Comments/Ratings for a Single Item

Something is the matter with the interactive diagram on this page. The spaces are all too small, and the pieces, not being resized, are cropped at their edges.

Fergus Duniho wrote on 2022-12-13 UTC

On the vertical line of movement, it shows the King able to make normal moves like a Lance, and when the Pawn is moved to the last rank, the King gets an extra capturing move. But the King's ability to move forward is only a checking power against the other King and does not allow it to actually move.

I finally understand what you mean now! This was all a problem of obsolete versions of the script. When I nowadays create Interactive Diagrams, I am careful to always append a ?nocache=true to the URL of the betza.js script. But the Eurasian Chess ID was posted long ago in a comment, and I just moved it into the article without scrutinizing it. It did not have the nocache attribute in the script URL. So it was not using the version from the CVP website, but the one at CloudFlare. Initially I was looking at it from a computer that did have a more recent version of betza.js (without nocache attribute) in its browser cache, and there the problem you described did not occur. But now that I looked at the page from another computer, that did not have the script cached, and thus was using the CloudFlare version, it exhibited the problem. So the solution was just to append the nochache=true argument to the URL in the article, to make it use the latest version of the script (which was already in my browser cache; the latter considers the file with and without nocache attriobute as different files).

I did make some modifications in the betza.js schript, though:

For moves that can only be used to deliver check, but not to actually move or capture there, I let the move diagram now display the X marker, which was already in use for pseudo-legal moves of a royal piece that stumbled into check. The explanation of this marker symbol in the legend has been modified to reflect this meaning.

The move diagrams are now able to combine captures and non-captures to the same square that resulted from a different atom in the XBetza description (rather than using the marker symbol corresponding to the last-encountered atom). So when you write mQcK the adjacent squares get highlighted by the yellow circle, rather than the red diamond. (The distant moves of course all remain green squares, for 'non-capture only'.) This only works when the captures follow the non-captures in the XBetza move definition. A nice example is the Diagram on Divergent Chess, where Knights move as Maos, but capture as Moas.

Hello! I know this may be a strange comment, but I am working on a project involving your micro-max engine and have a few questions. Believe it or not, this site is the only place it seems that I am able to contact you directly. Maybe you have contact info on your website, but I could not find it.

Anyways, my question is simple: is it possible to play as the black pieces on this chess engine? I have tried changing the "initial piece setup" line of code and swapping the king and queen, but that doesn't seem to work. What would I have to do to get this to work? Thanks!

The Diagram allows different piece types to use the same image. Even the same ID. And it is smart enough that when this ID is in the promoChoice, it would not allow promotion to the type that would itself morph to something else on that square. In the Elk Chess Diagram I did use different images, though. (Flush your browser cache if you want to try this, as it was a recent patch.) As a player I like that better than having to pay attention to square shade.

Other examples of Diagrams for variants with position-dependent moving:

https://www.chessvariants.com/large.dir/contest/bigouter.html

https://www.chessvariants.com/usualeq.dir/amalgamated.html

https://www.chessvariants.com/rules/crouching-stepper-hidden-rider

I think that there is no escaping to accept that it will always be possible to construct pieces with such complex moves that a move diagram, no matter how advanced, will be more confusing than helpful.

I completely agree with this.  And I think that 99% of the time the digram tells the story.  And exceptions to this are usually called out prominently in the rules.  (Perhaps we should bold the statements that are especially important to supplementing the diagram.)

The Diagram treats cases llike the Elk as two pieces. There is an 'Elk (light)' and an 'Elk (dark)', which are treated as different piece types, which morph/promote to each other. So  you can summon a move diagram for each of those, without the need to move them around. Likewise, in Xiangqi you would have a 'Soldier (crossed)' and a 'Soldier (own half)'.

