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
Avatar Chess. Game with avatars that can assume any piece of chess, depending on the fields of the board. (8x8, Cells: 64) [All Comments] [Add Comment or Rating]
H. G. Muller wrote on 2022-10-22 UTC
satellite=avatar ranks=8 files=8 maxPromote=0 promoZone=1 promoChoice=Q graphicsDir=/membergraphics/MSteammatechess/ squareSize=50 firstRank=1 fileOffset=0 darkShade= lightShade= rimColor=#FFFFFE coordColor=#000000 background='/membergraphics/MSinteractive-diagrams/diagram2.png' whitePrefix=w blackPrefix=b graphicsType=png symmetry=rotate useMarkers=1 borders=0 pawn::ifmnDfmWfceF::a2-h2 knight:N:::b1,g1 bishop::::c1,f1 rook::::a1,h1 queen::::d1 man::K:commoner: king::KisO2::e1

Avatar Chess

Despite the better evaluation the Diagram now makes on morphing pieces, it still cannot do any better than this: No extra avatars (drops are not implemented in the AI), and it cannot 'promote' the Rook on a castling move.

*** Nice graphical improvement ****

to the Diagram, though: I introduced a parameter background='...' , through which you can specify an image file that will be used as background to the board table. (You must make sure it has exactly the same size as the board, including the coordinate margins, though.) To make it visible you should define lightShade and darkShade as empty strings.

H. G. Muller wrote on 2022-10-12 UTC

How about this variant?

Greg Strong wrote on 2022-10-10 UTC

I just stumbled on a similar idea in Pritchard's Encyclopedia of Chess Variants. "Cataclysmo" by Bruce Trone, 1991. A piece's movements depends on the square it occupies. On even numbered ranks, a piece moves as a pawn. On odd numbered ranks, it moves as the standard piece that normally starts on that file (rooks on a and h, knights on b and g, etc.)

No mention of any special treatment for the king. Also no clarification of whether "even" and "odd" is in an absolute sense, or relative to the player but I think it must be relative -- that way, each piece moves from its initial position as it would in Chess. And nothing about castling, extended pawn move, or en passant.

Greg Strong wrote on 2022-10-05 UTC

For those interested in testing out this unusual game, I have a new build of ChessV that plays it. Everything is supported except the rule where pawn moves to the last rank give you another avatar to drop.

NOTE: You may need to copy that link and paste it into your browser. Since my site is not HTTPS, some browsers will not follow the link.

Extract the zip somewhere, run ChessV.exe, and double-click on Avatar Chess from the games index.

Kevin Pacey wrote on 2022-09-24 UTC

Testing the game first with actual play, say using a preset on this website's Game Courier, would give players and viewers much clearer ideas about how any complex strategies might unfold in typical games.

Because the game is not played on a board that uses plain graphics (like a chess board), you (or someone else) may need the help of a CVP site editor or the webmaster to get the graphics of the board onto this website, for subsequent use on Game Courier when you (or someone else) will go about making a preset.

Gerd Degens wrote on 2022-09-24 UTC

I would like to see similar discussions around 'AC'.

Gerd Degens wrote on 2022-09-14 UTC

Thanks Ben for your posting.
But is there still interest for that idea? And if so, in which 
direction does it go?

Ben Reiniger wrote on 2022-09-05 UTC

OK, with that I think the 2-avatar closing files mate works. Without loss of generality (now that everything is symmetric), say defending avatarius (I'm just going to call it "king" for the rest of this) is somewhere on files a-d. We should have plenty of time to set up two avatars on, say, h4 and g5, then can walk them to f6 and g5. The defending king is somewhere a-e.

Next we want to pivot from g5 to e8; we can go through h5, and we just need to be sure to defend e8. The f6 avatar covers most of the ground, leaving only d7, and zugzwang makes it easy.

Next to pivot from f6 to d1. The steps f5,h5 will do for travel, and again we just need to ensure safety of the landing square. This time it looks like we need to use the friendly king, using zugzwang to ensure d2 is not occupied so that our king can move to e1.

The hardest pivot might be from e8 to c3. I think the friendly king needs to get into the a-c files now in order to protect c3 while not blocking the d-file queen guard. I think this is possible through e1-d2-c2, keeping the e8 avatar to prevent escape while our king is on d2, using a bit of zugzwang to force our king's way in?

After that it's easier, with c3 protecting b4 for the next pivot (through e.g. d5, or f3/g4-f4), and then a final c3-a1 or c3-h8-a8 (depending on which half of the a file the enemy king is confined to).

