Disclosed herein are systems and methods for controlling display objects. Particularly, a body part of a user may move, and the movement detected by a capture device. The capture device may capture images or frames of the body part at different times. Based on the captured frames, velocities of the body part may be determined or at least estimated at the different times. A blend velocity for the body part may be determined based on the different velocities. Particularly, for example, the blend velocity may be an average of the velocities of the body part over a period of time. A display object may then be controlled or moved in accordance with the blend velocity. For example, an avatar's body part may be moved in the same direction as a recent captured frame of the user's body part, and at the blend velocity.