Comments by FergusDuniho

From the questions and expressions of confusion you got regarding what the rules are, it should be evident that you have not described them clearly enough. Also, you seemed to change your mind about some details in the comments. So, you should rewrite your rules to reflect anything you have changed your mind on and to supply the details that someone would currently have to read the comments to find out about.

the main feature of a switch that I'll retain is that spaces in the switch share some routes to and away from them, and movement from the narrow end can go in either of two different directions."

I'm sorry, but I didn't understand that.

I mean I am retaining the geometrical properties of the switches, but I am discarding the other rules regarding them.

Does that mean my variant is history? Or have I misunderstood something?

No, it means I am working on my own variant instead of putting pressure on you to change yours. You're free to go with the rules you want for your game. I'll see what I can do about programming it once you have settled on the rules and have described them clearly.

As I understand it, Fergus has decided to program a variant based on yours, and given it a different name to signal that it's not the same game.

Exactly this.

This game differs from yours only in that both a4 and 4, or both 5 and h5, can be occupied/passed through simultaneously.

There are some other differences. Here's a link: Reroute66. The rules are described below the board. I'll add individual piece descriptions to the rules later.

I'll admit I find it a little odd that such conditionally untraversible squares should be so difficult to implement (couldn't it be done with uncapturable dummy pieces that appear and disappear as the other square is occupied and vacated?)

That's the solution I already proposed. But I also think these rules are not essential to the core concept of the game, and I wanted to start with a variant that does not include them.

During the night I changed my mind on one of my rule suggestions. I had proposed allowing capture of a piece on the occupied space of a switch when a piece moves to the unoccupied space of a switch. This was to fit in with rules regarding no double occupancy of switches and blocking any movement through the switch when one space in it is occupied. These happen to be the most difficult rules to program, and I don't think they're really essential to the concept of a switch. So, I'm going to program a stripped-down version of Chess66 that I'll call Reroute66. This will treat each space as a fully separate space, and the main feature of a switch that I'll retain is that spaces in the switch share some routes to and away from them, and movement from the narrow end can go in either of two different directions. I'll get to work on it later in the day.

"A switch is an overlapping pair of spaces that in some ways operate together as a single space."

From my point of view it looks different. The square a4 of a switch is an independent square and is completely identical to the square a4 in normal chess. 4 (A4) is a composite square consisting of a half and triangle part of a4 and the new triangle due to board geometry. The new square 4 (A4) gets the same play options as all squares of the game board - 4 (A4) is considered equal.

This means that the squares a4/4 (A4) must first be seen independently. These independent fields get an additional function when they work together and act as a switch, as described.

I didn't say anything different. We're in agreement that the spaces composing a switch are separate from each other. But since they cannot be simultaneously occupied, and occupation of one space blocks movement through the other, and they share a common edge and a common corner, they also function in some ways like the same space.

It is possible to agree on the sides from which the switch should be accessible. We have clarified access from below and from the side, access from above is also not a problem and is already part of my proposal. The only question is whether it is access to the switch as a unit or to the individual square of a switch. According to my intention, the second applies.

Agreed. Use of the switch as a switch is possible only from below. That is due to the nature of how it works. If a piece moves to a switch through a diagonal move or a Knight move, it will have to stop on the switch, which will prevent it from using the switch to alter the direction of its movement. The same is true if a Rook or Queen moves to a switch from the side. It will have to stop there, which will prevent its use as a switch. And if a Rook or Queen moves to a switch from above, it will have only one path through the switch. So, it won't be utilizing the switch as a switch. This shows that access to the spaces composing a switch is a completely different matter than using it as a switch.

"Allow pieces access to the spaces of a switch from any direction, and when a piece moves to a switch, allow capture of any piece on the switch even if the capturing piece moves to the other space. For example, if a bishop is on A4, and a Rook moves to a4, consider the Bishop on A4 to be captured."

As described, access to the switch from all sides is not a problem. It is possible to agree on the proposal, but it does not fit my intention. However, if the game becomes more playable and programmable - so what.

What I'm proposing fits with the rule that both spaces of a switch cannot be occupied at the same time and the rule that occupancy of either space blocks vertical movement through the switch even when that movement would technically be going through the other space in the switch.

To the Knight: I can't allow anything here, but I can say what my point of view is. If we stick to the fact that the squares of a switch should be seen separately, then knight moves ending on the same line are not possible. However, as the game becomes more playable and programmable, compromises should be possible.

We could just say that because of the way that switches affect the geometry of the board, some spaces may be reached by either a vertical move or a Knight move. This would also correct the injustice done to the Knight of making the weakest piece even weaker around a switch while it gives all other pieces greater mobility.

"One last rule change I would suggest is to let Rooks, Queens, and Kings use their ability to move horizontally to switch between the spaces constituting the switch. This would basically involve lifting one more restriction on movement to the spaces making up a switch. "

I have problems with that. I have emphasized that the squares of a switch represent independent squares. This would rather mean that moves between the squares of a switch are possible.

