I am willing to include the channel matrix in my analytical model (which only includes pathloss and antenna gains). But from what I have seen, channel matrix can't be used to produce numerical results in an analytical model because it relies and should be based on real device tests.
Is my way of thinking wrong in this problem?