Regarding throretical versus real-world reflection coefficient curves

Could some electronics guru shed some light on this ? The curve of the absolute value of the reflection coefficient vs. frequency(derived from a Smith chart assuming lossless transmission line) for a simple circuit consisting of a s ignal generator driving a complex load through a transmission line and comp lex source impedance, looks exactly as expected, with the reflection coefficient going to zero at the operating frequency. OTOH, A simple SPICE simulation of the same circuit using the lo ssy transmission line model, generates a reflection coefficient curve that has a similar shape as the one obtained using the lossless transmission lin e model, but in this case the minimum value of reflection coefficient is not close to zero, and there is a large offset So how are these differences reconciled in ractice, given that real world devices are non ideal.