Comments/Ratings for a Single Item

Since these variants have so many piece types, I tried to pick mainly pieces that a player might already be familiar with. This is why I preferred WAD and FC over FAH and WC, and avoided bent sliders that turn after two steps. I had my doubts on FC, because I already have a pure C. But because it is such a well-known piece I considered it admissble. Other compounds of C plus a 4-target atom (such as WC) would suffer from the same lack of distinction with the pure C, and would not have the benefit of being somewhat familiar.

I could not afford eliminating the Camel itself, because there aren't too many totally symmetric 8-target leapers, and I needed everyone I could get to have enough knight-class pieces. But when I want to avoid including pieces that differ from each other by just having 4 extra moves, that leaves almost no 12-target leapers.

Thinking about the FAH (which I call Siege Elephant in my games) and the champion there is also the WL (That I call Mamluk) and the FL (the wizard). So you may have a variation on the game where the WAD and FL are replaced by FAH and WL.

See now why I prefer to make variations on the same theme? Or it could be just me being weird, I guess.

Actually I do not now how this goes with your other pieces. But if you replace the FA with WD it is probably the same.

Also, the craziness could go on, you know?

And another thing. Have you considered bent riders that bend after more then 1 move: the R2 then bishop or the B2 then rook? They probably work well in such variants.

Having most pieces promote makes the game more shogi-like than chess-like, and I would like to avoid that. For this reason I also specified promotion on last rank, rather than in the camp.

If mating potential would be a problem, I'd rather adopt the rule that a bare King loses. But I am not sure it would be a problem. Eliminating all Pawns when you are behind is easier said than done. When behind, you will not be in a position to gain them. So you would have to sacrifice them away, where each Pawn would cost you yet another piece. For the Rams that is fine, as these are practictically useless on a near-empty board, if not from the very beginning. But there are many more Pawns than Rams, and you would have to eliminate the Warriors too.

I have little doubt that 3 minors could checkmate a bare King, even on 18x18.

There's a large number of pieces which would struggle to checkmate a king. A player that's behind would be tempted to capture all of their opponent's pawns and hope that there is insufficient mating material.

One suggestion is for pieces to be able to promote to themselves plus a king movement when entering the opponent's camp, with the idea of king + any being able to checkmate a lone king.

Some pieces would benefit more than others, though. Also, the game would become more complicated as the piece values would change.

This is as large as I want to go: 2 x 96 pieces on an 18 x 18 board:

Rather than doubling many of the super pieces, I only doubled the Archbishop, and added a new (anti-trade protected) monster piece in a single copy (the Terrorino). Compared to Megalomachy there are two more Pawns (wider board!) and two more Warriors. The remaining 10 extra pieces are new and appear in pairs: The Spider (a new minor, AH), the Omega-Chess Wizard and Champion, the (lame) Squirrel (NAD), and the Snake (lame GHCZ), all rook class.

The setup has a few 'air holes', to limit the density to 60%. These should allow reasonably easy development of most pieces. E.g. after the Lion is gone, the Rhino and Amazon have a nearly free diagonal (assuming 5th-rank pieces are developed quickly). And on the other Wing the Griffon can leave over the m-file, and the Queen along the diagonal that this opens.

The Spiders start rather backward, but by their H move they can jump over the serial defense shield, to reach the Bishop squares (after these have been developed, which should have high priority, as those files are vulnerable to Cannon attack). After the Spiders move on, the Marshall, Terrors and Terrorino can use those squares, as these are all jumping pieces.

I was running out of ideas for adding other rook-class pieces. Leapers with 12 move targets can play that role, but I dislike having pieces that just add 4 moves to a piece that already participates. This is what you automatically get when you use an oblique atom in them, as these already have 8 moves, and the pure N, C and Z already participate. As an exeption I do allow the Omega-Chess Wizard (FC), because it is a well-known and popular piece, while the Omega-Chess Champion (WAD) is acceptable despite its overlap with WA and WD because the latter two in the variant I am contemplating would have the special ability to block flying captures, which makes them very different even from a normal piece that would move exactly the same. Since all combinations of pairs of W, F, D, A already do occur as minors, the non-inclusion requirement would limit sufficiently non-overlapping 12-target leapers to compounds that contain both G and H. This leads to unattractive and difficult-to-master move patterns.

