The international general game playing competition

General game players are systems able to play strategygames based solely on formal game descriptions sup-plied at run time. (In other words, they don’t know the rules until the game starts.) Unlike specialized game players, such as Deep Blue (Campbell, Hoane, and Hsu 2002), general game players cannot rely on algorithms designed in advance for specific games; they must discover such algorithms themselves. General game playing expertise depends on intelligence on the part of the game player rather than intelligence of the pro-grammer of the game player. General game playing (GGP) is in many ways similar to autonomous planning. Domain-independent problem solving is at the core of both. The description of a game in the game description language (GDL) (Love, Hinrichs, and Genesereth 2006) is similar to that in the languages used by planners (such as PDDL); and the overall goal is the same — to achieve a state with specified properties. One obvious difference is that, in GGP, there are opponents, which complicates the process of determining an ideal course of action. Another difference is that, in GGP, there is an execution environment, making it pos-sible for a game player to interleave planning and execution. Also, in GGP, there are time constraints, which make it essential for players to act even when they are unsure which courses of action are best. In the last two respects GGP is more similar to reactive than classical planning, in that it has to commit to