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
Carrera's Chess. Large chess variant from 17th century Italy. (10x8, Cells: 80) [All Comments] [Add Comment or Rating]
Fergus Duniho wrote on 2016-04-09 UTC

Errors now corrected on this page include the following:

  • The spelling of the title of Carrera's book is not Il Gioco delgi Scacchi.
  • The pieces in the diagram were all placed in reverse, and the positions given for Kings, Queens, Centaurs, and Champions were all incorrect.
  • The earlier text says (or at least misleadingly suggests) that Ben Foster wrote Thomas Pruen's An Introduction to the History and Study of Chess..
  • The earlier text suggested that the game included en passant. But it does not.
  • The earlier text says that Murray mentions the game with a somewhat different (possibly wrong) startup position. Actually, Murray gets the startup position correct, but he mixes up his descriptions of the Champion and Centaur.

Sam Trenholme wrote on 2009-10-21 UTC
Somehow, I get the feeling that this posting is an extension to the discussion we had in this thread.

The best way to really study the opening of a given chess variant is to have a lot of games played with said variant, and analyze how the games went. Since no chess variant has traditionally had the popularity to have enough games played to say anything meaningful about the variant’s opening, H.G.Muller suggested having a chess engine play thousands of games with a given variant to get a sense of the opening.

After some 30,000 Schoolbook games, I got two significant pieces of data from all this simulation:

  • Some strong black replies to certain opening moves by white. 1. f4 c5 looks good for black, and 1. e4 d5 2. exd5 Nb6 looks to equalize. It was somewhat problematic to find a fully satisfying reply to 1. c4, and, indeed, 20,000 of the games were played to see how to minimize black’s problems after white makes this move; I finally settled on 1. c4 Mh6. One line of research I was not able to complete in the timeframe I allocated this project was to see if black had a reasonable reply to minimize his problems after 1. c4 e6 2. g4 (1. c4 e6 2. Bc2 f5 equalizes for black, but we can not depend on white cooperating so); I was not able to find a fully satisfying reply for black.
  • Over 100 mating positions, which can be downloaded and looked at in Winboard; go to to download them. I hope to one day make an inexpensive book with nothing but Schoolbook mating problems, a la Reinfeld’s classic 1001 Brilliant Ways to Checkmate (without the tactical errors the original book had)

The nice thing about the technological age and the ready availability of powerful home computers (did you know that an inexpensive netbook has as much computational power as a then start-of-the-art Cray XMP from 1984?) is that we can research information that previous generations could never dream of. I would like to thank H. G. Muller for making all of the software freely available so I could do a meaningful in-depth study of a chess variant. Indeed, I have coined 1. c4 in Schoolbook the “Muller attack” since it was his software that first showed me how powerful this line is for white.

George Duke wrote on 2009-10-20 UTC
Occasionally for indifferent and casual viewers the history needs reminding that Carrera's, according to Murray and Gollon and so on, was the inspiration for Bird's and Bird's for Capablanca, and none of them are worth duplicating by some new array anymore, because there is ''Capablanca Random Chess'' by CVPage erstwhile Scharnagl. I cannot think of a worse compound than Rook plus Knight, Carrera Champion. Champion was original in its time, sort of a mild satirical protest 100 years after OrthoChess today was clearly being established paramount. In those days they did not make many CVs. In the 1960s only 40 years ago they did not make many CVs. They had too much sense and made problems of ''fairy'' themes and OrthoChess problems. CVs are a fad since about 1990 or 1995 or arguably another 20 or 30 years before that with the likes of Betza at which time Boyer and Parton still lived. CVPage Trenholme's 10,000 or whatever games, if you can call them that, are probably ten times more than have been played in all 400 years or Carrera-Capablanca. Trenholme should be commended, as for a significant statistical tabulation, with virtually no game annotated or particularly understood. For depth there would have to be following move by move by parties interested. Though Carrera gets credit for Centaur and Champion, I would say what we now call ''CVs'' got started in the 13th century with Grande Acedrex sponsored by King Alphonso, almost another 400 years before Carrera. That is, notwithstanding earlier Persian, Arabian and Indian tweaks and four-way play, that hardly count because of approximation to Shatranj itself.

