With the recent success of tablet devices a new device type became available for mobile interaction. Just as for mobile phones, touch is the dominant way people interact with tablets. In contrast to the much smaller phones a firm grip with both hands is needed to securely hold tablet devices. While a large body of work has investigated touch interaction on smaller devices, is little empirical research has been carried out on touch-based pointing while holding the device with both hands. To understand touch-based interactions using tablet devices, we conducted an experiment to compare four pointing techniques on both the front and back of the devices while it was held in landscape format. We compare direct touch with the following alternatives for selecting targets, indirect pointing on a virtual touchpad, an inverse cursor, and a miniature interaction area. While direct touch is 35% faster than the fastest alternative, only 74% of the touchscreen and 64% of a back-of-device can be reached by each hand. We show that among the indirect pointing techniques, the miniaturized interaction area is significantly faster and received the best subjective ratings. We conclude that a miniaturized interaction area is a viable alternative to direct touch especially on the backside of tablet devices.