In this paper, we evaluate the performance and experience differences between direct touch and mouse input on horizontal and vertical surfaces using a simple application and several validated scales. We find that, not only are both speed and accuracy improved when using the multi-touch display over a mouse, but that participants were happier and more engaged. They also felt more competent, in control, related to other people, and immersed. Surprisingly, these results cannot be explained by the intuitiveness of the controller, and the benefits of touch did not come at the expense of perceived workload. Our work shows the added value of considering experience in addition to traditional measures of performance, and demonstrates an effective and efficient method for gathering experience during interaction with surface applications. We conclude by discussing how an understanding of this experience can help in designing touch applications.