But then it occurred to me that pieces can also be weakened by making them lame, rather than reducing the number of move targets. A lame 16-target leaper might have a value similar to a true 12-target leaper. And lame leaps are not counter-intuitive. (In fact every distant slider move is one, and we understand sliding.) In particular, the following pieces seem nice:

• A (lame) Squirrel (nAnDnN, where the non-jumping N move is the multi-path Moo.
• The nHnGnCnZ ('Snake'), where the nCnZ part is George Duke's multi-path Falcon.

The heuristic of the interactive Diagram guestimates Champion, Wizard, Cannon and Bishop on 18x18 around 500 (where the 8-target leapers and Vao are around 300), the Squirrel at 600, the Rook at 650, and the Snake and Dragon Horse (BW) around 750. For comparison, Queen is valued 1350, Marshall (RN) 1100, Griffon and Lion around 1200.

Thus this Squirrel and Snake fit well in the set of pieces that can be traded 2-1 for a super-piece (plus Pawn). Or together with a minor for one of the weaker super-pieces.

FAH would indeed be a good choice for a 12-target leaper on a large board. In this particular case it has a bit too much overlap with the FH (Frog) and FA (Elephant) that I already have. When I used the WAD (Champion) it was as a replacement for WD (War Machine). Of course it still had overlap with the WA (Archer), but the latter is really a special-purpose piece here, because of its blocking power for flying captures. If I would have a need for upgrading another minor, the FAH would certainly be a prime candidate as a replacement for the Elephant or Frog.

Having two copies of most super-pieces really unbalances the piece-value spectrum a bit. So perhaps an 18x18 variant should have more 12-target leapers. On these large boards the leapers get weaker compared to the sliders, and consequently slider-leaper compounds also start to lag behind pure sliders. E.g. the difference between Marshall or Archbishop on the one hand and Queen on the other increases with board size. In particular pieces with 8 sliding directions become very strong, and Griffon and Rhino also belong in that class. So perhaps I should not have two Queens and two Griffons, and introduce another pair of rook-class pieces instead. A sliding version of the FAH (BH) might also be a good choice, to stay on the high-value side of the Rook. (The 12-target leapers also suffer from the large board, compared to sliders.)

HG,

If you use the WAD, you may also think about the FAH as it works somewhat the same but with a bit more of a diagonal flavor.

I wonder if there would still be market for an even larger chess variant. After all, Dai Dai and Maka Dai Dai Shogi have 2x96 pieces, which still surpasses the 2x80 of Megalomachy. (But 18th-century Buddhist Monks might have been much more patient than modern westerners...) On an 18x18 board, maintaining the 6-rank distance between the armies, there would be room for 6 x 18 = 108 pieces in each camp. But with so many ranks in the camp there is ample room for 12 empty squares in it without making it too easy to avert the danger of a suffocated mate. E.g. put the empty squares only on the 3rd and 4th rank, in the wings. That would bring the piece density to 60%, which I think is acceptable. The 5th rank, directly behind the Pawns, would remain fully filled, so the slow minors could be placed there to give them optimally fast access to the battle.

Compared to Megalomachy I could put two copies of all super-pieces other than Amazon (which then symmetrically pairs with the King). That would give an extra Lion, Queen, Griffon, Rhino, Marshall and Archbishop, while there would of course also be 2 extra Pawns. So that already makes 8 extra pieces. I think with so many pieces it would require an extra pair of ultra-powerful pieces (Terrorinos or Terrorissimos) protected by anti-trading rules to keep the game fluid . With so many super-pieces there would have to be a wider shield against flying attacks to shelter behind. (Just placing the shield more forward would make it too easy to slip behind it from the wings.) This would need an extra pair of pieces that block flying captures. The WD (which was not in Megalomachy but replaced by the Omega Champion WAD there) could be given this role. That leaves room for two more piece pairs. A good choice would be the Omega Wizard (as we already have the Champion) as a rook-class piece, and perhaps two extra Warriors.

