We study the behaviour of some tiling games by computing the average score a player can obtain, as a function of the probability to make a mistake. The idea originates from the well-known Tetris game and the problem is approached by using a method similar to the one presented in [MSV00], where it is described an algorithm to determine the number of ways a strip, fixed and can be tiled with some given set of pieces. In that paper it is proved that the set of possible tilings can be generated by a regular grammar. From the grammar it is routine to find the rational generating function counting the number of tilings of a strip with pieces. Here we show how similar arguments can be used to study tiling games having rules analogous to Tetris.