Comments/Ratings for a Single Item

The name of Lion Chess belongs to a game created in the 1950s by J. Boyer.
It is described in Pritchard's Encyclopedia of Chess Variants. It replaces
the Queen, Rook, and Bishop with the Leo, Vao, and Pao (Cannon). Maybe you
could call your game Chu Shogi Lion Chess.

Seems pretty complex for a single piece.  The combination of double-move
AND leap, and the nuanced anti-trading restrictions, jumps out at me as
being an evolutionary design and not very elegant.

Kind of like FIDE pawns (with initial double-move, en passant,
promotion...).  Pawns have been heavily tweaked over the years.  That
probably makes FIDE a better game, but I would be reluctant to import all
of those quirks into a new game--especially one where pawns weren't a
major focus and/or where the target audience wasn't already familiar with
them.  (And in fact, many casual chess players don't even know about en passant.)

Thanks for your comments! Perhaps I should explain something about the
background.

The Lion as described is a key piece in all large Shogi variants, starting
with the 12x12 Chu Shogi. These games have been around since the 12th
century, so one can indeed assume they are highly evolved. The purpose of
the proposed variant was to introduce this piece to players of the FIDE
game. So the piece was a given. One could argue that the possibility to
jump is an unnecessary complication, and that one-or-two King moves per
turn might have been enough, as most distant squares can be reached
multi-path fashion, so the chances of actually being blocked are not that
large. Jumping over the Pawn wall is a quite common opening move in Chu
Shogi, however, so I decided to not simplify the piece.

Chu Shogi also has anti-Lion-trading rules, of which those I phrased here
are a highly simplified version. For 'substitution variants', which are
FIDE augmented with only a single unorthodox piece, there is always the risk
that the players will quickly trade this piece, after which the game is
spoiled as a variant. The more powerful the piece, the larger the risk, and
in the case of the Chu Lion, this risk is enormous. (One centralized Lion
attacks 24 squares, i.e 37.5% of the board, and if both players centralize
their Lion, the Lions will almost always attack each other.)

I don't think 'highly evolved' is a disadvantage per se. The evolution took place for a reason, and if in a new game the same reason would apply, it seems a regression to throw away the advanced features. E.g. it would never occur to me to drop the Pawn double-push and e.p. capture in a Chess-with-Different-Armies sub-variant. These are needed there just as badly as in orthodox Chess, for the same reason, and dropping them would make an awful game.


As to the name: indeed I noticed there was a reference to Lion Chess in the
index of this site. (No rule description, just an applet.) As it didn't
seem to have anything to do with Lions, (it involves just Cannons of various flavors), it
seemed kind of a misnomer, and I thought the proposed variant was more
deserving of this name. It is not unprecedented to have two games with the
same name (e.g. Dragon Chess). I don't like to use 'Chu Shogi' in the
name, to most orthodox Chess players the words Chu and Shogi would not mean
anything. If a name change is desirable, something like 'Mighty-Lion
Chess' would appeal to me more.

Lion Chess has a piece called Leo, which is the astrological sign of the lion. That's what it has to do with lions. Also, I have used the Leo piece in Caïssa Britannia with its name changed to Lion. Given the theme of the game, it was a given that it would have a lion piece in it. I considered the Chu Shogi Lion, the Murray Lion, and the Leo as candidates and chose the Leo. One of my reasons against using the Chu Shogi Lion was that it moves twice per turn, which would have made it harder to program. (Since then, I have added commands to GAME Code for handling double moves, which could be used to program it.) Also, it's just a complicated piece to remember how to use. It's one of the reasons I have stayed away from Chu Shogi.

As for the no-trading rule, see also Angel Chess (Ed has a java app for
it).

Thanks for the reference to Angel Chess. It is interesting that there the
concept 'protected' is also defined in terms of pseudo-legal moves (which
might put your King in check). My first idea was to use legal recapture as
the as the criterion, but that would force an engine to think one extra ply
ahead to determine legality after the recapture, and the effect would not
be that different. In fact pseudo-legal recapture is closer to Chu Shogi,
where it is perfectly legal (but of course losing) to leave your King in
check.

