When chess master Garry Kasparov was defeated by IBM computer Deep Blue in 1997, AI programmer Omar Syed felt that it was not fair. He wanted a simple board game in which humans were at a clear advantage against computers. He also wanted to play it with his four year old son Aamir. Thus arimaa was born. It can be played on a standard chess board. Normal chess pieces of pawn, bishop, rook, knight, queen and king are replaced by or renamed as rabbits, cats, dogs, horses, camel and an elephant. This was unquestionably very entertaining for little Aamir (Infarct, arimaa is Aamir spelled backwards with an additional ‘a’ at the start).
Basic concept of the game is quite simple. The aim is to move rabbits to the end line of the board or to capture all enemy rabbits. But complexity of the game comes from the fact that its starting positions are not fixed. There are estimated 64 million ways to open a game. Similarly, the average possible moves are 500 times higher than that of a chess game. This branching factor makes it very difficult for computers. Since 2004, there is an Arimaa challenge in which human and computer programs compete for a 10,000 dollars prize. So far, humans have won the championship every single year. But it is predicted that by the year 2020, computers will have enough processing power to start beating human players. What will happen then? It is possible that another artificial intelligence expert would come up with something new to test the wits of machines. The standoff might continue into the future.
One might say that this is actually not a battle between human and computers at all. In reality, this is just one kind of human intelligence versus a different sort of human brainpower. A chess or arimaa mastermind is competing against carefully crafted human algorithms that are being run on human designed circuits. That is certainly a valid point, until humans give enough logic and processing power to computer that they become conscious.
Image credit: http://www.arimaa.com