Gerd Degens wrote on 2022-09-04 UTC

I did indeed overlook the lack of rotational symmetry of the board. I changed that - as well as the max. number (5) of avatars that can be added. Thanks for the input.

H. G. Muller wrote on 2022-09-03 UTC

Let's take the starting position with the black royal on e8 and white on e6 and b4. The white move a1 to a8 or h1 to h8 checkmates the royal. Or am I wrong? If not, it took 3 avatars to checkmate the royal.

You don't get the point. There is a difference between forcing mate, and showing that a mate position exists that can be reached if the opponent is so helpful to run into the knife. With a pair of knights there also are many reachable mate positions. (E.g. white: Kb3, Nd2, Ne3; black Ka1; Ne3-c2#) Nevertheless KNNK is a dead draw.

It is quite possible that avatarius plus two avatars can force checkmate. But to show that, one must start from a position like one with white on a1, b1 and b2 and black on e4.

Ben Reiniger wrote on 2022-09-03 UTC

It'd almost be straightforward to force mate with two avatars (plus the friendly avatarius) occupying rook or queen cells in consecutive files, except for forcing the king away from the e file.

The two sides aren't symmetric: an avatar on the black queen-start cell d8 can move and remain queen-powered at f6 or g5, but one on the white queen-start cell d1 cannot. I wonder if exchanging the king and queen starting cells on the 8th rank (completing rotational symmetry for the whole board) would be better for both points.

Gerd Degens wrote on 2022-09-02 UTC

I must have answered too quickly and was apparently asleep. Only the moves from a1 to a8 or from h1 to h8 are relevant. Please forget everything else.

Let's take the starting position with the black royal on e8 and white on e6 and b4. The white move a1 to a8 or h1 to h8 checkmates the royal. Or am I wrong? If not, it took 3 avatars to checkmate the royal.

"But why would the King ever move to e8 on the preceding halfmove?" I can't answer and only refer to the gameplay, which causes the initial situation I described, namely black royal on e8 and white on e6 and b4. Then follows the white move a1 to a8 or h1 to h8.

H. G. Muller wrote on 2022-09-02 UTC

The piece indicator is always redundant when you give the origin square, in variants where there can only be a single piece on a square.

I don't understand wfhat you mean by the mating example. What is a1-a8? A move? Is g5-g8 supposed to be an alternative for that? From g8 the avatar would move like a Knight, right? So that would not even be check, and black would just move e8-d8. On a8 the avatar would indeed remain a Rook. But why would the King ever move to e8 on the preceding halfmove? If it was on d8 it could have moved to c8 or c7, if it was on f8 to g8 or g7. This seems more like a helpmate.

Gerd Degens wrote on 2022-09-02 UTC

Does a royal moving as orthodox king also trigger a promotion when it moves to 8th rank?

No, this is not intended.

Securing a 'rapid-promotion factory' (e.g. a few white avatars in the trapezoidal region d7-e7-f8-c8 shuttling between 7th and 8th rank) might become the main strategic goal in the game.

The function of the opponent's pawn row is precisely to increase the dynamics of the game. The assumption that this results in a 'rapid-promotion factory' is only valid ceteris paribus, i.e. when the rest of the game is left aside. And even if it does, it is supposed to 'speed up' the game and make it more interesting.

I had already thought about 'forbidding' the return from the opponent's base line to the pawn row. But that would complicate the game and would certainly be inconvenient for programming - if it should come to that.

But I am quite with you, H.G., that the number of avatars that can be won during the course of the game must be limited. Here I will change the description and will include a limit of 5 avatars. This should be enough for the intended effect. many avatars would be needed to force checkmate on a bare royal (moving as orthodox King)?

Black Royal on e8, white on e6, b4, a1-a8 (or h1-h8/b6-b8/b3-b8/c5-c8/c4-c8/c3-c8/f6-f8/f5-f8/f4-f8/g6-g8/g5-g8/g3-g8). It takes 3 avatars to checkmate the royal in this example.

(Quick note: the chess notation changes in Avatar Chess; the initial letters of the pieces are no longer necessary).

H. G. Muller wrote on 2022-09-02 UTC

I am still a bit concerned whether this game is winnable. If the number of new avatars one can obtain through promotion is limited, and the limit has already been reached, how many avatars would be needed to force checkmate on a bare royal (moving as orthodox King)? Has this been investigated at all? It seems far easier to promote in this variant than in variants where pawns have to cross the entire board (and pass each other) to do it. So if there is any limitation on the number of extra avatars it is likely to be used up in an early stage of the game, and end-games would be conducted in the 'no promotions' regime. If there is no limitation, the game might never reach the end-game stage, as people would replenish their armies all the time. Securing a 'rapid-promotion factory' (e.g. a few white avatars in the trapezoidal region d7-e7-f8-c8 shuttling between 7th and 8th rank) might become the main strategic goal in the game.

