Peer-to-peer media streaming is a very attractive approach since the bandwidth available to serve content scales with demand. A challenge key, however, is making content distribution robust to peer transience. Multiple description coding (MDC) has proven to be very effective in problems related to packet loss, since it generates several descriptions that allow the reconstruction of the original information with any number of descriptions that can reach the decoder. In this way, multiple description may be effective for robust peer-to-peer streaming. This paper investigates the influence of MDC on the robustness of peer-to-peer video streaming. It is shown that not only MDC can really be effective for robust peer-to-peer video streaming, but also that varying the redundancy level of description on the fly lead to a better performance regarding an existent system where this parameter is fix.