Check out Grant Acedrex, our featured variant for April, 2024.

[ 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 EarlierEarliest
Chess and a Half. Game with extra leapers.[All Comments] [Add Comment or Rating]
H. G. Muller wrote on Tue, Oct 11, 2022 12:54 PM UTC in reply to H. G. Muller from Sun May 16 2021 11:30 AM:

There is an error in the piece list below the diagram (which is not Diagram-generated, but hard-coded HTML): clicking the Queen there shows the move diagram of the Nightrider. This is because the onclick handler calls ShowMoves with the wrong argument (should have been 10 instead of 3).

Greg Strong wrote on Sun, May 16, 2021 03:45 PM UTC in reply to H. G. Muller from 11:30 AM:

This has been updated

H. G. Muller wrote on Sun, May 16, 2021 11:30 AM UTC:

Can an editor (or the author himself) change the fmW* move for Pawn in the diagram definition to fmnW* ? Then I can drop the automatic assignment of e.p. capturability (i.e. without the n) for W* even for the first-defined piece in the diagram script. Apart from Wildebeest Chess (which was posted by me so that I can fix it myself) this was the only diagram I could locate that used the W* move.

Aurelian Florea wrote on Fri, Oct 27, 2017 02:59 AM UTC:

That is cool!

H. G. Muller wrote on Thu, Oct 26, 2017 08:46 PM UTC:

I found the Joker Chess discussion in the programming section of TalkChess:

Aurelian Florea wrote on Thu, Oct 26, 2017 07:34 PM UTC:

So you have a retrograde analisys program for variants, neat, but thing of course goes up exponentially. But you would probably find out if in apothecary chess 1 KA vs K aanca is a checkmate, or vs K griffin. I believe for apothecary chess 2 you have also verified via grand chess the major piece combinations :)! I doubt the queen can "beat" any of the four such strong piece. Thanks for the info. GL.

By the way I have started an Elven chess game today, sounds good :)!

H. G. Muller wrote on Thu, Oct 26, 2017 07:25 PM UTC:

Well, I have not actually calculated that, so it is just based on intuition. But my feeling was that the Bishops win this by covering so much of the board that the Knight has nowhere to go anymore. Which would get more difficult on a larger board.

I do have a program that can generate a tablebase for 3 vs 1 man on boards of up to 14x14 (limited by the memory size of my biggest computer). But not for 3 vs 2, which would require vastly more memory. The more man, the faster the required memory grows with board size. The program uses 1 byte per position, so for 4 men a 12x12 board requires 144 to the power 4 = ~410MB, which fits easily. But for 5 me it would be again 144 times larger, i.e. 57 GB.

Aurelian Florea wrote on Thu, Oct 26, 2017 07:02 PM UTC:


Why is board size affecting KBB vs KN. It seems weird to me :)!

H. G. Muller wrote on Thu, Oct 26, 2017 05:55 PM UTC:

In orthodox Chess there are also end-games that require more than 500 reversible moves to force a winning conversion. So going to an 80-move rule there doesn't really solve anything. The end-game that would probably be affected most would be KBBKN, which is generally won with an 80-move rule, but mostly draw with the 50-move rule. But it is very questionable if it would be a won end-game at all on this larger board.

Aurelian Florea wrote on Thu, Oct 26, 2017 04:50 PM UTC:


Funny 2500s joke, I was hopping for my apothecaries to go into the 2300 (as they are smaller). Maybe mister Spock can challenge captain Kirk to one of the apothecaries chess :)!

V. Reinhart wrote on Thu, Oct 26, 2017 03:37 PM UTC:

Thanks Nicolino for correcting the minor typos (even removing "revised" from the title). It's awesome!!!

About the 50 vs 80 move rule:

Even in normal chess a case can be made that the 50 move rule should be extended. Some endgame mates require more than 50 moves, and this game has all the normal chess pieces plus more, and the board is bigger so there may be even longer checkmate sequences than in classical chess.

Of course humans can't calculate this type of thing in OTB play, but these endings could have a consequence for people playing by correspondence, or in games between engines. It's doubtful any of it is applicable now, but people might be playing Chess and Half in the year 2500. It's somewhat arbitrary, but if Nicolino recommends 80 I would leave it at 80.

The other comments seem to be valid concerns, but this is really getting into minutiae. Maybe something for Nicolino to work on, and finish before he dies, just in case the game is still being played in 2500.


H. G. Muller wrote on Thu, Oct 26, 2017 12:12 PM UTC:

Some minor points: your text says the Pawn has an initial move of 2-4 steps. But the Betza description uses W*, which means that it also has a 2 or 3-step push on 3rd rank, and a double push on 4th rank, while the Pawn can no longer be virgin in these locations. (And it is a bit pointless to put W* instead of plain W for the Super-Pawns, as these can only exist on the enemy half anyway, where W* also means single step. Likewise, it is pointless to give the normal Pawns e.p. capture, as they can never be in a location where they can do that without becoming Super-Pawns first.)

