I did not think that matching depended on amplitude as the purpose is to provide optimum energy transfer... can you elaborate?
You're using the circuit as rectifier to detect a 433 MHz signal.
Have a look at the datasheet of the diode you're using (PDF, and attached PNG, package parasitic effects are omitted) where you also can see the AC circuit model of a Schottky diode. The datasheet actually calls it "equivalent linear circuit model" but really, the junction capacitance is also a function of junction voltage (so matching depends on the junction voltage and other stuff like the SOT package parasitics).
The current through Rj is voltage dependent Id(Vj)=Is*(e^(α*Vj)-1). (no latex addon?)
Further, the diode junction capacitance (varactor) is Cj(Vj)=C0/((1-Vj/Vo)^gamma)
This is what the spice model is doing for you (compare the values from the datasheet with the spice model):
(HSMS285X.PNG)
*SPICE model for HSMS-285x
*The parameters are for a single diode (HSMS-2850). Parameters also apply
*to the individual diodes within multiple diode configurations.
*
.SUBCKT hsms 1 2
DCD1 1 2 DMOD
.MODEL DMOD D(IS=3E-6, CJO=0.18E-12, VJ=.35, BV=3.8, IBV=3E-4
+ EG=0.69, N=1.06, RS=25, XTI=2, M=0.5)
.ENDS
But you also need to simulate the package parasitics and the layout will also affect the matching.
I don't have the time right now to do redo an old design for 433MHz but here's what the dependency looks like for a different design frequency. You can clearly see how the optimum detection frequency changes with different input power:
Harmonic balance is necessary in order to simulate the rectified voltage (and all the harmonics) correctly. It may not make too much of a difference but since you're interested in very low powers I would say it is good to include harmonic balance in addition to just the linear solver. Especially when you want to plot voltage over the load resistor vs. input power to the detector.