George Duke wrote on 2008-03-20 UTC
This is where early-17th-century RN (Champion) and BN (Centaur) actually come from, only one century after Mad Queen itself. Henry Bird in 1870's and Capablanca in 1920's have gotten more credit than they deserve. One of the two inventors of Seirawan Chess(2007, and lately commented) expresses surprise, saying approximately, ''apparently RN and BN have been around 400 years.'' They must have thought these compounds originated with Capablanca. It is really sorry commentary, speaking volumes on lax research into many CVs, not that one alone. Pietro Carrera made this breakthrough, and all since are tinkering around the edges. Betza, Karakas, and now the duo Seirawan-Harper would maintain the better embodiment for the two new strong pieces requires 8x8 not 8x10.

George Duke wrote on 2005-01-25 UTC
What was really written 'years ago': Carrera's Chess was written up
about 1617. This one's eight-piece mix is still being debated. I guess it
stood the test of time. 'Mad-Queen', now 'FIDE' or Orthodox, Chess
developed about 1480. Mediaeval ingenuity had more common sense than we do
today. Or, think about the heyday of Chess Variants and novelties (before
these poor degenerate times): Sam Loyd (1841-1911), Henry Ernest Dudeney
(1847-1930), Thomas Raynor Dawson(1889-1951). Use the hat-pin method to pick
any game developed after 1950. Would Loyd or Dudeney or Dawson devote more
than an hour or an evening, if that, to 99% of them? They did not need a method of
exhaustion to allocate their time, because they still had standards.
(Actually Dudeney is fond of looking at every possible permutation for his
number puzzles.)

Mark Thompson wrote on 2005-01-25 UTC
Touche! :-)

I wrote that years ago and have forgotten the wording enough that when I
reread it nowadays I keep thinking, criminy, what pompous a$$ wrote this

George Duke wrote on 2005-01-24 UTCExcellent ★★★★★
Mark Thompson in the classic article 'Defining the Abstract' uses the term 'their present perfection' referring to families of games including chess. Unfortunately CVP has never reprinted that article. I probably would have changed use of verb 'perfect' if Thompson's comment had not intervened. I mean 'to make better'. I think of the various Carrera embodiments as like rolling terrain with higher hills hard to discern with our tools which protrude, but with limited time we make the best estimate, and give supporting facts, including game scores.

David Paulowich wrote on 2005-01-14 UTC
'Carrera Random Chess' has become 'Pairwise Drop Chess'. Two games were played on the CV Game Courier in December 2004. You start a game by choosing one of five pairs: R+R, N+N, B+B, K+Q, Archbishop+Chancellor and placing the two pieces on your first rank. Bishops must be placed on opposite color squares. Each player copies the opponent's drops, placing pieces on the same files, and then chooses another pair. This process uses up the first three moves of the game.

Each king may 'free castle' once in the game with either the nearest rook on its left side or the nearest rook on its right side. This variant idea comes from the Kibitzer web article 'Bring Back Free Castling!' by Tim Harding. Note that free castling simply switches the King and Rook, if they are adjacent at the start of play. A Pawn promotes on the last rank to a Chancellor, Queen, or Archbishop of the same color. Nothing else.

Thanks to George Duke for reminding me of The Kibitzer #31, in his 2004-09-24 comment to Grotesque Chess.

Fergus Duniho wrote on 2004-10-08 UTC

In an earlier comment, I wrote:

According to Pritchard, the Champion (R+N) goes on the King's side and the Centaur (B+N) goes on the Queen's side. He also mentions that Murray mixes the names. Since this page describes the positions of Champion and Centaur as the reverse of what Pritchard says, and given what he says about Murray mixing the names, it looks like this page gets the setup of the game wrong. It would seem that it is my Capablanca variation, not Aberg's, that has the same setup as Carrera's Chess.