As to the 80-move rule: it talks about Pawn pushes. I suppose that does not include sideward moves of the Super-Pawns, as otherwise you could prolong the game indefinitely without any progress by just occasionally moving a Super-Pawn back and forth between two squares. But the text doesn't make that clear. It also seems to me that any promotion should be considered a sign of progress, and reset the counter.

I am not sure why you extended the 50-move rule to 80 moves. This game doesn't look slower to me than orthodox Chess. Did you have any particular end-game in mind that would take more than 50 (but less than 80) moves to force a win?

💡📝sirius628 wrote on Thu, Oct 26, 2017 02:36 AM UTC:

Sorry for the confusion. Only the radial leapers can multi-capture. Queen castling is removed now, because it was honestly just changing for the sake of changing, and didn't improve the game at all. You didn't mention this, but King-to-Kraken promotion is also gone now, as it forced players to keep playing in positions that would normally be insufficient-material, but would allow a win through King promotion. I have changed the rules to force promotion of non-pawns. And finally, "not mostly not" was a simple error, thank you for pointing it out.

Aurelian Florea wrote on Wed, Oct 25, 2017 08:58 PM UTC:

HG, I have searched for like 5 minutes the discussion you mentioned and failed, but if the need occurs I know where to find you. Moreover I remember now better, it was not the joker, not initially at least but my weird promotion rule, as fairy max, at least at that time could not neatly make different promotion zones. Thanks anyway it help a lot in measuring piece values though, although I'm left wondering besides the obvious problem of the value of the joker, how the introduction of joker influences the value of ohter pieces. Stronger pieces are weakened by the presence of the joker as they give chaces for an powerfull attack to an inherently weaker piece :). Have a good one!

Aurelian Florea wrote on Wed, Oct 25, 2017 08:49 PM UTC:

@HG, I did not knew that, I'll maybe reconsider :)!

H. G. Muller wrote on Wed, Oct 25, 2017 08:18 PM UTC:

Yes, the diagram has limitations. It was developed for the purpose of illustrating Betza moves. The Joker is a piece without a fixed move, mimicking moves of other pieces, which is outside the scope of Betza notation. I remember I actually did make a special version of Fairy-Max once that could handle a Joker. I think I simply hard-coded one of its 15 piece types to move as a Joker. Someone on the TalkChess forum was interested in that. It is probably still posted somewhere on my website (without link to it).

[EDIT] It seems the link is, and it was put there March 30, 2014. From the size it seems it is just a zipped (modified) Fairy-Max + game-definition file. I have no idea anymore what had to be done to let it play that variant in WinBoard. (Perhaps nothing other than select it.) It must have been discussed around that time in the TalkChess forum, but I have no time to search for that now.

Aurelian Florea wrote on Wed, Oct 25, 2017 07:22 PM UTC:

@HG & @V. Reinhardt

HG, You did tell me once (about a year ago) that the foul/joker/jester/imitator can't be represented, also Fairy-Max can't do it either (the reason I did not continue using Fairy-Max to continue developing my apothecary games). So I guess there are limitations.

H. G. Muller wrote on Wed, Oct 25, 2017 07:00 PM UTC:

Well, I am open to suggestions, but I could not really think of something that would still be palatable. For normal leapers and sliders it is easy enough, because intuition is shaped by experience with orthodox Chess, and you expect stretches of consecutive destination squares to be slides that can be blocked anywhere between the source and the destination. So showing those on an empty board is fine, and showing virtual captures in another color even though the board is empty solves the problem of divergent pieces. Chess players still tend to think of empty squares diagonally in front of Pawns as being attacked by that Pawn.

The problem is with lame leapers (the marker indicates the moves don't jump and can be blocked, but where?) and hoppers (locust capture really being a special case of hopping). There occupying a square does not only have effect on that square (as it would have for a potential capture to that square), but can enable a etire series of moves (e.g. capture to all squares on the ray behind it for the XQ Cannon). And in principle you can have hoppers that turn corners (see for instance the 'sandbox' diagram I posted with the article on Betza Notation), so that occupying a different square on the primary leg would enable an entirely different series of actual destinations. In general occupying a square can enable multiple destinations, and a destination can be enabled by occupying multiple squares. It is really a mess. In an interactive diagram the reader can always set up a position he wonders about, though, and see what moves the diagram allows there.

V. Reinhart wrote on Wed, Oct 25, 2017 06:43 PM UTC:

It does say that when looking at the board, but it goes away when looking at the move definition of pieces.

While on this topic, I wonder if there is a way to always uniquely specify a piece with a move diagram without requiring it to be interactive. Unless a piece is very complicated, I would think there is a way to do it, but haven't thought it through yet. If there is such a way, it would be my preference (why animate a specification if a static diagram would work)?

Anyway, thanks for your thoughts.

H. G. Muller wrote on Wed, Oct 25, 2017 06:24 PM UTC:

Well, above the diagram it says in fat letters against red background: "you can move the pieces". I really don't know how this could be made more clear than that. Not far from now such diagrams will also be able to make moves for itself, btw, similar to what is already possible in the Peace Chess article.

