I would like to have some feedbacks from you about KingsDrop, my new chess variant.
While I was designing KingsDrop I wanted to create a game with this main goal: the game should have a big strategic depth. (A minor goal was to keep the regular equipment of chess.) I divided this main goal into three sub-goals: 1. huge game-tree 2. great clarity, that is easy to look ahead or easy to see down the game tree 3. more global winning conditions To reach the first sub-goal my game has: a. multi-move turn b. not fixed setup c. drop rule To reach the second sub-goal my game has:d. pieces with easy-to-visualize-movement e. short range pieces f. limited squares where to drop the pieces g. multi-move with each piece can move just once per turn To reach the third sub-goal my game has: h. a player can win by getting one of the minor pieces to the last rowi. there are two kings to protectKingsDropAll rules of Orthodox Chess (FIDE Chess) apply, but with the following modification. PIECES(number of pieces for each player)4 Spearmen: move one square straight forward or sideward4 Archers: move one square diagonally forward 2 Rooks: like FIDE Rooks but they move up to three square orthogonally (no castling)2 Bishops: like FIDE Bishops but they move up to three square diagonally 2 Queens: like FIDE Queens but they move up to two square orthogonally or diagonally 2 Kings: like FIDE King, they move to any orthogonally or diagonally adjacent spaceSETUP The game begins with an empty board. White player places in his first two rows all the pieces. The order of the pieces is up to White. When White has finished, the same procedure is followed by the Black.A possible configuration:
TURNA player can make up to three moves per turn. He has to do the maximum number of moves available. If he has no move available, he passes.One move consists in one of the two following actions: - dropping a new piece in the board from the reserve; - moving a pieces on the board that it is not moved or dropped in the present turn. Just one drop per turn is allowed.EXCHANGE When a player captures a Spearman, an Archer or a King, this piece is permanently removed from the game; but the other pieces (Bishop, Rook and Queen) go into the opponent's prison. At the end of the turn, if a player has in his prison a type of piece already present in the opponent's prison (Rook and Rook; Bishop and Bishop; or Queen and Queen), there is the prisoners' exchange: the white piece in the black prison is moved to the white reserve, the black piece in the white prison is moved to the black reserve. The prisoners' exchange is mandatory and it is not considered a move. DROPPING Dropping consists in moving a piece from own reserve to the board, placing the piece in an empty square adjacent to one of the two Kings.THE KINGS - No check or checkmate. - The King can move into a square under attack. - The King can be captured like the other pieces. OBJECTIVEThe winner is the first player who reaches one of the two following objectives: - capturing one of the two opposite Kings; - getting a Spearman or an Archer to the last row (the eighth row for the white player and the first one for the black player). The three times repetition of the same position (considering the pieces on the board and in the reserve) is a loss.
The repetition rule ('first to repeat 3rd time loses') is usually very unsatisfactury. It often occurs that you can force the opponent to repeat first by perpetually checking him. Games employing this rule therefore usually exempt perpetual checks (making the checker lose in this case, no matter who repeats first).
The exchange of prisoners is an interesting idea for allowing Shogi-like drops with Western Chess material.
I think the first and second sub goals are contradictory. The bigger your game tree is, the harder it is going to be to look down it. That's the usual idea of increasing the size of the game tree: to make it hard on computers that rely on looking down the game tree.
This shows in how you achieve those sub-goals. For the first one you add multiple moves and drop moves, both of which increase the number of moves possible on each turn. Then for the second sub-goal you use short pieces, limited drops, and limit the multiple moves; all of which decrease the number of moves possible on each turn.
I would switch the second sub-goal to making the game more strategic. Limiting the range of moves works there. If you can't shift your attack wildly, it becomes more important to put them in the right place to start. But maybe drop the need for easy to visualize moves, allowing for some odder short range pieces. Especially ones that can't go in certain directions, because that makes their moves have greater impact on the position in the long term.
HGMuller, thanks for your reply. I consider you to be an expert about the chess variants subject. I know your great work on chess engines.
I happy you found the exchanging rule interesting.
About repetition rule and perpetual check, I'm open to solutions. Do you have any idea?
For now, I think this is not a problem. With 3 moves per turn and no check, I think that situations is extremely rare, to say the least.Anyway, to my knowledge KingsDrop is the first member of the chess family to have 3-move per turn. Do you know other chess variants with that protocol?
The game propriety clarity (https://webcache.googleusercontent.com/search?q=cache:jBacr89P5lYJ:www.thegamesjournal.com/articles/DefiningtheAbstract.shtml+&cd=1&hl=it&ct=clnk&gl=it) refers about the human skill to look ahead, a psychological capability, in contrast with computer power.
My personal belief is that it is difficult to design a deep game (big game-tree) where you can see several moves in the future. Difficult but not impossible.
Here a interesting link (http://www.logicmazes.com/games/tree.html) where the father of Ultima talk about the lacking of clarity of Ultima.
From a psychological point of view, pieces with difficult-to-visualize movements (and capture mechanisms) increase drastically the game obscurity.
The key point here it is, I think, that the human brain makes shortcuts using rule of thumbs (most of the time they are not verbalized).
The short-range pieces is another factor helping to improving the clarity of the game. In a game with long-range pieces it is very difficult to foresee the general board configuration because you cannot make many assumptions. But in a game with short-range pieces you can make a lot of assumptions.
Look at Arimaa. A very difficult game for computer (because its big game-tree) but “easy” for humans.
Look at Othello/Reversi. Many people say Othello is a very obscure game but the game-tree is relatively small. This is because the flipping mechanism is counterintuitive and it changes a lot the board configurations (so it is difficult to look ahead.)
But you're addressing the issue of clarity mainly by reducing the search tree, which is counter to the other work you are doing to increase the search tree. My point is that if you give up on some of your clarity you can move closer to a game where the rules of thumb become more important than the tree search, which seemed to be your point.
I'm familiar with Arimaa, and it was obvious you were taking inspiration from it. It's just not clear to me that the claims Arimaa fans are making have really been tested. And I've never found Othello to be that hard to visualize. It may not be the way we're trained to think, but that doesn't mean you can't learn to think appropriately.
I cut the parts of the game-tree that I think cannot be easily visualizable and intuitive. I think the actual version of KingsDrops has already a huge game-tree, so it is not my main goal to increase it.
Anyway, your inputs are making me thinking. I'm taking some time to ponder abot them.
If you have other suggestions, let me know. I posted KingsDrop here for finding inspiration, so any feedbacks are very welcome.
I would really like if someone could play KingsDrop and post impressions here.