It is not clear to me if 'visually guarded' would include X-ray protectors.
With the rule I use, which measures this after the capture, they would
count. Otherwise the piece attacking its unprotected counterpart could be
attacked from behind on the same ray by a slider, after which trading would be possible.

Like in Chu Shogi, the rules I used also forbids indirect trading, so that
you cannot 'protect' your Lion by launching a counter attack on his Lion
(or soft-pin the piece he attacks you with on his Lion).


As to the name, it still seems 'Cannonized Chess' would have been a more
apt name for the other game, as there are 2 other exo-pieces apart from
Leo. But I am also happy with Mighty-Lion Chess, so I will use that name in WinBoard/Fairy-Max.

Note that the double move of the Lion only manifests itself when the first
leg was a capture, because every possible destination can also be used
through a direct leap. And in that case it has the character of e.p.
capture: a piece disappears on a square other than the to-square. Compared
to regular e.p. capture the to-square isn't necessarily empty (a
complication that also strikes Alice Chess), so that you can have two
capture victims.

Lion is a well known piece of fairy chess, and means an extended
grasshopper.
Piececlopedia:
"http://www.chessvariants.org/piececlopedia.dir/lion2.html".

> Lion is a well known piece of fairy chess, and means an extended
grasshopper.

Well, the piececlopedia page says there are more pieces known by this name,
and the Chu Shogi Lion (which is the piece I described) seems to predate
any other version by at least 600 years.

> It also seems logical that a variant whose distinctive piece has been known as a Lion for so long should start with Lion as that would be the obvious place to look for such a game in the index pages.

Well, an index in general should list items on key word, which is not necessarily the first letter (which might come from an article). Even if that is not policy here, it is questionable whether limitations of a particular website should have an influence on variant naming.

Of the possibilities that have 'Lion' as first word, my current favorite would be 'Lion-terror Chess'.

I've used a weak form of the lion in a number of games, mostly as the
leader piece in a series of games ranging from Chieftain Chess through my
current assault on the size of chess variants and the sensibilities of
chess variantists. It steps one or two squares, may change direction
between the first and second step, but may neither jump nor return to its
starting square. (Leaders aren't allowed to run around in circles!) But
the first game I used it in was Lemurian Shatranj, as the queen analog. It
was quite a strong piece there, and it was matched up against some rather
strong pieces on a small (8x8) board. It was the other strong pieces which
balanced it out.

When I look at Chu Shogi, I see both a 12x12 board and 8 pieces/side of a
rook or greater in value, most of them sliders. Both these serve to dilute
the power of the lion. Both are lost in bringing the lion to the queen's
place in FIDE. It seems to me you are metaphorically bringing a ninja piece
into a game of epee and main gauche... I have a terrible urge to suggest
taking a leaf from Noble Wing chess and adding a non-capturing ferz move to
the queen's rook and king's knight, and a non-capturing wazir move to the
king's bishop as an alternate variant that might reverse a little of the
gain the lion has made in the transition to a smaller pond. Some may find
it interesting.

True, the immense power of the Lion could be a concern. In Chu it is worth
over 1.5 Queen, which would put it around 15 pawn units. (Even without the
double-capture power, a WFDAN leaper would be worth about 11 pawns in a
FIDE context.) Perhaps it is indeed a bad idea to use the Lion as Queen
replacement, because that leaves only a Rook as second-strongest piece, and
thus virtually no possibility to trade the Lion for other pieces.

With Queens, there is at least the possibility to trade L for Q+R. One
possibility would be to replace a Knight with a Lion, rather than a Queen
(e.g. on b1/b8). The downside is that this would make an extremely large
'power density' on a relatively small board. (Not worse than Seirawan
Chess, however.)

I don't want to introduce any other unorthodox pieces, and certainly not
upgrade them, as this would create even more power density. All FIDE pieces
(except K and P) can already attack a Lion from a safe distance (i.e.
without exposing themselves to hit-and-run captures).

Perhaps a better design would be to switch to a 9x8 board, with a setup
similar to Angel Chess, replacing the Angel (worth about 13 pawns) by a
Lion. Non-standard boards tend to strongly decrease the willingness of
people to play a variant, however. That would also be an argument against
using 10x8, replacing C and A from Capablanca Chess by L and an extra N.
(To keep the power-density low: L+N ~ C+A, value-wise.)

