Technically speaking, in order to have a fully redundant system from a reliability standpoint, we should have two completely independent control trains. In other words, two batteries, two switches, two recievers, and they should be from different manufacturers/designs. That is to further reduce the likelihood of a "common mode failure" experience by two identical components in separate trains.

Since we don't have any real probability numbers, or failure rate data for any of our components, we are left with qualitative statements about "the most common source of failure". Some believe LiPo batteries are the most likely failure. Others think switches. Still others think servo connections. Bottom line is that for the basic trainer setup, there are enough of them flying with just a standard setup (one battery, one switch, one reciever) to demonstrate they are reliable, and good preflight checks will reveal nearly any failure in the system (regular voltage checks, known battery discharge curve, range checks, routine assembly procedure, inspection).

The most common failue IMHO is the loose nut holding onto the transmitter.

Brad