511 Week 3

7
Games and Game Complexity Alan Nochenson IST 511 Sept 8, 2012

Transcript of 511 Week 3

Page 1: 511 Week 3

Games and Game

ComplexityAlan Nochenson

IST 511

Sept 8, 2012

Page 2: 511 Week 3

Combinatorial game theory Sequential games

Perfect information

Win, lose, or draw (ordinal utility)

Examples include chess, checkers, tic tac toe, nim

Page 3: 511 Week 3

State-space complexity The number of legal

positions possible in a

game

For tic-tac-toe, it is 765

(removing rotations)

For chess, it is ~10^43

Page 4: 511 Week 3

Game tree size The number of leaf nodes

in the tree representation

(end-states reached in

different ways are counted

twice)

26, 830 in tic-tac-toe

(removing rotations)

Page 5: 511 Week 3

Decision complexity How “hard” it is to decide

on a given move, or how

many moves deep in the

tree you need to look to be

sure of the outcome

Page 6: 511 Week 3

Game tree and Computational

complexity Game tree complexity

The number of leaf nodes of a full-width game (equivalently how

many leaf nodes are in the level you need to reach to get a

minimax score)

GTC = b^d where b is the average branching factor, and d is the

game tree depth (in plys)

Computational complexity

Asymptotic difficulty of a game (big-O notation)

Play a generalized version of the game (n x n)

Page 7: 511 Week 3

Works Cited General information from http://en.wikipedia.org/wiki/Game_complexity

State-space complexity:

Facts from http://www.methodshop.com/games/play/tictactoe/index.shtml

Picture from http://www.gameideasforkids.com/images/tictactoe.JPG

Game tree size picture from http://scienceblogs.com/goodmath/wp-content/blogs.dir/476/files/2012/04/i-30601eb10fe21a4ce5a4f2f92e80eb10-tic-tac-toe.png

Decision complexity

picture from http://www.cs.berkeley.edu/~ddgarcia/teaching/CS3Gamesman/assignment/tttbranch.gif

Fact from http://www.nature.com/news/2007/070716/full/news070716-13.html