater than'. However, this needs not to be true, see here for a simple game with dice:

http://en.wikipedia.org/wiki/Nontransitive_dice

So here is a new chess variant challenge: Chess with nontransitive armies

Design a chess variant with different armies such that, whatever army your opponent chooses first, you can choose another army having an advantage over your opponent's army. (To avoid the first move problem assume 2 games where either army moves first once)

Die Schwalbe is one of the leading problemist's journals (hey, they have journals allowing them to standardise on names!), therefore I consider their usage as well established. The Schwalbe reference gives no date when the name Ibis was assigned to (1,5) leaper, but here you can find a problem from 1999 using the Ibis:

www.softdecc.com/pdb/search.pdb?expression=CREATIONDATE%3E=20101230

(a quick google search with 'ibis chess problem' turned up this page as first hit).

To Christine: My first name is Jörg iwth an O-umlaut as second letter.

First, look at the recognised chess variants here:

http://www.chessvariants.org/rindex.html

Then, all national chess variants with a player community and some
tradition; i.e., Makruk, Sittuyin, Shatar (Mongolian Chess).

You may also look at contest-winning chess variants and contest finalists,
see here:
http://www.chessvariants.org/index/mainquery.php?type=Contest&orderby=LinkText&displayauthor=1&displayinventor=1&usethisheading=Contests

Unfortunately, no one seems to know the circular chess variant I have
sketched in the previous message of this thread. The board was similar to
this one:

http://www.chessvariants.org/index/msdisplay.php?itemid=MSknightsofthero

and the rules for going through the center were similar like in this
variant

http://www.chessvariants.org/shape.dir/globe.html

References

• Index of Ralph Betza's work on piece values
• A Pawn is as Strong as the Hand that Holds It Part I
• A Pawn is as Strong as the Hand that Holds It Part II
• A Pawn is as Strong as the Hand that Holds It Part III

Joe Joyce wrote in this thread:

> Jeremy Good expressed the wish for totally different armies, including 
> pawns, in our conversations on shatranj-style armies, where he wanted 
> for a long time to dump the knight from the shatranj CwDA. And someone
> has tried Chess with Different Pawns, but I cannot remember where.

In fact, there is an old experimental army for Chess with different armies, namely Ralph Betza's Jovian army, see

   http://www.chessvariants.org/d.betza/chessvar/cvda/wow-jup.html

This army may be an interesting opponent to the Spartans.

A very interesting note worth reading again from time to time.

What makes me wonder here are the remarks about old shatranj and that
K+4F (same colour) should win against K+R. I tried this by hand and found no way of confirming this claim. 5 Ferzes of the same colour are a different thing: A chain of Ferzes along the long diagonal can confine the rook in one diagonal half of the board (with the help of the King and noting that it is forbidden to give permanent check in Shatranj). Then, the net of Ferzes can be tightened until the rook is caught in a net and can be taken by the King.

There is still considerable counterplay by an active defending King, and exact play is needed.

I think, the mentioned scenarioes are worth a computer analysis.

--Jörg Knappen

A great idea and a great story together with the traces of playtesting. This makes an excellent new chessvariant.

Of course, a cooperative game can be regarded as a puzzle for one player only (fighting againts the dice), but my experience (from playing Rengo) shows that reading the partner's ideas adds another twist to a game.

In fact, I don't know immediately how to save this game. The problems are obvious: A blocked pawn threatened by a knight will bring the game to an early end, and white has the initiative which brings free moves for white while black is forced to answer all the threats by the white knight. The key is the forking power of the knight.

So there is much room for experimentation and playtesting. Ideas include:

* Make pawns insensitive to the knight

* Allow the capture of the knight if there are no forced moves left after
 the knight is removed

* Allow suicide of a stopped pawn

* Use multiple moves

* Tune down the knight to a Mao, a Moa, or a heavenly horse (vN)

Each idea and combinations of them need intensive playtesting, many of them probably must be discarded as the original sketch.

Is it allowed to move to a square threatened by a knight (of course you have to move away in one of the following moves, but maybe you can just oscillate between two threatened squares in order to avoid to be forced to move a blocked pawn?)

At last, a more precise description of the interaction between 'check' and forced moves is required. Does a piece give check and mate when it cannot move because other pieces have forced moves? What happens if your king is in check but other pieces have forced moves (you loose, you may move the king or remove the check in another way)?

Again, the rules need playtesting to see whether they result in an interesting game or not.

The Quintessence lacks the can-mate property on regular rectangular boards. The necessary piece of analysis goes as follows:

There is only one mating position with King and Quintessence vs. King, namely Black: King a1, White: King a3, Quintessence d2. Note that the position with a white King on b3 is a stalemate because the white King now blocks the check against the black King.

To force the the black King to a1, the position prior to the mate must look like Black King b1; White King a3, Quintessence controls squares c1 and c2. This can be done with Quintessence on b4 or d4. However, there is no quintessential move from b4 or d4 to d2. Therefore no mate can be forced.

I don't think that the rook 'reappeared' in Shogi, because it was never gone. Makruk, one antecessor in the genealogical line of Shogi, also has it.

What is striking is that no widely played chess variant has lost the knight completely despite the fact that it is found difficult by many beginners and even intermediate level players of chess. The knight certainly adds 'flavour' to chess and to any chess variant where it is in. 

The rook is another constant in chess, being there from the very beginning.

