Static Evaluation Function for Checkers

Question

I am trying to write an evaluation function for a game of checkers that I am developing, but I can not find the documentation I need. I have read several documents on the Internet that describe various methods for writing or resolving a computer (using genetic algorithms or Bayesian learning), but they are too complicated for beginners like me. All documents contain a link to "Some research on computer training using the game of drafts" from ALSamuel, but I still could not understand it :( I just read the following: "Some research on computer training using the game of drafts -II" and found there is some good information, but it does not explain what the eval parameters mean (I think I do not have the whole article).

Answer 1

I would start with something dead simple: material difference. Which is equal to: (# of my checkers on board) - (# of checkers of opponents on board). Then you can add additional functions and start weighing them, for example, “the number of checked drafts”, “# protected drafts” or, possibly, “the number of squares controlled in the middle of the board”, etc. Talk to a domain expert (i.e., the Drafts Player) and / or refer to the drafts book to find out what will work well.

Answer 2

AI , , , ( , )

" , , , , , ( , ) ". - Wikipedia

Ai at https://en.wikipedia.org/wiki/Chinook_(draughts_player)