The human visual system can only attend to a limited number of regions within an interface at one time. Several methods for predicting the deployment of attention have been suggested. We describe an interest point method in which participants identify five “interesting” points in a display. This method successfully predicts attentional selection in scenes and websites as reflected by eye movement data. First, we developed a tool that allows interaction designers to rapidly employ this method for interface testing. Second, we investigated the usefulness of combining the traditional squint test (by blurring the image) with the interest point method. Results indicated that blurring an image disrupts the typical deployment of attention in an interface, therefore squinting or blurring cannot be used as a proxy for interest points.