The major problem with tuned RF design is you don't know where the optimize point is for minimum reflection and maximum power transfer. You have 3 variables, L, R and C and you have more than one optimized point depends on what is the other requirements. Smith Chart show you in graph every single point so you can choose the RLC values at the same time. It can get very complicated to vary 3 values in LTSpice.
In another words, LTSpice is good if you know the values and see what is the outcome. But it is not good if you are searching for the optimize values of 3 components or more.
Now, for wide band circuits, Smith Chart is useless. I think that's where LTSpice shines.
Bottom line, I think it's the "iteration" thing that's where Smith Chart simulation is strong.
