discussion needed: match network from antenna to switch

I'm trying to build a match network from an antenna to a switch whose impedance is 50Ohm. To realize that, I first got the return loss values of the antenna in the whole frequency band, namely from 800M to 900MHz. Then I import the data into a port in ANSOFT designer. Then, I choose the center frequency of the span to make a match by the Smith Chart. The match network thus conducted is added to between the measurement port of 50 Ohm, which simulate the switch, and the antenna port, which is simulated by its RL values.

When I simulate the new circuit with match network and draw out the RL curve , it is found the the frequency point at which I have done the match is not at the valley of the curve. Does anyone come across the similar phenomenon as mine?

Can someone tell me if there is anything wrong in my way of simulating?Or can anyone provide a good method to simulate a match circuit?

You have to take also in account the parasitics of matching network. These will make your design to be a little off.