It’s a “robotic” board that moves the pieces by itself.
You can sometimes find “untested” (i.e., broken) ones on eBay for a reasonable price, and if you’re lucky they’re an easy fix. Mine was stuck because the lock slider had wedged something and the repair took all of 10 minutes.
Very clean engineering: a few screws gets you in, there’s a remarkably small PCB, few wires and mechanical pieces: the main mechanism consists of two orthogonally mounted sliders with a stepper motor and belt each.
I don’t even play chess, but it’s amazing to watch it play both sides.
They also use a clever algorithm to route pieces around other piece since (obviously) the pieces can’t jump over other pieces given that they are moved by a magnet under the board.
Wow! Very cool. How did it manage to get a knight out as an opening move, for instance? Moving between the pieces, or moving other pieces out of the way?
Oh, and (more often) they move pieces out of the way before the real move. First time you see that it looks like an illegal move but it’s all good because it gets moved back :)
Very cool! Wouldn't it be even cooler if the museum could score a couple of the very oldest machines? I'm talking about the El Ajedrecista machine (1912)[1] and Caissa [2][3] (named after the goddess of Chess[4]) built by Claude Shannon.
We had a Fidelity Chess Challenger 7 when I was a kid.
I was a horrible chess player but painstakingly worked out a way to win as white, keeping a detailed log of my experiments in a notebook. The first couple moves were wildly out of book (because I didn't know book), and the computer with its limited Z80 processor always computed the same moves after that. Some googling [1] shows the board's Elo is 1300ish.
To illustrate the state of the art in 1979, the manual [2] explicitly calls out that it understands en passant and castling.
From the same time period there's also Atari Video Chess for the VCS (aka 2600). It had to fit in 4K of ROM and 128 bytes of RAM yet also had en passant, castling and prevented illegal moves.
I realize this museum is more about hardware but software that implements the game also deserves its place in history. To that effect, here is a writeup on Microchess made for the Kim-1 microcomputer:
Ah—my favorite is in there:
https://tluif.home.xs4all.nl/chescom/EngExcPhanF.html
It’s a “robotic” board that moves the pieces by itself.
You can sometimes find “untested” (i.e., broken) ones on eBay for a reasonable price, and if you’re lucky they’re an easy fix. Mine was stuck because the lock slider had wedged something and the repair took all of 10 minutes.
Very clean engineering: a few screws gets you in, there’s a remarkably small PCB, few wires and mechanical pieces: the main mechanism consists of two orthogonally mounted sliders with a stepper motor and belt each.
I don’t even play chess, but it’s amazing to watch it play both sides.
They also use a clever algorithm to route pieces around other piece since (obviously) the pieces can’t jump over other pieces given that they are moved by a magnet under the board.
Wow! Very cool. How did it manage to get a knight out as an opening move, for instance? Moving between the pieces, or moving other pieces out of the way?
One of the tricks I observed is that they move pieces half-way a square.
Oh, and (more often) they move pieces out of the way before the real move. First time you see that it looks like an illegal move but it’s all good because it gets moved back :)
Very cool! Wouldn't it be even cooler if the museum could score a couple of the very oldest machines? I'm talking about the El Ajedrecista machine (1912)[1] and Caissa [2][3] (named after the goddess of Chess[4]) built by Claude Shannon.
[1]. https://www.chessprogramming.org/El_Ajedrecista
[2]. https://www.computerhistory.org/chess/stl-430b9bbe92716/
[3]. https://mitmuseum.mit.edu/collections/object/2007.030.005?qu...
[4]. https://en.wikipedia.org/wiki/Ca%C3%AFssa
Technically, these electro mechanical machines may not qualify as computers, but still, what a scoop it would be to get them!
We had a Fidelity Chess Challenger 7 when I was a kid.
I was a horrible chess player but painstakingly worked out a way to win as white, keeping a detailed log of my experiments in a notebook. The first couple moves were wildly out of book (because I didn't know book), and the computer with its limited Z80 processor always computed the same moves after that. Some googling [1] shows the board's Elo is 1300ish.
To illustrate the state of the art in 1979, the manual [2] explicitly calls out that it understands en passant and castling.
[1] https://www.spacious-mind.com/html/chess_challenger_7.html
[2] https://ia902902.us.archive.org/20/items/mame0.211manualsful...
From the same time period there's also Atari Video Chess for the VCS (aka 2600). It had to fit in 4K of ROM and 128 bytes of RAM yet also had en passant, castling and prevented illegal moves.
It plays well enough that it beat ChatGPT 4o last year. https://en.wikipedia.org/wiki/Video_Chess
I love that the TEST PROGRAM at the end to verify the computer is working is to basically scholars mate yourself.
I'm so glad this exists!
I'm also surprised that I never bothered checking the make of my own chess computer from the 80s.
So I now have. It turns out I own a Novag! https://tluif.home.xs4all.nl/chescom/EngCCmk2.html
I realize this museum is more about hardware but software that implements the game also deserves its place in history. To that effect, here is a writeup on Microchess made for the Kim-1 microcomputer:
https://www.benlo.com/microchess/index.html