Modelling market sellers in World of Warcraft

The virtual economy operates like the one in the physical world, which is realized by players with two functions–demand and supply, only using the form as avatars. Only that in virtual worlds, the prices of virtual items is decided by the players instead of by its quality or scarcity, so the player behavior analyses became an important tool to monitor the persistent virtual worlds. In order to achieve the idea of a more balanced virtual market, we analyze the trading behaviors in three points – sellers, prices and items, and expect to figure out the selling motivation, pricing rules and the over/underpricing conditions. In this paper, we use the auction house data in World of Warcraft (WoW) to analyze these seller behaviors. After applying clustering and classification, a market seller model of WoW is presented.