I just came back to this page, because I was reading Parton's description of this game in Dasapada. Parton gives the same description as Pritchard, but I have just noticed that the King and Queen are reversed in the diagram provided on this page. The White King is to the Queen's left instead of to her right. Thus, the descriptions given by Pritchard and Parton actually do fit the diagram given here. But I think confusion has arisen from the general understanding that Queen-side means White's left and King-side means White's right. In this game, it is the reverse. It turns out that the setup shown on this page is just the mirror image of the array used for Aberg's Capablanca variation, and they are mathematically identical.

David Paulowich wrote on 2004-09-21 UTC
My favorite opening setup in 'Carrera Random Chess' is: CRNBAKBNRQ. To be frank, I prefer it to any variant with rooks in the corners invented in the last four hundred years. Might put a strain on the royal marriage, but we do not have to worry about that kind of symbolism here. <p>How do we deal with a computer program that insists on a single piece drop by each player per move? Suppose, in a game that resulted in CRNBAKBNRQ, White started by placing the knights and Black replied by placing the rooks. Start recording the game as follows: 1. Nc1 Nc8 2. Nh1 Ri8 3. Ri1 Rb8 4. Rb1 Nh8. Black has simply delayed the required second knight drop until both rooks have been placed, thus creating the illusion of alternating single moves. Problem solved.

David Paulowich wrote on 2004-09-21 UTCExcellent ★★★★★
'Carrera Random Chess' - all your variants in one game!

The game starts with only the pawns on the board. Each Player divides the remaining pieces into five pairs: R+R, N+N, B+B, K+Q, and C+A (using the names Chancellor and Archbishop). White begins by choosing a pair and placing them somewhere on the first rank. Black copies this placement on the eighth rank and then places another pair of pieces somewhere on the eighth rank. White copies and then places another pair. Black copies and then places another pair. White copies and then places the final pair. Black copies on the eighth rank, resulting in an opening setup with every Black piece on the same file as the corresponding White piece. No castling allowed and no need to place the rooks on either side of the King. The only special rule is: Bishops must be placed on squares of opposite color. Naturally this also means you are not allowed to fill all the available squares of one color before the Bishops have been placed. The game should be played with the modern rules for pawn movement and capturing. Promotion to Archbishop, Chancellor or Queen works perfectly. Back on August 18, I discussed the myth of underpromotion in games with this piece set - see the Comments to 'Mainzer Schach'.

Fergus Duniho wrote on 2004-09-21 UTC
I was thinking of the same thing. It would be easier to program not only
for Game Courier but for Zillions too and maybe ChessV as well. I had
thought I had seen a new ZRF for Aberg's, and I was hoping to download it
and use it to test whether free castling gives White too great an
advantage, as some people have claimed it does. But I guess it doesn't
exist. I had never programmed free castling into my ZRF for Aberg's,
because I didn't know it was part of the game.

I was tending toward free castling, because there is already a name for
it. Perhaps I shall use a half-enhanced castling and call it flexible
castling. It would work like this. The King would move two or more spaces
toward the Rook, and the Rook would leap over the King to the immediately
adjacent space. By omitting the option of moving one space toward the Rook
and moving the Rook to the King's original space, a program doesn't have
to distinguish between a King's usual one-space move and a castling move,
and it will remain less confusing for players to enter notation into Game
Courier. It also helps to avoid making the castling move too powerful. I
think one of the complaints against free castling was that it would allow
the Rook to go directly to the King file. This may make castling too
powerful by increasing the opportunity of using castling to check the
enemy King. The flexible castling I'm suggesting won't allow the Rook to
move to the King file.

Here's another thing I like about flexible castling. It forces the player
to make a choice between one thing he wants and another. Is it more
important to move the King to a certain space or to move the Rook to a
certain space? He sometimes has to give a little to get what he wants. He
can't have his cake and eat it too. In contrast, free castling is
basically two moves in one. I think it's best to avoid that. So, I think
I will use flexible castling in Grotesque Chess. However, I'm now toying
with different ideas for the opening setup. I'm going to try out some
different setups with a Gothic Chess set and see which I like best.