Yes, I knew this but called it out anyway because this is an important distinction.  ChessV also considers the Elks two types that change as they move, but I'm not sure how I'd handle the Xiangqi Soldier.  I think I'd view that more as how a pawn can move two spaces if on the starting rank.  But I think, in the context of this conversation, it's important to keep this dichotomy in mind.  We can go one of two ways and the choice we make dramatically affects how the diagram will operate.  In the case of Elk Chess, the author of the page did not describe them as two separate piece types.  If we diagram them as different types because it's easier, we have already deviated from the written description of the game.  (That said, for implementation purposes, that may well be the best way to go.  In the Elk Chess case, by making them different types, the icon of the piece changes when it moves so the player can see how it moves on the given square without having to think about it, so I think this was the correct decision.)

If one rally wants to have separate move diagrams for a 5th-rank King, the ID can provide that: you just have to define 'King (5th rank)' as a separate piece type from 'King', and make King oromote to 'King (5th rank)' when it reaches 5th rank, and the other way around when it reaches 4th rank. I don't think that would be particularly illuminating, though. Writing a sentence "the King is confined to the 3x3 Palace" right under the clickable list would be much clearer. I would prefer such a warning for every case where the normal move diagram would not tell the whole story.

I agree with this 100%. I've been working on this exact solution for ChessV -- identifying all the situations where the diagram doesn't tell the whole story and providing a simple explanation.

I think that there is no escaping to accept that it will always be possible to construct pieces with such complex moves that a move diagram, no matter how advanced, will be more confusing than helpful. In the variants I have made Interactive Diagrams for so far, I encountered pieces that could jump over any number of contiguous friendly pieces, that could capture linear groups of enemy pieces, that could slide only over occupied squares, and capture anything on the squares it passed over... This is why the primary function of the Diagram is to experiment with the piece on a fully populated board, surrounding it by friends and enemies in all kind of ways. If someone wants to study how the move pattern of a King is different on 5th rank, he can just drag the King to 5th rank, and then move it from there.

The move diagrams are already coupled to the mouse, for materializing an enemy Pawn by hovering, in order to reveal hopping and lameness. On an empty board a Grasshopper would not have any moves. The move diagram indicates where a mount would activate some moves. But you would never know whether these moves are like a Grasshoper or like a Korean Cannon without probing it. We cannot use the mouse for moving both the piece and its potential mount or victim. What if the piece is a Korean Cannon with position dependent moves? (Oops! Janggi actually has that!)

The Diagram treats cases like the Elk as two pieces. There is an 'Elk (light)' and an 'Elk (dark)', which are treated as different piece types, which morph/promote to each other. So  you can summon a move diagram for each of those, without the need to move them around. Likewise, in Xiangqi you would have a 'Soldier (crossed)' and a 'Soldier (own half)'.

For confinement there is another mechanism, though. In principle you could implement confinement as location-dependent moving: a Xiangqi General has different moves on d0, d1, d2, e2, f2, f1, f0 and e0/e1, none of which happens to step out of the Palace. So we could have 8 different piece types, all promoting into each other when they move. If you really did that, you would have separate move diagrams for each location. But it is simpler to consider the General as a Wazir that is confined to a 3x3 sub-board. Because the moves of a General on, say, d0 are not really different from a Wazir. They just have some moves clipped that would leave the sub-board. So "Wazir that cannot leave the Palace area" is a much clearer description.

If one rally wants to have separate move diagrams for a 5th-rank King, the ID can provide that: you just have to define 'King (5th rank)' as a separate piece type from 'King', and make King oromote to 'King (5th rank)' when it reaches 5th rank, and the other way around when it reaches 4th rank. I don't think that would be particularly illuminating, though. Writing a sentence "the King is confined to the 3x3 Palace" right under the clickable list would be much clearer. I would prefer such a warning for every case where the normal move diagram would not tell the whole story.

Ben wrote:

with the ability to move the Elk in the movement diagram the user can verify that it's working as expected / that the user has understood the text.

Yes, these are both good points.

1. This feature would help the creator of the diagram confirm that the correct Betza code is being used for the piece.
2. This feature would help a reader verify that he understands the rules.

Greg wrote:

They would need to know to do that

A line of text telling them that would help with that. For example, "Drag the piece around to see its legal moves from different spaces."