The Terrorino could be a Queen that could also capture as a limited-range Locust: on the 1st or 2nd square in each direction, by landing directly behind it (i.e. a tuned-down version of the Teaching King from Maka Dai Dai Shogi). Or, if that makes it too strong, just on the 2nd square, and then also a non-destructive jump to that.

OK, thanks! I also thought about 'Meteramachy' for 'Mother of Battles', but I suppose this is syntactically incorrect, because it would need the 2nd declination plural of 'machy'. So I suppose it should be 'Meteramachoon'.

(I googled for the declination of Greek nouns, to see how words ending with the letter eta would fare. And, belief it or not, the first hit I got actually used the word μάχη as example! :-) )

dear HG. No pb with your choice of names. After Rhitmomachy, Ouranomachy, Metromachy, we are into a secular tradition

Based on the experience of setting up the 14x14 variant, I would now do the 16x16 like this:

I was never happy with the Steward, so I replaced it by a novel piece that I called 'Scout'. I want this to be worth about 2 Pawns, so that it can take on a Pawn protected by a Pawn (if it has backup). But I am in doubt what move I should give it. Pieces with 4 captures and 8 noncaptures (e.g. Mat Winther's Alpaca) have this value, but are necessarily divergent. The alternative is a piece with 5 or 6 normal moves. But that would have to be asymmetric. In the Diagram above I chose mNcA, to not make it too slow, not make it color bound, give it two forward captures rather than one, and allow it to attack a Pawn without being attacked back by it. (The most suitable asymmetric alternative would be fhNbD.)

Perpetual checking is also a concern. A flying Queen would be much to good at that. In Tenjiku Shogi that is not a problem, because repetitions (other than check evasions) are forbidden. (But even then it can take a long time before a flying Queen runs out of new checks.)

Something would have to be done against this, or players would just preserve their Eagle until the board population thins enough to draw by a perpetual, when they are in danger of losing. Where a Rook on a near-empty board cannot check a King forever (so that the Raven is not really a problem), and the color-bound Bat is no problem at all, we know that even a single piece to shelter behind is woefully inadequate defense against checking by a Queen.

A similar problem could occur through perpetual chasing. But only the Terror is worth more than an Eagle; other pieces can simply be protected when the Eagle attacks them. This is a second reason to extend the anti-trading rule of the Terror to capture by an Eagle: if capturing a protected Terror is forbidden for Eagle as well as Terror, the latter can be protected too to put an end to any chase.

The Eagle should definitely be able to check a King, though; that was the entire point of including it. To not interfere with normal chess-like play too much, the repetition rule could state that it is forbidden to repeat a position through a move with an Eagle that delivers check.

I am still a bit in doubt with regard to the Eagle (= flying Q) vs Terror piece value and the anti-trading rule. If the values are too close, trading Eagle for Terror might be common. And this makes the anti-trading rule for Terrors ineffective; both the Eagles and the Terrors would disappear from the board fairly quickly. Only when the Terrors stick out so much that you would not trade them for anything else, a ban of trading them for each other would result in a prolonged existence.

Now in the late end-game, when little is left to jump over, the Eagle would become similar to a Queen. Which is definitely a lot worse than a Terror, which in addition has Knight moves and hit-and-run capture. But in the opening it is not so clear-cut. The Eagle has an enormous number of capture targets on a crowded board. If it would initially be worth more than a Terror, and its value would drop as the board gets emptied, there would be a point where the values are similar, and trading would be a good option.

One could argue that even though the Eagle can strike deep behind enemy lines, most of these strikes are useless, because pieces there tend to be protected, and the Eagle is too valuable to be traded for those. And even when the targets are not protected, capturing those is risky: although the Eagle can enter the opponent's camp easily to capture, it cannot easily withdraw to safety afterwards with a non-capture. But it would usually be able to sacrifice itself for a second piece, and two Queen-class pieces might be a good trade for it. Values guestimated by simple move counting doesn't take account of such subtleties.

If the Eagle turns out to be too valuable, it should be subject to an anti-trading rule as well. Or at least the rule that Eagles are also not allowed to capture a protected Terror. I wouldn't mind so much if they capture each other; it is the Terror I want to survive.

