Creating quality artificial intelligence for computer games is now a real problem. Only one sound and graphics, even though they are very, very best, can not guarantee the ultimate success computer games. Gameplay – that's what collects huge crowds of fans, and that makes the game a true masterpiece. That's why, today, the attention of developers are increasingly focused on improvement of artificial intelligence in computer games. Probably, the penetration of the problem and led to the tournament BotPrize. BotPrize – an annual three-month competition among computer "bots" for Unreal Tournament 2004.
The behavior of the game character controlled "bot" is estimated by the expert group. More precisely experts are trying to determine who controls the character of a person or artificial intelligence. At the heart of BotPrize is one of the modifications, so called "Turing test". Turing Test, as its name implies, was proposed by Alan Turing in 1950, to determine whether the machine intelligence. The classic interpretation of the Turing test is as follows way. Man-Expert communicates through a keyboard with two companions, one man and one computer.
It is believed that the computer will be a test for intelligence, if within a specified time expert can not clearly define who is a real person of the interlocutors, and who is machine. According to Associate Professor School of Computer and Information Science at Philip Hingston Australia, the tournament BotPrize is of great importance for artificial intelligence in games. BotPrize shows that computer games can be made much more interesting and enjoyable, if a qualitative model of artificial intelligence. In 2009, attended by fifteen BotPrize applicants from around the world: from Japan to Brazil. The choice of game to create "bots" fell on the Unreal Tournament 2004. For the "purity of judging," every means of communication among the players deactivated. This allows you to evaluate the intelligence of "bots" only based on their behavior. Each judge at the same time meets the two rivals. He does not know how the character manages a real person, and how artificial intelligence. At the expiration of ten to fifteen minutes game expert you must determine which of the characters is a "bot". If the program to control the bot can fool at least 80% of the judges, the creator receives a grand prize of $ 6,000. In BotPrize, held in 2008, do it no one has. Then the creator of the "human-like" artificial intelligence gave the founders of the tournament consolation prize of $ 1,700. In conclusion it should be noted that attempts to draw attention to the problem create a quality game AI does not end the tournament BotPrize. In August 2009 a group of scientists in the field of Computer Science has announced a competition to win the game Super Mario Bros programs Artificial Intelligence. Scientists do not accidentally select this version of Mario. The fact is that in this version of the game, each level is randomly generated, and the number of such levels can be infinite. The founders of the tournament promised the winner that development, which will be more on all levels. True, the prize of this contest is far less than BotPrize and is only $ 500.