In this seminar, researchers from the GRIP LAB at the University of Michigan will present ongoing work on “Anticipating Student Work,” a virtual teaching simulation designed to support mathematics teachers in selecting student work for classroom discussions in problem-based instruction. The simulation’s context is a geometry lesson, in which a problem about a circle is used to teach a theorem about tangents to a circle through an exterior point. In teaching through problems, teachers need to be able to notice features of student work that may support the class’s work toward the goal of the lesson. The simulation includes activities in which participants sort, select, and sequence student work that provide teachers with opportunities to learn what kinds of ideas students bring to the problem and consider how those ideas might support the classroom discussion. We will discuss the design of the simulation and share how we have been tracking participants’ learning as they complete this simulation.