Variants, way to add new pieces without changing board size

[musketeerchess]

Hi
In the rules of musketeer chess, we add two new pieces to a classic chess set of pieces. The game is played on 8x8 board.

We put each of the 2 new pieces behind one of the classic pieces. They are immediatly dropped on the board when the classic piece moves for the first time. For more detailed rules see www.musketeerchess.com

This specific rule was not accepted easily by the chess community.

Thus i'd like your help to see how we can smoothly add the new pieces, but by "defining" a clear strategy as in classic chess, where the initial placement of the pieces is known.

[hgm]

Of the options you give I favor the Seirawan gating method (1). It gives an interesting dynamic between revealing your intentions, and keeping your options open but exposing yourself to tactics that might destroy your gating opportunities before all pieces are in. The always-drop or initial-substitution method don't have anything comparable. A fixed square of introduction is a bit static, and leads to a smaller opening game tree.

A fifth method that comes to mind is to allow the King to always 'gate in' a piece when it moves. You will always have a King, so gating opportunity can never be destroyed. On the other hand, you wold not want to move your King too often. Presumably you would make your first gating on e1 when you castle, and the second on c1 or g1 stepping towards the corner. This might make Q-side castling more popular, as b1 is an excelent place for the King, and you have to spend a tempo on the King move anyway to get your second piece.

[musketeerchess]

I didn't remember the exact rules in Seirawanchess.

When you see the following link http://www.seirawanchess.com/ you'll see that Seirawan introduces his pieces almost in the same manner as i do in Musketeer concerning gating of the new piece. The only difference i think is that in Musketeer Chess i "define" clearly the strategy of introduction of the new pieces, which can lead to some subtle tactics (when a piece is captured on it's initial square before it moved for the first time or the king is checked and it's obliged to move, the right of gating is forfeited!

[hgm]

... (when a piece is captured on it's initial square before it moved for the first time or the king is checked and it's obliged to move, the right of gating is forfeited!

Yes, but these are purely illusory complications: as you have to fix in advance where the piece will be gated, there is no point in delaying it: the opponent already knows. So you will in general gate at the earliest opportunity, long, long before there is any chance the piece can be threatened at all. And no one would be so stupid to choose gating at the King square, as you cannot quickly develop the King.

In Seirawan Chess you are under pressure to delay the gating, so that the opponent has to reckon with the pieces appearing anywhere, and while you develop other pieces the number of gating opportunities shrinks. This gives a real dilemma. Therefore I consider the Seirawan gating rules as highly superior to the Musketeer gating rules.

[Jhoravi]

Therefore I consider the Seirawan gating rules as highly superior to the Musketeer gating rules.

I suggest we start with Seirawan as a foundation and fix some of its flaws I would like to mention.

1. Seirawan gating forfeit may lead to confusion when a piece like Knight or Bishop is moved back to the first rank as it often happens in many openings. When it moves again it's Difficult to track whether it's the initial move or not. My Suggestion is ability to introduce the new pieces even if it's not an initial move from the first rank.

2. When castling, the gating should EXCLUDE h1, a1,b1 because reasonably and visually these square are not in the castling path. But my stronger reason is that h1,a1,b1 strengthens the kings defense too much which discourages exciting kingside attacks.

[musketeerchess]

Hi Seirawan chess is a far older variation then mine. So the rules of gating in Seirawan Chess got sufficient time to enroll your approval.

I think i'll change the rules of Musketeer Chess to stick to those of Seirawan Chess. I also think that the rules of my pieces (Leopard and Cannon in particular) are by far more interesting than those of Seirawan (rules of Chancellor and Archbishop actually in Musketeer Chess). That's because my pieces are strong enough to challenge Queens and Rooks, but also "fragile" which allows the opponent to attack them with Bishops and Knights.

This makes the game more subtle. On the contrary, the tactical possibilities are probably less spectacular than in Seirawan Chess where we litterally got 3 Queens on the 8x8 board which imho is too much

[hgm]

I agree on your assessment of Seirawan Chess: 3 Queens is too much for such a small board. In Gothic Chess (which has a 10x8 board) this is much better, and I don't rank Seirawan Chess as a particulary good variant. Below average, in fact. Weaker pieces are often more interesting.

As to remembering which pieces are virgin, one also has to do that in orthodox Chess for King and Rooks in connection to castling, and this is in general not seen as a problem. It doesn't happen very often that you retract your Knights or Bishops to their starting squares, and when you are forced to do so you typically remember. Especially when for every move of a virgin piece you must have thought about the decision to gate or to postpone it.

I am not sure what you want to say about a1, and h1, or when b1 even gets mentioned. In Seirawan Chess you can gate with a castling on the King or the Rook square (but not both). Gating on a1 or h1 is in general a very poor move, however, as the new piece is locked in by pieces you cannot move without weakening your King fortress, and has only a single escape through its Knight jump, which can be kept easily under cover by an opponent minor, so that his minor neutralizes one of your super-pieces for the rest of the game. It is just one of these rule complications that doesn't affect the game at all. I also consider that an indicator of poor design.

[Evert]

1. Seirawan gating forfeit may lead to confusion when a piece like Knight or Bishop is moved back to the first rank as it often happens in many openings. When it moves again it's Difficult to track whether it's the initial move or not. My Suggestion is ability to introduce the new pieces even if it's not an initial move from the first rank.

This really changes the dynamic of the game considerably, and I think it's a bad idea. Yes, you have to remember if a piece moved or not, but in principle the same applies to the king and rooks for the purpose of castling rights. In practice, you need to gate the pieces in when a piece is first developed and you should develop your pieces before you start shuffling them back to the back rank, so it shouldn't be a real problem.

2. When castling, the gating should EXCLUDE h1, a1,b1 because reasonably and visually these square are not in the castling path. But my stronger reason is that h1,a1,b1 strengthens the kings defense too much which discourages exciting kingside attacks.

You cannot gate on b1 when castling.
a1/h1 are also poor squares (compared to e1) for the hawk/elephant, so I don't think being able to gate there is much of a concern. Yes, your kingside defence is stronger, but these are super (queen-class) pieces; you're better off using those to attack than to defend.

[Evert]

Hi Seirawan chess is a far older variation then mine. So the rules of gating in Seirawan Chess got sufficient time to enroll your approval.

I think i'll change the rules of Musketeer Chess to stick to those of Seirawan Chess. I also think that the rules of my pieces (Leopard and Cannon in particular) are by far more interesting than those of Seirawan (rules of Chancellor and Archbishop actually in Musketeer Chess). That's because my pieces are strong enough to challenge Queens and Rooks, but also "fragile" which allows the opponent to attack them with Bishops and Knights.

I think having essentially "Seirawan chess with different pieces" is better than another (complicated) gating rule. I don't actually like the gating rules in Seirawan chess very much (I think they're clunky) but they're a good compromise to just starting with extra pieces in hand and dropping them whenever and just swapping pieces in the initial array.
The problem with Seirawan chess is that the board gets very crowded: it's just too small to accommodate three queen-class pieces. It's much more interesting if you have more minor (knight-class) or medium (rook-class) pieces.

Not to mention, if you have Seirawan-style gating rules, it becomes relatively easy to adapt any engine that can play Seirawan chess to play this variant.

This makes the game more subtle. On the contrary, the tactical possibilities are probably less spectacular than in Seirawan Chess where we litterally got 3 Queens on the 8x8 board which imho is too much

Agreed.

[musketeerchess]

Hi Evert
I'm still looking for programmers to help implement Musketeer Chess.

If you can help, it'll be a pleasure.