I can top that. My first TRS-80 had 4K of memory. I actually wrote a Kriegspiel referee program in TRS-80 Level I BASIC. Each player, in turn, would enter his passnumber, then see his position (his own pieces only) on the display, and enter his attempted move.
If the move, because of the positions of his opponent’s pieces, was illegal, the program would say “ILLEGAL” and ask for a different move to be entered.
If the move was legal, the program would respond with “OK” or “CAPTURE”, as the case may be, and would also display “CHECK ON THE FILE”, “CHECK ON THE RANK”, “CHECK ON NE-SW DIAGONAL”, “CHECK ON NW-SE DIAGONAL”, and/or “CHECK BY THE KNIGHT” as appropriate. Then the player would press ENTER, the screen would go blank, and the opponent would be asked for his passnumber.
The first player would then step away from the computer, the opponent would step up and enter his passnumber, see the current position (his own pieces only) along with the same “CAPTURE” and/or “CHECK” messages the first player had just seen (along with an asterisk on the capture square, if any), and be asked to enter his own move, etc.
I did not implement that feature on my TRS-80 version. There’s a limit to what even we 4K wizards can be expected to do in 4K!
The Hoyle’s Games book I read in the 1950’s mentioned that rule, except that “Any?” did not mean “Do I have any legal captures?”, but rather “Do I have any legal captures with my pawns?”. It was designed as a convenience feature, so that the player wouldn’t have to try every possible pawn capture. Accordingly, the rule was that if you ask “Any?” and the answer is “Yes” (or “Try”), you must keep trying pawn captures until you find the legal one (or one of the legal ones).
Note that with pieces other than pawns, if you try a move you believe will be a capture, and it turns out not to be, your attempted move will (in all probability) still be legal, in which case you are stuck with it.
Actually, my Hoyle’s Games book modified the “Any?” rule a bit, so that the player asking it was only obligated to try one pawn capture, and if the first try failed, the player could abandon the search at any time and play a different move. This made the “Any?” rule more than just a convenience rule; now it was a strategy rule too, i.e. it could be used to find out information about your position.
When I was in college, we had a number of variants like
Scotch chess. White makes 1 move, Black makes 2, White makes 3, etc. The series ends with a check and the first move must be to get out of check. After 1. e4 2. e6 Be7, Black threatens a number of checkmates. Best is 3. d4 Bg5 Bxe7 and both sides are in the game.
Two for one. White starts with four center pawns on the 3rd rank and king at e2. White may step into check as long as he is out of check at the end of his move. Black wins but only after really good defense.
Kriegspiel. Requires three sets and a referee. Players are told if a move is legal and, if a check, where it came from: the rank, file, long or short diagonal, or with a knight. They are told where captures happen and if a pawn can capture anything or if a move is legal.
Kamikaze. Players arrange their pieces on their side of the board behind a partition. The partition is taken away and the clock started.
Kamikaze-Kreigspeil. From the people who brought you WW II the above two are combined.
Giveaway. Nothing sacred about the king here. Captures are mandatory, but a player can choose which one. It resembles pool more than chess.
Cylinder. Pretend the board is a cylinder in that the a file is just to the right of the h file. A knight can move from g1 to a2 or a diagonal can run from f1-g2-h3-a4-b5-c6-d7-e8. Note on an open board a bishop on f1 controls the same squares as on b5, but can’t be captured there.
Most of them were from the old Harkness Handbook, the tournament director’s guide then.
My earliest programs related to chess was one to proofread (play) a game in AN, solve 8 queens, knight’s tour, and RR pairings.
Kreigspiel played over the Internet would be easier, because the two players are already separated. (I’d be amazed if SOMEONE hasn’t already implemented it.)
When we used to play it at the Northwestern chess club (with Bill Smythe as the referee), players were permitted to put pieces on the board where they thought their opponent’s pieces were. The guesses were interesting in their own right.
We once played around with chess on a sphere, with the pieces arranged around the two poles.
This turns out to have some interesting behavior for pieces that move over the poles. (Try it and see.)
I’ve thought about how to play it on other shapes, like a Mobius strip, but figuring out how to place the pieces initially is challenging.
One of my first large-scale programming projects (on a Control Data 6400 computer, the same one used by Larry Atkins and Dave Slate to win the world computer chess championship) was multi-dimensional tic-tac-toe. 3D tic tac toe (on a 4x4x4 cube) wasn’t too hard, 4D (on a 5x5x5x5 hypercube) started to get challenging, and 5D (on a 6x6x6x6x6 polycube) got time consuming, even when the computer was making all the moves. I was working on playing a game in 6D when the school year ended. Once you’ve got the basic structure set, adding another dimension wasn’t difficult.
