Aside from being good at English, I am also obsessed with math. I recently created a draft on how to create a “computer” to beat Chess by learning from multiple games from the opponent.
Firstly, a “computer” does not need to be filled with technological pieces to, by definition, a computer. The technical definition of a computer is something that processes inputted data and outputs the, well output. In this case the inputted data is the opponents move, processes it by randomly picking a random possible move and out puts it. This setup can be done with a actual computer or by hand. First we need a lot of containers, like plastic bags or matchboxes. With pictures of the turn move and the layout of the board. It must also show the possible moves the “computer” can do marked by colored arrows. In the container there should be some indicator of color that is randomly taken out correcponding to the current board layout. When that color is chosen, move the piece the color corresponds to to the place where that colored arrow points to. Every time the “computer” loses, take out the previous color indicator which made it lose. When every color indicator for a chess layout is taken out, also take out another color indicator, the one which got the piece in the losing area in the first place. This way the “computer” learns the more it is played against.
You will need to know every possible chess move and draw that layout on the containers.
This setup was inspired by a similar game called hexapawn, which is a more simple way of playing chess. The idea is that you play on a 3 by 3 tiles area with 3 pawns and the enemy with 3 pawns as well. To win the game you are supposed to either, get a stalemate, kill every piece or get one of the pawns to the other side. Making a “computer” set up to win this would be easier because there are less possible moves.
Amateur writer.com 