Precisely.

The independence of the squares on the one hand and the functionality of a switch on the other compete with each other. Regarding the direct change between the squares of a switch, I tend towards the superordinate function, so a direct change should not be possible.

Allowing a piece that can move horizontally to move from one space to the other in a switch as a normal move does not affect the functioning of the switch. While the piece is on either space, other pieces can't pass through the switch, and once the piece leaves the switch, pieces will be able to pass through it again.

Move 17: N a4-c3 doesn't seem legal, but N a4-c2 would be legal.

It looks legal to me. Why wouldn't this move be legal? Meanwhile, a4 and c2 are along a diagonal, and N a4-c2 would definitely be illegal.

Move 25: N f5-c4 doesn't seem legal in same way, but N f5-d4 would be legal again.

N f5-c4 is a perfectly legal Knight move, and I see no reason why it would be illegal. Meanwhile, f5 and d4 are diagonally adjacent, and N f5-d4 is clearly illegal.

When you're examining the moves, be sure to click on Record, paste the moves into the box, and click on View. This will let you view each position. I think you made some wrong judgements, because you got confused about how notation for this game works. Vertical movement shifts to an adjacent file as it goes between the 4th and 5th ranks. So, the space immediately above d4 is e4, for example.

I have started writing up a brief summary of the rules to appear underneath the board for Game Courier.

According to my definition, the switch can only be operated from 'below'. After that, the move N c5-A4 would not be possible.

A switch is an overlapping pair of spaces that in some ways operate together as a single space. By their very nature, they operate as a switch only from a given direction. From below, a piece is given two choices for how to continue its vertical movement. However, this property of a switch is not a property of the individual spaces whose overlap constitutes the switch. So, restrictions on movement to A4 have nothing to do with your definition of a switch. They are simply arbitrary restrictions on movement.

There are already other examples of A4 being reachable from above. For example, a Rook or Queen on a8 could move to A4, and a Bishop or Queen on e8 could move diagonally to A4. You have also agreed that it makes sense to let it be reachable from the side.

I'm now going to propose some rule changes. Since the switch can make a piece more powerful, this should be balanced by making a piece using a switch more vulnerable. Allow pieces access to the spaces of a switch from any direction, and when a piece moves to a switch, allow capture of any piece on the switch even if the capturing piece moves to the other space. For example, if a Bishop is on A4, and a Rook moves to a4, consider the Bishop on A4 to be captured.

Currently, the Knight is the only piece whose powers of movement are restricted in the vicinity of a switch. Correct this by allowing a Knight to move to any space that can be reached by a step in an orthogonal direction followed by a step in an outward diagonal direction, or by a step in a diagonal direction followed by a step in an outward orthogonal direction. This is the normal definition for how a Knight moves in Chess and other variants, and it would allow the Knight to also increase its capabilities in the vicinity of a switch.

One last rule change I would suggest is to let Rooks, Queens, and Kings use their ability to move horizontally to switch between the spaces constituting the switch. This would basically involve lifting one more restriction on movement to the spaces making up a switch.

I'm done working on Chess66 for Game Courier today. The directions are looking good, but I still have to adapt the Pawns to work with switches, and I still have to prevent double occupancy of the switches. Since Firefox won't remember this after I shut down my computer, here are the moves I've been making to test things out. In the meantime, feel free to check it out, and apart from the issues I just mentioned I still need to work on, let me know if you find any illegal moves it permits or any legal moves it doesn't permit.

1. P c2-c4 
1... p f7-f5 
2. Q d1-A4 
2... q e8-H5 
3. Q A4-a7 
3... p f5-e4 
4. N b1-a3 
4... q H5-e5 
5. P c4-d5 
5... q e5-d4 
6. N a3-c5 
6... p e7-e6 
7. Q a7-a3 
7... b f8-e7 
8. N c5-A4 
8... b e7-g4 
9. P b2-b3 
9... b g4-H5 
10. P b3-b4 
10... q d4-b4 
11. Q a3-g3 
11... p g7-g5 
12. P d5-e6 
12... p g5-f4 
13. Q g3-h7 
13... r h8-h7 
14. N A4-d5 
14... b H5-g6 
15. N g1-h3 
15... n b8-c6 
16. N d5-a4 
16... b g6-h5 
17. N a4-c3 
17... q b4-c5 
18. P e2-e3 
18... b h5-h4 
19. P a2-a3 
19... q c5-h5 
20. N c3-a4 
20... q h5-h6 
21. N a4-c6 // - Check! -
21... p b7-c6 
22. N h3-f4 
22... q h6-e6 
23. N f4-h6 
23... q e6-d5 
24. N h6-f5 
24... n g8-h6 
25. N f5-c4 
25... b h4-e8 
26. P a3-a4

En Bw

I assume you mean me? That's one correct letter out of four, with two more misplaced ;)

The other one was turned upside-down but left in place. So, they're all there with some mix-ups. :)

I'm testing my code for Chess66 on Game Courier right now, and with a Knight on a3, I currently have its legal moves as a5, c5, c4, c2, and b1. Because 4 is right above it, it looks like b5 could be another legal move by going over 4 and a5 and turning right to b5. It could be argued that this is illegal because there is a single-direction path from a3 to b5. But if that rules that out, it might also rule out a5 as a legal move, because there is a single-direction path to it through 4.