I’ve seen some non-mathematical discussions on a theory for N-dimensional tic-tac-toe proposed, but if there’s been a proof of it, I haven’t seen it. I got started on the 3D program around 1970 after playing many games of it with some dorm mates, and deciding that it was probably a draw, though we saw wins for both colors. We simulated around 10,000 games, with two of us writing competing modules to choose ‘optimal’ moves, and concluded that with perfect play it was probably a draw. Not a proof, though. I think we only got about 300 4D games completed, and a few dozen 5D. I don’t remember if we actually completed a game in 6D.
It may not be an ‘interesting’ mathematical exercise, compared to, say, Fermat’s last theorem, the proof of which has opened up some interesting areas in mathematics.
I was guessing that, because in your examples you always had m=n+1, you had proved (or speculated) that m>n was a necessary condition for the game to be a draw with best play, while with m=n (or of course m<n) it was always a forced win for the first player.
If 4x4 tic-tac-toe is any indication, it is too easy for players to block all winning positions when M>N+1, but I haven’t looked at game strategy since I was in grad school, and even then it wasn’t part of my PhD program, even though MIS was one of my minor areas.
Yes, one rule was that all conversations with the (human) referee had to be heard by both players.
Sometimes a player would ask me (as referee), “If I were to play [ such-and-such a move ] , would it be a capture?” and I would have to respond, “Sorry, I can’t tell you that. You’ll just have to try it and see. And then, if it’s legal, you’ll be stuck with it, whether it’s a capture or not.”
I once saw a Kriegspiel game begin 1. Na3 g6 2. Rb1 Bg7 3. b3 Ba1 4. Rxa1. Apparently black had tried that “bishop takes rook” stuff once too often in their previous Kriegspiel games.
Yes, my TRS-80 version implemented that too. Saying IMPOSSIBLE instead of ILLEGAL tends to thwart the idea of trying a zillion impossible moves to make your opponent think you have a zillion pieces.
Your opponent knows EXACTLY how many pieces you have, because captures are reported to both players, he just doesn’t know which ones you still have or where they are.
True, but some players don’t bother to keep count. As a human referee, I would always allow the player to ask me for the opponent’s piece count, and I would respond to both players. My TRS-80 version always displayed the opponent’s piece count along with the diagram of the player’s pieces.
I recall those early TRS-80 computers. Storage of data was on cassettes that would operate with a normal cassette recorder. That storage method was iffy. Black and white screen in a big gray box terminal with an attached keyboard. 8088 chip. 4K of memory was pretty small. Was the program written in BASIC? As I recall, drives to handle 5.25 floppy disks were not added until the machines were given 64K memory. There were no programs at first. You had to write your own program to do anything at all. A math teacher friend tried to write a chess program. He was successful in getting it to generate random moves that were legal, but not much more.
As far as variants go, I think giveaway chess has a lot of appeal to beginners. The forcing nature of the game and the ability to think in planning chunks to give away pieces is an interesting precursor to actual chess thinking. Kids discover early that blocking lines, short moves, and putting knights in odd places are good defensive strategies and completely opposite to regular chess. There are in between move captures when two or more captures are possible, requiring calculation to determine which move is best. Giving away bishops and the queen is easy. If you make a mistake, your rook may decimate many pawns and pieces in a windmill like action. The king may die early, especially if you play moves like 1. f3 2. Kf2 3. Kg3 as an opening. It is not a game that I would recommend between rounds of a tournament. After 1.d4 Nf6 2. Bh6?? your OTB regular chess playing opponent in the next round will look at you quizzically and then take the bishop. All you can do is sigh.
If the keyboard was attached, you had a Model III TRS-80, probably with 16K and Level II BASIC – quite advanced compared to my Model I with 4K and Level I BASIC. I later upgraded to a Model III, and wrote an improved version of my Kriegspiel referee program – still in BASIC, though.
None of the TRS-80’s (Model I or Model III) ever had an 8088 chip. They were Z-80’s, which had a much better instruction set.
My bad about the chip. It is over 30 years since I last saw a TRS-80 Model III in use. I believe the 8088 chip was in the clones. The TRS-80’s I remember were used to teach programming skills to high school kids by a math teacher. In my library at home, I still have some books on learning BASIC, how to install a hard drive, and a large 3 ring binder with info on Deskmate for the allegedly portable computer Tandy sold. It weighed around 40 pounds, two 5.25 disk drives. There is a box with a bunch of empty floppies and a box of 3.5 diskettes, still in the shrink wrap. I also have a book with programming for Sargon, an early chess program. At one time I picked up a talking Fidelity Chess Challenger at a chess club auction for $10. The device was in a small briefcase. It was not too bad. The kids in my class liked to play it. It still works, but the voice chip went kaput a long time ago.
Upon further review (of my fading memory), I believe that the Z-80 chip and the later 8086 chip (and its 8-bit cousin, the 8088) were both extensions, in different directions, of the old 8080 chip, which was in turn an extension of the even older 8008. (Or something like that.) All of those came from Intel, except for the Z-80 which came from competitor Zilog. I don’t believe there were ever any “TRS-80 clones”, although there were some Z-80 clones (chips). Programmers of assembly language loved the Z-80 because of its extra registers and expanded instruction set.