I think the Elk is a great example where being able to move the piece around would be really helpful. No, the user might not start with the diagram and quickly figure out the pattern, but a static movement diagram leaves the user with the impression that it's just a rook (assuming it started on light). Reading the rules is required in either case, but with the ability to move the Elk in the movement diagram the user can verify that it's working as expected / that the user has understood the text.

Now, the actual ID is sufficient for this, and the movement diagram mode doesn't need this additional feature. But it seems an incremental improvement if it's not difficult to implement and doesn't confuse the interface too much.

How about letting someone drag a piece around the board to see how it can move from different locations?

This does not seem like a good solution to me.  They would need to know to do that (most pieces in most games don't move in different ways from different spaces) and dragging the piece around willy-nilly to see what happens doesn't seem very efficient.  A note saying "the King cannot cross the river" would be the efficient way to communicate this.  It isn't really possible for move diagrams to show the nuances of special rules.  In Elk Chess, for example, the Elk moves like a Rook on light squares and a Knight on dark squares.  Expecting the user to drag the piece around until he notices the pattern would not be reasonable.  The diagrams give a quick overview of how the peices move, but to understand all the details, you still need to read the rules.

I suppose this is about the King's move diagram, when 'materializing' the Pawn by hovering over a square.

Yes, all my comments were about that diagram.

This is a general problem with pieces that have a location dependent move: these would need different move diagrams for each location.

How about letting someone drag a piece around the board to see how it can move from different locations?

It would be far more confusing to omit the King's three forward moves in that diagram; people would think the King could never move forward, no matter where it is.

My preference is for not being confusing at all. I don't want to have to choose which of two confusing options is less confusing.

I don't completely understand what situation / position you talk about. Is this all in the move diagrams, or is it when you are moving around pieces in the original board diagram? So the following commenst might not be to the point.

1) An X means the King is not allowed to go there, for reasons of check, but indicates that it would otherwise have been a legal destination. (So that indeed it could deliver check there, as for capturing a King you are allowed to move into check in orthodox rules.) When you say 'protected by a Pawn' you mean 'attacked by an enemy Pawn'? I don't understand what the other King has to do with it.

2) I suppose this is about the King's move diagram, when 'materializing' the Pawn by hovering over a square. This diagram shows the three forward slide as red diamonds, which is the marker symbol used for 'capture only'. That is indeed to generous, because it should really be 'King capture only'. (In Betza notation: k mode rather than c mode.) I only could imagine so many marker symbols, and I did not have a dedicated marker for this particular mode (which AFAIK only occurs in Xiangqi and Eurasian Chess). So I used the same marker as for normal captures.

When you actually put an enemy Pawn in the path by hovering, you will see that the Pawn cannot be captured. I don't understand what you mean by "extra capturing move". As far as I could determine the move diagram highlights all squares up to, but not including the Pawn with red diamonds. When selecting a King in the main board (rather than viewing it in the move Diagram) only the actual pseudo-legal moves will be indicated. So a square reachable by the forward slide will never get highlighted, unless the enemy King is on that square.

Come to think of it, though, the X would have been a more logical choice to indicate these forward slides, as moves that are forbidden because they stumble into check also effectively do have k mode, as they can still be used to capture a King.

What is really needed here is a better specification of what the symbol means; the standard legend (summoned by clicking on 'here' two times) does not mention that the red diamond has an ambiguous meaning in this case. This could be solved by writing the extra sentence (under the clickable list of pieces next to the board) "The forward distant moves in the King's move diagram can only capture a King."

3) This is a general problem with pieces that have a location dependent move: these would need different move diagrams for each location. The move diagrams always give the 'nominal move' (i.e. unrestricted by confinement), and the board of such a diagram should not be considered the real 'absolute' board, but rather a 'relative board' centered around the piece, wherever that may be. It would be far more confusing to omit the King's three forward moves in that diagram; people would think the King could never move forward, no matter where it is.

I'm not an expert on Betza notation, but from examining the interactive diagram for the King, it looks like some things are incorrect.

