How scary is this?

[jshriver]

Very neat. I went to lookup chu shogi and spent almost 2hours on wikipedia reading about shogi and the myriad of variants. Neat new world to explore.

[Kirill Kryukov]

Hi H.G.,

Both sets look great, however I vote for mnemonics - much easier to pick up and understand the game. Although pictograms are no doubt prettier.

Perhaps, show pictograms first to impress the user, then show a huge red message - "if you want to also understand what's going on, switch to mnominics".

Congrats and thanks for continuing to push computer shogi forward!

Kirill

[lucasart]

I am thinking of how to best add Chu Shogi to XBoard. Using the traditional Japanese kanji pieces does not seem a good way for enticing people to try it. That leaves a representation with pictograms, or a 'mnemonic' represetation that makes it obvious how the pieces move.

The advantage of the pictograms is that people will recognize the symbols for the pieces that participate in Chess, for which they know how they move. (And to a lesser extent the pieces of regular Shogi, if they tried that before with XBoard.) They would not have a clue how the unorthodox pieces move, however.

For the mnemonic representation it is pretty obvious how everything moves, once you realize how the system works (which should take about 30 seconds). There is absolutely nothing familiar about it, however: even Rooks, Bishops and King look strange.

What do you think?

I prefer the third secret option, that you didn't mention. It's the best of option 1 and 2:

use pictogram (pretty), but display the list of possible moves when selecting a piece (or hovering over a piece, whatever you prefer).

Of course, expert players know the moves, and don't need to be reminded, so they will want to be able to switch off the square highlight feature.

That's the way most normal chess GUIs do it: highlight possible moves to help beginners, but with the ability to switch off the option for more "advanced" players.

[hgm]

Sure, such a highlight feature is very helpful when playing variants. But that is already standard in XBoard, as the option 'Show Target Squares'. It applies the highlights to the board as soon as you grab or select the piece, and erases them when you release it:



There is a snag, however: it only shows the moves you can actually make. Not the 'virtual moves' the piece would be able to make on an empty board, and not even moves that protect your own pieces. I don't know if you meant it should also show these moves.

For learning how the piece moves, showing only its legal moves is less convenient, as you will almost always see only a subset of its capabilities. OTOH, some pieces might acquire moves by the presence of others (e.g. Grashopper). I could make a mode where the legal moves are shown with yellow and red dots, but virtual moves the piece would have on an empty board by open circles.

The latest development version of XBoard even allows the engine to apply the markers when XBoard doesn't know how the pieces move, so that the game has to be played with legality testing off.

[hgm]

Making use of the new GUI infra-structure that was necessary for Chu Shogi, it was trivial to also add a much more modest variant, which I called 'Mighty Lion (Chess)':



Like in Chu, the Lion moves as a King, but twice per turn. (Squares from which it can make a second step, but where it is important to 'make a stop on the way' because you want to (e.p.?) capture something there, are marked in cyan.)

Normally such variants (with only a single 'exo-piece') are not very interesting, because people immediately trade that piece, leaving them an ordinary Chess game. But the Chu-Shogi rule that forbids Lion trading is also active here: a Lion is not allowed to capture a (pseudo-legally) protected Lion. (From a distance; when in contact, the protection in meaningless, as the capturing Lion moves on for its second King step, taking you 'en passant'!) There even is a rule against indirect trading: If your Lion is captured by a non-Lion, it is not allowed to capture the opponent's Lion on the immediately following move. Those rules make it very difficult to get rid of the Lions.

I am still wondering what would be the best initial setup. Currently XBoard has this:



Would it be better to not replace the Knight, but to just add the Lion as a 17th piece? (Chess with only a single Knight just doesn't seem the same... )



What do you think?

[Daniel Shawul]

Thanks for this. I will be implementing this variant soon and comment on it by then. So far I did not have a multi-leg movers in alien variants. Inclusion of this will allow games like Arimaa, which I wanted to add some time ago but couldn't. This game is probably more closer to checkers anyway since only some pieces can multi-move in certain situations, so may be it is better to call Arimaa multi-turn. This will affect the design of whether to use a multi-move data structure like checkers' or just add support for the same side to move twice. Anyway I will try out and see.

[hgm]

I implemented it internally in the regular version of XBoard as a form of e.p. capture, indicated by a 'promo-suffix' ;EP_SQUARE, rather than as a multi-leg move. Click on cyan is used for indicating the 'e.p. square', not terminating the move.