You might replace a rook with the lion, instead. This would keep more power
on the board, but would mess somewhat with castling. 

One thing I neglected to mention in my last comment is that the weaker lion I use captures only once/turn, and stops on that square.

Joe Joyce's piece is almost the same as the Warlord in my game For the
Crown (the Warlord is allowed to return to its own square, though I don't
think I've ever seen someone exercise that option).  I priced it slightly
lower than the Queen, but that's only because I believe long-range moves
are more important in For the Crown than in FIDE; I will still often
promote to Warlord over Queen (since the promoted piece is already on the
enemy back rank).

I've found it's important to plan your defense against a charging Warlord
at least a turn or two before it arrives.  (Though in For the Crown, that
usually means "drop some pawns in front of your King", which doesn't
translate to FIDE.)

Or what if you replaced a rook with Jeremy's weak version, and allowed
castling with it, to start? Then you could upgrade to the full lion later,
after a player has had experience with the weaker version. Maybe use 2,
replacing both rooks and leaving the queen. If you did that, you could
dispense with those I can't capture your lion if you capture my lion with
your lion... rules.

Well, the fact that the Lion can move on after it captures is all the fun.
Otherwise it would be a pretty dull piece. I'd rather drop the jumps than
hit-and-run captures. But I don't expect that to have much effect on the
value, so it would not be worth it.

It is true that replacing a Rook in stead of a Knight would weaken the army
by 2 Pawns. As you say it causes some castling trouble. But what might be
worse is that it makes it more difficult to trade a Lion for other
material, as the only reasonable path for that would be L vs Q+R, and
halving the number of Rooks halves the number of opportunities you will get
to do that. OTOH, removing a Knight leaves plenty of possibilities to trade
Rooks for weaker material (e.g. R vs 2 minors), as there are still 3 minors
left.

So I expect the game to work better it the Lion replaces the Knight rather
than the Rook.

If your main goal is just to incorporate "hit-and-run" attacks, you could
put them on a less powerful piece.  For the Crown also has a piece called
the Charger that moves as a R3 but can "overrun" an enemy piece and
continue moving up to its maximum range (and even capture multiple pieces
in a row).  I've seen a couple of other variants with similar pieces.

I estimated its value in a FIDE context as ~6 pawns, but that hasn't been
tested and might be far off.  If you happened to produce any new
information on its value, I'd be interested.

(Trivia:  The last expansion for For the Crown originally had another unit
called the Behemoth with the same "overrun" mechanic, but on a Q3 instead
of a R3.  It got scrapped because it made it too easy to obtain a perpetual
check.)

The main goal was to introduce people to the Chu Shogi Lion in a context
that is as familiar as possible to Chess players.

I think you are most likely right then, that introducing it on a knight
square is probably the best option. Imagine you'd replace the king's
knight.

I am not sure about that. It is true that the Queen is also on the left
half of the board. (But only barely so.) Putting the Lion on the King side,
however, would force the players to develop it before they can castle in
that direction. (And it is often much more convenient to castle there, as
on the Queen side you would need an early Queen move, and your King ends up
in a worse place too. But developing a Knight in front of your Pawns is
comparatively safe, as it is well protected, and only has to fear Pawns.
Putting a Lion there makes it a sitting duck for all kind of slider
attacks. The only way to prevent that would be to develop the Lion to e2,
which makes for much less rich opening theory. OTOH, a Lion is very agile,
and perhaps developing it to f3 or g3, and moving from there to a more
central position  might not be very inconvenient. So I am still in doubt.

Couldn't you set up a round robin on your chess engines and play each
different location against the others? I know it'd be something of a pain,
but it might tell you a lot about the various positions for the lion.

Eventually I will be doing that. For now the engines I normally use for
this (like Fairy-Max) cannot handle pieces with double-moves at all,
however. The only engine I have that can handle it is my Chu Shogi engine
HaChu. But that cannot do castlings and e.p. captures, and in addition it
is excessively weak, as I just started writing it, and it is almost without
anything yet, just the most basic fixed-depth search.