1. On the diagonals, it has an X on a space protected by a Pawn, but a King cannot avoid check from the opponent's King by moving to a protected space. So, the King still has the power to threaten check on that space against any King that might otherwise move there.
2. On the vertical line of movement, it shows the King able to make normal moves like a Lance, and when the Pawn is moved to the last rank, the King gets an extra capturing move. But the King's ability to move forward is only a checking power against the other King and does not allow it to actually move.
3. Finally, the King diagram shows the King at the edge of the River with the ability to cross the River, yet it is illegal for the King to cross the River.

I moved the Interactive Diagram into the article, after changing its appearance to the original style. I hid the ascii board for those that have JavaScript on; the Diagram had to be larger than the original static image in order to use the Eurasian .gif pieces.

About Vao: maybe Dawson gave that name because it was phonetically from the same family than Pao, and the V because this letter is made of diagonal strokes. Maybe it is not that, but it can be used as a mnemotecnic mean. Remark, it could have used Xao as well, that would have been looking more Chinese.

When I first made the Interactive diagram for this variant, I somehow overlooked the rules for King confinement and King facing. I now corrected that, so that the diagram indicates the proper King moves.

The confinement was easy to do: the diagram already supported a mechanism for this, by allowing the user to add a a small JavaScript function BadZone(file, rank, pieceType, color), which returns a non-zero value ('true') if a piece of the mentioned type and color is not allowed to move to the mentioned square. In absence of such a user-defined function, the diagram supposes that all pieces can go anywhere. This feature was needed to handle Xiangqi properly.

The non-facing rule was more troublesome, though. In the Xiangqi diagram I had implemented it by a trick, giving the Kings an extra move like a kind of lame 'Ski-Rook', which could not go to the first two squares, but could be blocked there, and relaxing the confinement rule for Kings to allow King capture anywhere. In combination with confinement to the Palace, this would prevent the King could capture anything other than a King with these move.

For Eurasian Chess this trick did not work, because the Kings can approach each other more closely than the number of ranks they are confined to. So any move allowing the Kings to capture each other could also be made entirely in the confinement zone, for capturing other pieces than Kings (which can go there). And for distant moves this should of course not be possible.

To solve this I made an ad-hoc extension of the Betza notation used by the interactive diagram: the modifier 'k' (which so far was not used) indicates the described move can only be used for capture of Kings (or, in general, royals). This makes it possible to describe the move of the King in Eurasian Chess as KkRkB, and that of Xiangqi as WkR.

I believe you're right Fergus, that pieces values are useful fictions (in chess-like games). Yet it's also true that equivalency formulae for tradeoffs of various piece [+ pawns(s), optionally] combinations don't seem to always hold in such games, as a general observation (it also seems to me such formulae are often implicitly based on presumed average piece values). This can depend on whether the position is an endgame, open or closed, and all sorts of other quirky characteristics. In the case of chess such characteristics have been extensively thought about, or have arisen time after time, but for a less explored game like Eurasian Chess I'd suppose there is still a lot to be learned.

Still, people (beginners at least) crave having a simple, single set of average case piece values to go by, for any such game. I based my values for Eurasian Chess heavily on the endgame, which admittedly may not be entirely correct, but it's a jumping off point if nothing else. For average middlegame (if not opening) piece values, if it turns out that there's an important distinction that should be made (as by a Chinese Chess book I've seen, regarding some of that game's piece values for the endgame vs. earlier phase), at the moment I have less of a clear idea about what the average values of some of the pieces should be in the earlier part(s) of a Eurasian Chess game, if the values are indeed significantly different than in the endgame.

I guess the reasoning is that a Knight plus any other piece is sufficient mating material, but a Bishop must be paired with the right piece to have sufficient mating material. While two Vaos on the same color would be equal in value to a Bishop when paired with a Bishop on the opposite color, this is an unnatural combination that would occur only through promotion, and a single Vao in combination with a Knight or Cannon is sufficient mating material. So, I would not value a Vao at exactly half of a Bishop. I would value it at more than half a Bishop. In the middle game, its value may be close to that of a Bishop, but in the endgame, its value will decrease. In general, piece values are useful fictions, and we should consider values of piece combinations more important than the values of individual pieces.