If the diagram does not show any capture symbols on jumped-over pieces, you can assume that they don't capture any. The Speedy Knight is a normal Nightrider. This is also what the move definition in the piece table says (NN). So I guess the accompanying text is misleading, a legacy from before the Knight could promote.

I agree with you that this seems a very interesting game.

@Nicolino: I see that you also allow the Queen to castle, but this is causing problems in the diagram. Not because it would not understand the ON symbol on a non-royal, but because the Queen has so many normal moves that have the same destination square as the castlings. For O1 castling with the King I now solved that by requiring the Rook being clicked as destination. But then there the destination square is defined as the adjacent square. The diagram currently handles capture of an own piece in a non-final leg of a multi-leg move ('da') as a castling attempt, but it would be very cumbersome to specify that on such a big board, and even more cumbersome to specify it for so many different castlings. I have no good ideas how this could be solved.

V. Reinhart wrote on Wed, Oct 25, 2017 05:53 PM UTC:

Oh, that helps. I didn't know that the interactive diagram is itself interactive (I thought it's interactive only because it appears when selecting on the piece name). There does still seem to be a minor discrepency in my view. The rules say "The new leapers/jumpers are capable of an optional multi-capture". I take this to include the speedy knight, and strictly speaking the knight is new too because it's not the same knight as in classical chess. But the diagram does not show capture symbols in the intermediate squares.

Again, these are minor points that should be cleared up before starting a game. As a whole - this is a very well-written game description and one of my favorite variants (although I have not played it yet).

H. G. Muller wrote on Wed, Oct 25, 2017 05:17 PM UTC:

Even if you don't understand the Betza move descriptions that are give in the piece table below the diagram, you can easily answer such questions for yourself by just trying it out in the diagram. This would allow you to conclude that a Knight isn't able to do anything against the pieces it jumps over.

V. Reinhart wrote on Wed, Oct 25, 2017 03:31 PM UTC:

Another comment, and again this is a stylistic idea, and not very important (but I'll mention anyway):

Would the page look better if "(revised)" was pulled out of the title, and instead (optionally) a very short "change history" added at the bottom?

V. Reinhart wrote on Wed, Oct 25, 2017 03:25 PM UTC:

A few questions/comments:

About the optional multi-capture: Rules say it is for new leapers/jumpers, so this includes knight, speedy knight, cat, star-cat, and eques rex, corrrect?

(I'm assuming the knight is regarded as new because of it's different allowed initial move, and promotion rule compared to classical chess).

The knight does say "may not double capture during the 2-step move". I take this to mean the initial 2-step move, and not its normal 1-step L-shaped move.

When a knight makes an L-shaped move, there are two ways to move over squares it jumps over. The optional capture allows TWO pieces on EITHER one of its two path to be captured, plus the destination square, correct? (so capturing three pieces possible in one move).

(or can the knight capture all 4 squares it jumps over, plus the destination, for 5 total captures)?

Last, I'm wondering if the very first sentence should be changed a little:

Instead of:
...but not mostly not because of it's L-shaped move...
better as(?):
...but mostly not because of it's L-shaped move...

Sorry to bring up minutiae, but I really like this game. If I didn't like it, I would't be asking these specific and detailed questions.:)

H. G. Muller wrote on Tue, Oct 24, 2017 08:14 PM UTC:

Option capture-by-overtake: The fact that capturing pieces passed over by a cat or star cat is optional will require specification of how these moves are annotated in game scores, and some very-customized UI for any computer interface. 

This is not that hard. The usual method for the locust capture is to put the 'to-click' on the victim. The UI would know that this is a locust victim, save the square coordinates as locust square rather than to-square, and continue waiting for the to-click. In the case that both a normal capture and a locust capture of the same victim is possible, the UI should assume locust capture when that victim is clicked, and then change that to normal capture (invalidating the locust square) when the victim is also clicked as to-square. (I.e. the normal capture is considered a two-leg move with a null move as second leg.) This is the most natural way, which people expect. ShogiVar works that way for Lion moves, and I now do that in WinBoard to (as well as in the interactive diagram).

Of course in the case of the Star Cat it is a bit more complex, as there can be up to two locust squares.

The way it works for applying move highlights is that for moves that can be made to completion, only the first capture on the way that was not yet entered by a mouse click gets highlighted (in red if it is the final destination, in cyan if it is a locust victim). Moves the initial part of which does not match the already entered squares are not generated at all. So for the Lion, as soon as you click it, the K squares occupied by opponents would get highlighted in cyan, NAD squares occupied by opponents in red, and empty KNAD squares in yellow. Clicking the red or yellow squares would define the move as a normal one. Clicking the cyan squares would define the first leg as the locust capture, and then highlight the surrounding squares, which are the possible end points of all moves with that first leg, in red/yellow, and the locust victim itself in red.

Most intuitive move notation is to explicitly mention any locust victim that is not implied by the move, in addition to the to-square. Like Lxe4-e3 or Lxe4xd5.

25 comments displayed

Later Reverse Order EarlierEarliest

Permalink to the exact comments currently displayed.