@Jean-Louis: I have been thinking about how to name these large chess variants, and want the name to already suggest they are large. One idea for this was using Makromachy or Megalomachy. But since these are obviously inspired by Metamachy, I first want to check with you if you are OK with that.

Only now I notice that in the 14x14 setup white can threaten mate on the first move: 1. i5 ... 2. BAn9#. But this can be easily thwarted by 1... m10. Even if white would manage to back up the Bat by an Eagle (the only other flying piece that can move on the white diagonal), pushing up l11 would provide a second (X-ray) defender of n9 in the Archbishop. Both these 'forced' Pawn moves would be healthy development moves, liberating the Bishop and Dragon Horse. So nothing would be gained by white in pursuing this.

The players would have to be careful not to push their m-Pawn more than a single step, however; this would leave them with little defense against the BAn9/n6 threat. (The Archbishop alone is not sufficent defense, as it is worth more than a Bat, and the Bat would arrive on n6/n9 protected by the Terror.) At least until the l-Pawn gets pushed. After that DHl11 (l6) would provide an alternative defense.

So I don't think this is a problem. It just shows that the presence of the flying sliders keeps the players on edge, even early in the game when the King is still deeply buried. Which was exactly the point for including them. I suppose this is the equivalent of fool's mate in orthodox Chess, where the players have to mind the weakness of f2/f7.

I like it too. Maybe I would consider switching Cannons and Vaos. (To avoid a threat on a11 and n11).

Funny you should say that, because swapping those was actually the last change I made before posting. The threat to a11/n11 is not really a serious threat, right? Those are doubly protected Pawns, with a low-value protector (Knight). And even if they would have been unprotected, b11-b10 (m11-m10) would solve it through a healthy development move. It is probably worse that it also attacks b10/m10, but it is unlikely you would want to put the Knight there anyway.

Besides, the diagonal d1-n11 is not an easy exit for the Vao. It has a Rook in the path, which isn't going anywhere for some time, while moving the Archer would severely compromise the shielding of the noble pieces from flying attack. And the diagonal d1-a4 is blocked by a Warrior that will have a low priority in moving up, and a Dragon Horse, which will likely also be 'stationary developed' by just moving up the c-Pawn. From e1 the Vao could develop along the diagonal e1-a5, where the Knight and Zebra are likely the first pieces that will leave the camp, while b4 will be pushed to let out Ba3. The Vao will then be able to attack the opponent camp from a5, b4, c3 or d2 just as fast as it would be able to attack n11 from d1. But achieved through healthy development moves, rather than cumbersome manoeuvres that only weaken the position.

I worried much more about Cannon attacks. These cannot be solved by simply pushing a Pawn. And the two central files are very vulnerable, with 3 high-value pieces that can be skewered. Even the Bat is worth a Cannon + minor. It was a headache to make sure some minors could be put in front of the Bats in time to block the skewer.

I like it too. Maybe I would consider switching Cannons and Vaos. (To avoid a threat on a11 and n11).

I also like the fast castling which seems well suited for these large CV.

A somewhat smaller version (56 pieces instead of 80) could look like this:

The piece density has been reduced to 56%. To simplify the learning curve the number of piece types has been reduced from 33 to 24, and apart from the Warriors and the Terror there are no pieces that you would not encounter in other variants. The Ram, Steward, Prince, Frog, Nightrider, Leo and Spearman have been eliminated, and the War Machine has been replaced by the (upward-compatible) Omega-Chess Champion to rebalance minors vs Rook-class pieces. From the 'nobles' the Lion and Amazon were eliminated, and the Eagle and Terror only appear in a single copy.

The initial position is vulnerable, because there are several pieces more valuable than a Cannon directly behind the Pawns (Bishop, Dargon Horse, Bat, Archbishop, Marshall), and Cannon attacks are not so easily disarmed as Vao attacks (where one can interpose a neighboring Pawn). Cannons and Vaos start on the back rank, though, so it will take some time to mount an attack with those. This allows the players to prepare their defenses for such attacks, but puts them under pressure to do this quickly.

The Marshall and Archbishop can simply move away when they get attacked; there is only a Zebra behind them, so they cannot be skewered. The Bats are in serious danger, though, as the most valuable pieces shelter behind those from flying attack. It is important here to prepare robust interposition of low-valued pieces to block Cannonattacks on them. After moving up the center Pawns the Zebras and Elephants can be used for this, and can be 'secured' by Vao or Champion protectors.

