Input impedance matching on the ADRV9009 at low frequencies

I would like to optimise the input impedance matching of my circuitry to the ADRV9009. I do not have access to simulation software so I have to do this through experimentation and analysis of the input impedance Smith Chart. Below is the ADRV9009 RX input impedance Smith chart taken from the datasheet:

If I look at this Smith chart, I see that at most points the Impedance magnitude is close to 100R (certainly at frequencies below 1GHz). However, in multiple locations on the EngineerZone forum I have seen instructions to treat the RX input impedance as 50R. Why is that? It doesn't tie up with what is detailed on the Smith Chart.

This is further proven by the fact that the ADRV9009 evaluation board has a 1:1 balun populated. Surely if the input impedance is close to 100R, a 2:1 balun would be better for impedance matching and would provide some zero noise voltage gain?

Please advise as to the reason behind treating the RX input impedance as 50R when, as far as I can tell, this doesn't tie up with the Smith chart. Unless I am reading the Smith Chart incorrectly.

Thank you.