Suppose I have a Rook on d4. Could it move to 4? Or could it only move as far as a4? Likewise, could a Rook on d5 move to 5, or could it only move as far as g5?

The move sequences given by Gerd in his comment are B d1–4; B 4–b5 and B d1–5; B 5–g6.

So, I gather that the following are all legal moves:

• B d1-4
• B 4-d1
• B e8-5
• B 5-e8

I think I misread an earlier answer. This actually describes one move on black spaces and the other move on white spaces:

If the bishop is on 4, then his path goes via b5, c6, d7 to e8. In the other direction it goes via b3, c2 to d1.

So, it looks like the diagonal from 4 to the first rank is a different color than 4, and the diagonal that goes from 5 to the last rank is a different color than 5, and that is how the color changing is done.

Does that answer your question?

No, it does not. I specifically asked for a series of moves, and I didn't get that. What will answer my question is a series of moves in the generic rule-blind style of notation I used earlier. You do not have to write anything in English to answer this question.

By what series of moves can a Bishop change to spaces of the other color?

Since a Bishop that moves to 4 could immediately switch to a4, is the reverse also true? Could a Bishop moving to a4 immediately switch to 4? More generally, is switching between these spaces available for any piece or only for Bishops?

I have begun to program Chess66 for Game Courier. So that 4 and 5 have separate file labels, I'm calling them A4 and H5 and distinguishing them from a5 and h5. I'm programming it with logical directions, which lets me define named directions as linked lists of nodes. The way directions are defined, the same direction cannot have two destinations from the same space. This means I've had to define additional directions to handle some lines of movement. However, the same direction can go from different spaces to the same destination.

One of the main difficulties I'm coming across is that sometimes it makes sense to treat A4 and a4 or H5 and h5 as separate spaces, and sometimes it makes sense to treat them as the same space. One idea would be to put an invisible dummy piece on the unoccupied space when the other one is occupied. That way, movement through a switch would be completely blocked.

At present, more work needs to be done to get the directions working properly.

I gather from the comments that my version is not easy to understand.

I don't know if it's that or if you just haven't explained it clearly.

But if I see it correctly, then there are no bugs in the variant and the set of rules is consistent.

With respect to the Bishop, you told me that it can occupy only 4 or a4 or only 5 or h5. I presume the same holds for other pieces. So, a Rook on one of these spaces would not command two files. Likewise, a Knight on one of these spaces could not move as though it were on the other space. However, that looks like what the Knight is doing in your example. It's on 5, and its ability to move to f4 or g3 makes sense in light of that. But its ability to move to e4, f6 and g7 would make sense only if it were on h5 instead of 5.

Can a Rook move through a switch on a single move? Which of these moves could be legal?

• R a1-a8
• R a1-b8
• R a8-a1
• R b8-a1

I think we've clarified that a Bishop cannot pass through a transfer field on its move. To switch color, it must stop on a transfer field, and then it may use its next move to switch to spaces of the other color.

My next question concerns which directions of movement are available to a Bishop on a transfer field. Let's say a Bishop is on a4/4. I presume it can move to a5, b5, c5, or c3, but it cannot move to a3. Likewise, a Bishop on h5/5 could move to h4, g4, f4, and g6 but not to h6. Is that correct?

I have indeed trouble with your question. I don't understand how "B b4-b3" or "B b3-b5" has anything to do with a bishop's moves.

You may be familiar with abbreviated algebraic notation, such as Bb3, which would mean to move a Bishop to b3. This abbreviated notation depends upon knowledge of the position and knowledge of the rules. Without these, you don't know where the Bishop is, and if a player has more than one Bishop, you don't know which Bishop to move there.

Since Chess variants are played by a variety of different rules, it helps to use a generic rule-blind notation that can work equally well for different variants. The notation I used explicitly spells out a move in a manner that someone could follow without knowing the rules of the game in question. First of all, "B b4-b3" is a typo. I meant to write "B b5-b3", which means a move by the Bishop on b5 to b3. This is a possible move only if there is already a Bishop on b5, and whether it's a legal move depends on the rules of the game. In Miller's Spherical Chess, for example, it would sometimes be a legal move. Likewise, "B b3-b5" means a move by a Bishop on b3 to b5. It should now be evident what these have to do with Bishop moves.