Greg Strong wrote on 2004-09-21 UTC
Hmmm...  I do like the idea of the Enhanced Castling presented here, but it
might be a bit too powerful.  How about a half-enhanced castling, in which
the King can go as far as desired, but the Rook must then always be placed
just on the other side?  That should be a little easier to program, and I
think I like the rule better, too...  I don't like giving the King a free
choice of any square, *and* letting the Rook pick the open file, all on the
same move.

Fergus Duniho wrote on 2004-09-21 UTC

I'm going to abandon my old Capablanca variation, since it is nothing
more than the Carrera and Aberg setup with the same rules as Capablanca's
Chess. But I am working on a new one, which you might be happy to include
in ChessV when it's ready. I was inspired by some other comments you wrote
recently to create a game I am calling Grotesque Chess. For the present, I
will just leave this name as a riddle to be figured out.

I have decided on the setup for the game, but at present, I am still
debating over whether to include free castling, as Aberg's does. It seems
like a nifty idea, but it will make it much harder for me to implement the
rules with Game Courier. I'm going to sleep on it and see what I can come
up with tomorrow.

Greg Strong wrote on 2004-09-21 UTC
Ohhh... Thank you for pointing this out!  I will have to update ChessV to
correct this.  Actually, I'd like to add support for the Duniho variant,
too (now that I'm aware of its existance.)  Is this OK Fergus?  and what
are the additional rules?


LATER EDIT:  I had posted this before Fergus had posted his additional 
comment.  Now I'm really confused ... :)

Fergus Duniho wrote on 2004-09-21 UTC
According to Pritchard, the Champion (R+N) goes on the King's side and the Centaur (B+N) goes on the Queen's side. He also mentions that Murray mixes the names. Since this page describes the positions of Champion and Centaur as the reverse of what Pritchard says, and given what he says about Murray mixing the names, it looks like this page gets the setup of the game wrong. It would seem that it is my Capablanca variation, not Aberg's, that has the same setup as Carrera's Chess.

Bernhard U. Hermes wrote on 2002-10-26 UTC
Dear William,

I consider your effort to prepare Unicode support for Fairy Chess and
chess variant pieces very laudable. 
The naming of the pieces, however,seems a little problematic. Personally,
I prefer calling the Carrera Champion a Marshall and the Carrera Centaur a
Cardinal, others call them Chancellor and Archbishop. 
A Champion in Omega Chess is a different beast altogether.
Neutral Names, IMO, would be rook-knight and bishop-night for the naming
of the positions in the code, the writer then could in the text of course
use whatever name suits him.

Bernhard U. Hermes

William Overington wrote on 2002-06-20 UTCExcellent ★★★★★
Looking at which is
the Carrera's Chess page I notice that both Queens are on a square of their
own colour at the start of play.

Looking at the start up position in the Java applet of I notice that neither Queen
is on a square of her own colour at the start of play.

Does anyone by any chance have access to reproductions of any illustrations
from the 1617 book please so that the matter may be determined historically

I have recently been devising some code point and illustration pairings of
Unicode code points and chess symbols. 

I have now added some more code points so as to include the Champion and
the Centaur of Carrera's Chess.

I have published these in the hope that that might lead to some Unicode
compatible chess founts being produced, perhaps including the Carrera

The first document is at the following address.

The second document is at the following address.

Both documents are available from the index at the following address.

The index page of the webspace is as follows.

As I am interested in typography I am interested to know what designs the
1617 book suggested for the Champion and the Centaur (if indeed any).  Does
anyone know please?

William Overington

20 June 2002

William Overington wrote on 2002-06-19 UTCExcellent ★★★★★
Does anyone play Carrera's Chess please?

I like the idea of playing this game simply because it is so old.

I have seen the note in the older feedback which leads to mention of a 19th
Century translation of his book.  Does that book deal with ordinary chess
or with Carrera's Chess, or both, please?

Are there any documented games of Carrera's Chess please, in particular
from the 17th Century?

William Overington

19 June 2002

[email protected]

19 comments displayed

Later Reverse Order Earlier

Permalink to the exact comments currently displayed.