In the current study we examine whether, in a virtual environment, self-controlling the difficulty level of a game will result in worse performance compared to a fixed difficulty level. Our hypothesis is evaluated using a novel VR game. Forty-five participants were randomly assigned to two between-participants groups: the Experimental group, in which the participant determined the desired level of difficulty in all five game sets by themselves, and the Controlled group, in which the level of difficulty was set to medium for all participants in all game sets. The results demonstrate that the mean difficulty level selected by the Experimental group was lower than medium in the first two sets, and higher than medium in the last three sets. In the last two sets, an earlier termination of the game and lower scores were demonstrated for the Experimental group compared to the Controlled group. The conclusions are that the participants in an unfamiliar virtual environment might not be able to select their difficulty levels to match the task at hand and might not optimize their learning zones.