[Daniel Shawul]

I implemented this variant 'mightylion' now. There may still be bugs to debug but the framework for lion kind of moves is in place. I think I can move variant 'jetan' to the alien version now. It has many pieces which move like the lion that I had trouble with before. The moves are represented with all the intermediate square as b1c2c3.

Code: Select all

            a b c d e f g h
            * * * * * * * * * * * * * * * * * * * * * * * * * * * *
            * * * * * * * * * * * * * * * * * * * * * * * * * * * *
            * * * * * * * * * * * * * * * * * * * * * * * * * * * *
            * * * * * * * * * * * * * * * * * * * * * * * * * * * *
          8 r l b q k b n r * * * * * * * * * * * * * * * * * * * * 8
          7 p p p p p p p p * * * * * * * * * * * * * * * * * * * * 7
          6 . . . . . . . . * * * * * * * * * * * * * * * * * * * * 6
          5 . . . . . . . . * * * * * * * * * * * * * * * * * * * * 5
          4 . . . . . . . . * * * * * * * * * * * * * * * * * * * * 4
          3 . . . . . . . . * * * * * * * * * * * * * * * * * * * * 3
          2 P P P P P P P P * * * * * * * * * * * * * * * * * * * * 2
          1 R L B Q K B N R * * * * * * * * * * * * * * * * * * * * 1
            * * * * * * * * * * * * * * * * * * * * * * * * * * * *
            * * * * * * * * * * * * * * * * * * * * * * * * * * * *
            * * * * * * * * * * * * * * * * * * * * * * * * * * * *
            * * * * * * * * * * * * * * * * * * * * * * * * * * * *
            a b c d e f g h

                [Material: 6168 6168 ]
rlbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RLBQKBNR w KQkq - 0 1

go
[search_time = 5557ms, max_time = 29250ms , moves_left 10, max_nodes 0]
3 20 0 133  g1f3  g8f6  f3d4  EBF=4.72
4 0 0 268  g1f3  g8f6  f3d4  f6d5  EBF=3.75
5 28 1 1309  g1f3  g8f6  d2d4  f6d5  c1d2  EBF=3.96
6 -18 1 4381  g1f3  g8f6  c2c4  d7d5  f3e5  c8e6  EBF=3.85
6 0 2 7179  e2e4  d7d5  d1e2  d5e4  e2e4  c8e6  EBF=4.20
7 0 3 15905  e2e4  a7a5  d1f3  b8a7b6  f1c4  g8f6  g1e2  EBF=3.81
8 -102 9 51157  e2e4  c7c6  d2d4  b8c7d6  f1d3  d6c7b6  c1f4  b6c5b4  EBF=3.73
8 -14 10 54645  g1f3  g8f6  c2c4  d7d5  f3e5  c8e6  d1a4  c7c6  EBF=3.76
9 -62 16 99726  g1f3  a7a5  g2g4  h7h5  g4h5  b8a7b6  h5h6  h8h6  f3d4  EBF=3.46
10 -90 32 222899  g1f3  a7a5  g2g4  h7h5  g4h5  b8a7b6  f3d4  g8f6  c2c4  h8h5  EBF=3.30
10 -60 62 427470  a2a3  c7c5  b1a2b3  b8c7b8  g1f3  g8f6  b3c4  e7e6  d2d4  d8e7  EBF=3.54
11 -78 69 481976  a2a3  c7c5  b1a2b3  b8c7b8  h2h4  g8f6  b3c4  e7e6  b2b4  d7d5  c4b5  EBF=3.17
12 -94 86 597344  a2a3  c7c5  b1a2b3  b8c7b8  h2h4  g8f6  g1f3  b8c7b8  b3c4  e7e6  d2d4  d8e7  EBF=2.93
13 -108 165 1171366  a2a3  c7c5  b1a2b3  b8c7b8  h2h4  g8f6  g1f3  b8c7b8  e2e3  b8c7b8  b3c4  f6e4  f1d3  EBF=2.83
13 -88 237 1753140  b2b4  g8f6  g1f3  a7a5  b4a5  a8a5  b1b2  b8a7a6  d2d4  a5h5  c1e3  a6b5c4  d1d2  EBF=2.93
14 -96 331 2540756  b2b4  g8f6  c2c3  e7e6  f2f4  f8d6  g1h3  a7a5  a2a3  a5b4  a3b4  a8a1  b1a1  b8a7b6  EBF=2.78
nodes = 4341896 <73 qnodes> time = 5567ms nps = 779934
splits = 0 badsplits = 0
move b2b4

setup rlbqkbnr/pppppppp/8/8/1P6/8/P1PPPPPP/RLBQKBNR b KQkq b3b4 0 1