Does a royal moving as orthodox king also trigger a promotion when it moves to 8th rank?

Gerd Degens wrote on 2022-09-01 UTC

There is nothing to add to this!

Aurelian Florea wrote on 2022-08-30 UTC

I think Smess, which has a similar theme is a better designed game!

Christine Bagley-Jones wrote on 2022-08-30 UTCExcellent ★★★★★

Hi Gerd, hope your fine and well. I'm sure you haven't upset or hurt anyone!! I understand what you said, and it is extremely interesting idea indeed how you describe your game about the board etc.

I don't want to go on about the pawns, but still, I have to say, the game would maintain your 'theme' even if the pawns remained pawns. However, the game as you have it must have a unique feel and play to it!

Would be fun to see the game in action.

Gerd Degens wrote on 2022-08-29 UTC

Thank you Christine and thank you Greg,

let me be a little provocative: In normal chess, the board is - please don't be bad - a hole nut. The board is where the game happens, but where the board has almost no influence on what happens in the game.

In Avatar Chess, the pieces are only substitutes and the board determines the rules. This means at the same time that the structures of Avatar Chess are no longer comparable with normal chess.

Especially the mechanism of the pawns, which you mentioned, is cancelled. Pawns have a barrier function in normal chess, in Avatar Chess pawns acquire pieces to maintain the playing power.

Perhaps you can say it like that: Normal Chess is a top down game, where the pieces dominate and where the board plays no role.

Avatar Chess is the opposite, namely a bottem up game with rules from the underground (a little theatricality must be). Maybe a little less predictable, but a little more exciting.

From Magnus Carlsen's environment I got the indication that the game could be chaotic. Maybe that's true, maybe not.

But one thing has to be mentioned: In normal chess there is a tendency to protect strong pieces (e.g. queen). This tendency does not exist in Avatar Chess.

An avatar on a queen square that is captured can easily be replaced by another avatar on the same or similar square. The sturcture of Avatar Chess is quite different.

I hope I haven't upset or hurt anyone.

Christine Bagley-Jones wrote on 2022-08-29 UTC

It is an interesting idea, though in it's form right now, I think Greg is right, would be hard to see many moves ahead, and also his comment about the pawns.

Pawns make a barrier, they hold the game together, stabilize it. The game would have to be play tested. I feel the game could be good without the pawns being able to change. Pawns should always be pawns, like the King always a King.

But hey, it's your game and I could be completely wrong!! Interesting idea you have, well done.

Greg Strong wrote on 2022-08-28 UTC

With a game as unusual as this, it's difficult to say without playing it, which I'm not able to do at present. My guess is that it would be difficult to see more than a move or two ahead with the pieces changing types on almost every move. Also, in chess the slow-moving interconnected pawn chains create a kind of terrain the pieces must move around. In this game, as soon as a pawn moves it's no longer a pawn, so the center of the board will be a free-for-all of powerful pieces. Still, it might be fun.

Gerd Degens wrote on 2022-08-27 UTC

Does no one have an idea? I do not want to disturb, but why does no one comment? Lack of interest or there are neither positive nor negative comments? Am open for any criticism. My last comment on the subject for now.

Gerd Degens wrote on 2022-08-23 UTC

An expected indication, H.G. The assumption that an avatarius can hardly be checkmated as a normal avatar, I had already suspected.

So let's exclude this possibility and assume that the avatarius has only move possibilities like a king in normal chess. Then we have normal chess enriched with changing move possibilities.

Theoretically and exclusively in the short term, Avatar Chess can have 6 queens, 12 rooks, 12 knights and 12 bishops in play. 

Is that too special or is that a crass challenge?


H. G. Muller wrote on 2022-08-23 UTC

I had not scrutinized this variant much, because I dislike games were pieces change their move too much. But now that I give it a closer look, I start to wonder whether it is actually winnable. The version where the king moves as a normal avatar seems completely hopeless in this respect; you just keep shuttling your royal avatar between high-mobility squares. How many avatars should the opponent be ahead in order to checkmate a bare royal?

Ben Reiniger wrote on 2022-07-16 UTC


I'm reminded of the #Powerup-Zone tag, but this doesn't fit there. There's also Lumberjack which is sort of a simpler version of this, and of course Smess.

25 comments displayed

Later Reverse Order Earlier

Permalink to the exact comments currently displayed.