The Eagle (flying Q) has been moved to the back rank, to provide an extra protector for all other pieces there. So only the flying R (renamed to Raven as a mnemonic for this) requires an aerial blocker to prevent immediate trading, and the Archer (WA) satisfies that need.

The d-, e-, j- and k-files can be used for developing orthogonal sliders, the short diagonals from f1 and h1 for diagonal sliders. The Griffon and Rhino are placed behind the shield of flying pieces so that they can smoothly slide onto those rays. The King is placed off-center to get the Dragon King with which it (fast-)castles next to such an 'exit corridor'.

The Warriors now only have to move up a single step to join the Pawn wall, which then still leaves their backward Knight captures cover the back rank. This might be a good way to build a 'cannon-proof' castle for the King.

Would it break the game to allow 2 moves/player after white's initial turn 1 single move?

Probably not more than it would break most other variants.

But the goal was to design something with the character of Chess, but (much) larger. Introducing weirdness that has no bearing on size only makes it more alien. Perhaps the Terrors, with their hit-and-run capture and anti-trading rule are already too much.

Would it break the game to allow 2 moves/player after white's initial turn 1 single move?

Indeed, I agree with what you say here. I like to have a better balance between light and heavy pieces. It is easier to gain a weak piece than a strong one, and if there are too few weak pieces (or none left) then it becomes difficult to increase your advantage. This is one of the reasons end-games without Pawns are so drawish.

The large Shogi variants are overdoing it, though, by having too many excessively weak pieces (like the Shogi Knight, or Stone General, with just two moves). And many of the somewhat stronger pieces are steppers, which take a long time to move to where the action is. There is something to be said for Camels and Zebras on large boards!

I am starting to feel quite good about this latest setup. I should still play test it though, at the very least as computer self play. I can try to make the Diagram play against itself, but I am not sure it would be strong enough at affordable thinking times.

@HG,

A long time ago you have told me in the context of Grand Chess that having more heavy pieces shorten the game. That is true both there and here. But from my experience it takes a lot more to proper calculate when many heavy pieces are present. This gives to me rather a sentiment of randomness. To do it right you need a lot of time each turn even for computers. I want to argue that although having a lot of heavy pieces shortens the game as number of turns, it does not on time on the clock. That being said, you have made a nice pyramid here. But I do hold against Grand chess that is has few light pieces.

Well, the Vaos were put there for a reason: to prevent diagonal attack on the Falcon and Eagle (flying R and Q) behind them while the latter have no room to move yet. The Interactive Diagram estimates the value of Falcon and Eagle similar to Lion and much better than Amazon, respectively, while the Bat, which could make such attacks, is estimated as weaker than a Nightrider or Dragon King. E.g. 1. l7 followed by 2. ARi6 would threaten 3. BAp9, gaining more than the value of a minor. (Or BAo8 after protecting that square, to gain more than a Queen value.)

I understand the principle of not allowing the Vao to quickly trade itself for something more valuable. I always considered the hoppers only about half as valuable as the corresponding sliders, which would make the Vao worth much less than a typical minor. If I should believe the heuristinc of the Interactive Diagram, the situation is different on 16x16, though. (I have no empirical values for 16x16.) And this stands to reason: the Vao (as well as Cannon) is an unlimited-range piece, and the value of such pieces should go up with the board size relative to the short-range leapers. According to the Diagram the value of the Vao is about equal to that of Knight, Camel, Zebra, Elephant, Frog.... (The Bishop, being a slider, is of course worth much more.) So there is no advantage in quickly trading a Vao for any of those.

On the contrary; the Diagram estimates piece values at a total piece density of 25%, and the value of hoppers tends to decrease towards the end-game. So in the opening, with a piece density of over 50%, the Vao would probably be worth significantly more than any minor, with the exception of the Bishop. This is basically the same argument as why it is bad to play 1.CxH in Xiangqi.

So I think having the Vaos hitting the opponent camp in the initial position is no longer a problem in a variant this large. As long as there aren't many very valuable pieces directly behind the Pawns.