Theories of motor imagery conflict in their account of what happens during an imagined movement, with some suggesting that movement is simulated while others suggest it involves creating and elaborating upon an internal representation of the movement. Here we report evidence that imagery involves the simulation of a movement and that it varies in accuracy. Two groups of participants performed a motor task focused on challenging movement execution either overtly or via motor imagery. Overt performance was used to model expected performance given required movement characteristics (i.e., speed, complexity, familiarity), which was then compared with self-reported accuracy during imagery. Movement characteristics had a large effect on self-reported accuracy compared with a small effect of imagery vividness. Self-reported accuracy improved across trials with familiar movements compared with novel movements in a similar manner for each group. The complexity of the imagined movement did not influence movement time during imagery or overt trials, further suggesting that imagined movements are simulated rather than abstractly represented. Our results therefore support models of motor imagery that involve the simulation of a movement and its viability, which may be the basis of imagery-based motor learning. (PsycInfo Database Record (c) 2022 APA, all rights reserved).