Historically, the design of antenna arrays has evolved separately from Shannon theory. Shannon theory adopts a probabilistic approach in the design of communication systems, while antenna design approaches have relied on the deterministic Maxwell theory alone. In this paper, we investigate an information-theoretic analysis approach which we apply to evaluate the design of a dual-band, dual-polarized multiple-input multiple-output (MIMO) array on a cellphone. To this end, we use ANSYS HFSS, a commercial electromagnetic (EM) simulation software suitable for the numerical optimization of antenna systems. HFSS is used to obtain an accurate model of the cellphone MIMO antenna array and HFSS SBR+ is utilized to obtain channel matrices for a large number of users. Taking advantage of linear and optimal processing at the cellphone, we estimate the outage probability curves. The curves are then used to determine the diversity gain in a moderate signal-to-noise ratio (SNR) regime and the multiplexing gain at a high SNR regime. This approach is then compared with the method of estimating the diversity gain from the envelope correlation coefficients or the beam-coupling matrix showing substantial differences in the two methodologies.