The european invention of the queen was precedented by the Japanese invention of the 'Free King' in large Shogi variants (like Chu shogi) by some centuries. What is more striking in this context is that the european obsession since Carrera, namely the Chancellor/Marshall and the Janus/Paladin pieces, does not occur in asian chess variants. This says---IMO---something about the quality of the pieces: The Queen/Free King is a perfect chess piece while the other two leave something open.

Back to Eurasian chess: It has a nice piece and rule mix and makes a great variant (learning from several other excellent games). For my taste, the Eurasian pawn is a bit too complicated and the rules concerning the pawn could be simplified. Promotion to captured pieces only has an old-fashioned look, at least.

Yes, it is. Currently I think that the Chancellor/Marshall is somewhat (maybe 0.5 pawn units) weaker than the Queen. The Fers move adds about 1.5 pawn units of strength, plus another 0.5 pawn units for curing the specific weakness of the Chancellor/Marshall, leaving the Archchanchellor at about 1.5 pawns stronger than the Queen. This difference should be noticeable during the game, but it is probably not enough to decide the pawnless endgame K+Archchancellor vs. K+Q in favour to the Archchancellor.

Note that this comparison is made for the 8x8 board. On larger boards the Queen gains strength compared to the Archchancellor because all her moves are long-range, but most of his moves are short-range.

The specific weakness of the Chancellor/Marshall is that the King can directly attack it. Thus adding a move-only Fers to it does not cure this weakness. But adding it may equalise the Chancellor with the Queen (on 8x8, of course).

There is a minor error about the WL (Wazir+Camel) compound: This piece cannot checkmate, because it does not control two orthogonally adjacent squares. Therefore it is not a major piece under the definition used by Fergus Duniho.

On the other hand, a Wazir and a Camel and a King (three pieces!) can mate a lone King.

So when I have to name a piece, I always look for an established name first. Often I find one I like and use it, sometimes I want to differ and choose another name.

I use the lists of Derzhanski, Jelliss, Grimbert and Poisson (in approximately this order of preference) along with notes I have taken from several chess variants. A new list is published by the problemist's journal Die Schwalbe, compiled by Gruber and can be found here: 

http://www.dieschwalbe.de/lexikon.htm

Truelove's list only gives names and games, unfortunately, and I have never done the task of annotating it with move patterns.

I do not use the names made up by Gilman for several reasons:

1. They are not based on exhaustive research, in particular they do not incorporate the four lists I have cited above. So he does not give the relevant names to known pieces and he sometimes assigns a new movement pattern to an already used name.

2. I need translatable names, i.e., names that work in languages other than english. Gilman's names rely heavily on untranslatable puns and anagrams.

3. For use in game notation, the initial letters should be different for the pieces drafted. Gilman's names tend to start in the same letter for many pieces.

I agree with him on the value of the Eohippos. It is one pawn above the knight, giving it a value of 4. When I designed the game, I assumed that the value of the Chancellor or Marshall were equal to the value of the Queen. Taking the horizontal movent away should cost about two pawns leaving it with a value of 7. However, I don't think any longer that the equality of Chancellor and Queen holds true. There is something in the Queen making her a perfect chess piece and giving her a higher value than the comparable chancellor. So a realistic judgement of the Fischer's value lies between 6 and 6.5 pawn units---not enough to win against a rook in an otherwise equal setting.

In the setting against the fabolous FIDEs, the levelling effect or elephantiasis correction strikes the queen. This might still even the game.

This game should be judged by its design criteria: Given a set of pieces (in hardware), create a solid and playable chess variant for them. It think it fits this purpose well, allthough I think keeping the original movements of the Omega chess pieces makes this a rather slow game, because the pieces are short-range on a 12x12 board.

On the Omega wizard: A pair of wizards and a king cannot mate a lone king (on any conventional rightangular board). The reason is that they must switch between odd and even ranks and files all the time. On the big board the wizard is clearly weaker than the bishop; on 8x8 it may be equal or slightly stronger because of its higher mobility and forking power.

Correction: NW is emperor, prince is the NF in problemist's tradition.

The Rhino is an interesting and inspiring piece and this article is well written.

The Teutonic Order did not have archchancellors. Their leaders are (they are still extant!) titled Hochmeister which is translated into english as grandmaster---not a name suited for a chess piece. The Holy Roman empire had three archchancellors, the most important of them was archchancellor of the germans. This post was filled by one of the 7 electors, the archbishop of Mainz. He was responsible (among other duties) for the organisation of the emperor's election and he substituted in emperor in the time of an interregnum (together with two vicars).

No, I have not been aware of the reference you mention, but now I know it and I shall build it into the notes when I revise the text.

I am looking for a certain circular chess variant I have seen in the 1980s.
It was published in a physics students' journal (if I remember correctly
from Bochum) in germany. The board consisted of a full circular disk. The
'squares' in the innermost disk were triangular sectors. It used standard
chess pieces.

The most striking feature I still remember is that the bishop changes the
square colour while crossing the center. Does anyone know this variant?

This is a forward reference to X Chess by Jeremy Gabriel Good

http://www.chessvariants.org/index/msdisplay.php?itemid=MPxchess

where many pieces of Seenschach are set on another innovative hourglass shaped chessboard.

In fact, mating with the mirror-rhino can be more complicated than mating with a simple FN, because
the additionaly attacked zebra square provides lot of stalemate traps. A key position in the mirror-rhino's mating manoeuvre is the following:

Black King on a1,a2,or b1; White king on c3, mirror-rhino on f6.  Black to move.

Only now the white king can complete the confinement of the black king

--- BTW, the rhino is not colourswitching at all, it is a slider continuing its way after WN squares to
C NN2 .... Adding the wazir to the Gnu essentially gives it the can-mate property.