Kevin, I gather that you think the powers of the King in this game somehow reverse the values of the Knight and Bishop, but I'm not yet clear on your reasoning behind this.

Glenn, ask your question in a comment for an affected page and give some more detail there.

The layout of pages setup by using MS Word documents has changed and has resulted in something of a mess for my pages.  What has happened?.  Can this be put right?  I would very much appreciate an answer.

For Chess-playing ability and hard work, Bobby Fischer, Garry Kasparov, and Magnus Carlsen have my respect indeed, and I have no doubt their Chess ability would manifest itself in any sensible Chess-variant given the proper time to learn and study such Game.  As I have said before, let's drop the word "Variant" for any sensible Game e.g. Capablanca-chess, since these Games are still Chess. Yes, sensible is subjective, but top Chess-players would, I think, form a consensus of opinion on any particular Game or Set of Games.

As an aside I think that Bobby Fischer's ELO rating of 2785 achieved in July 1972 is, in reality, the highest achieved and that there has been, as former world champion Anatoly Karpov has stated, inflation of ELO ratings. 

I've edited my previous comment somewhat extensively.

For what it's worth, here is what I'm using:

I was mainly going by parts of the commentary on your Eurasian Chess page, Fergus, i.e. about the relative values of the various pieces (given at one point in decreasing order of value, though excluding any numerical values) . This was based on the various listed bare means of delivering basic Eurasian Chess checkmates. Correctly or not, I tried to take these listed bare means heavily into account myself when putting a N as slightly ahead a B (in numerical terms, in my case).

By way of comparison purposes, for my own 10x10 Sac Chess variant, I estimated a B at 3.5 and the N at 3, i.e. the reverse of what I've estimated their worth at in Eurasian Chess, which is in line with the premise your question (greater B mobility on a 10x10 board should favour it on average) as far as that game goes. However, Sac Chess is played with Ks that are like chess Ks (unlike in Eurasian Chess), so even for an average case Sac Chess endgame, I'd suppose the relative values of B and N would not be in any way affected. It may seem I'm going into contorsions a little to justify these values, but fwiw it seems to me one often has to use somewhat convoluted thinking to come up with estimates of the relative values of fairy chess pieces, unless one trusts fully in computer studies (in my case, I'd prefer to at least have a program used for a study that has a strong chess rating, if nothing else).

Note that even for Eurasian Chess I thought a B should be worth at least 3 pawns, but neither it nor a N should be worth 4 pawns. I wasn't going to have a Cannon worth 4.5 pawns (I think it was valued this in a Chinese Chess book I've seen, for what that's worth, though that book put R=9 to give context).

Note that a drawback of having V=2 is that three Vs=6 (greater than a R) yet any colour combination of these would never suffice to mate a lone K, not even by a helpmate.

I must point out that there's a slight discrepency between my Sac Chess value Q=10 and my Eurasian Chess value Q=9.5 (with the formula Q=R+B+P used for both games), given that the Eurasian Chess value for the Q really ought to be ten pawns too, in my view, but since in general reality cannot be perfect in every way, I wasn't going to quibble over 1/2 a P in value for such a high value piece as a Q (noting, though, that 2Rs=Q+P is a material equivalency formula that perhaps ought to still be valid in practice on average, in the event that that material balance happens in either game).

P.S.: Partly to avoid the slight discrepancy mentioned in the previous paragraph, here's my latest (possibly more accurate) set of relative values for the Eurasian Chess pieces, albeit with more (and uglier) fractional values included: P=1; V=1.75; B=3.5; N=3.75; C=3.75 (but just 2.75 in an endgame); R=5.5; Q=10 and K has fighting value=5. This set of values has the added point that 3Vs are now worth less than a R, though at the moment I'm feeling slightly uncomfortable with having such a relatively low value for a V, since for one thing it's possible 2V+K, or perhaps even V+K, can routinely hold a draw against B+P+K with ease (e.g. perhaps just by parking a V, if of the opposite colour of the B, on a square on its own side of the board in front of the P, and then only moving the defending K from then on). Also, is a N really worth more than 2V on average? Granted, this piece plus a V mate a lone K, unlike 3V...