Although the time‐averaged shear‐wave velocity down to 30 m depth (VS30) can be a proxy for estimating earthquake ground‐motion amplification, significant controversy exists about its limitations when used as a single parameter for the prediction of amplification. To examine this question in absence of relevant strong‐motion records, we use a range of different methods to measure the shear‐wave velocity profiles and the resulting theoretical site amplification factors (AFs) for 30 sites in the Newcastle area, Australia, in a series of blind comparison studies. The multimethod approach used here combines past seismic cone penetrometer and spectral analysis of surface‐wave data, with newly acquired horizontal‐to‐vertical spectral ratio, passive‐source surface‐wave spatial autocorrelation (SPAC), refraction microtremor (ReMi), and multichannel analysis of surface‐wave data. The various measurement techniques predicted a range of different AFs. The SPAC and ReMi techniques have the smallest overall deviation from the median AF for the majority of sites. We show that VS30 can be related to spectral response above a period T of 0.5 s but not necessarily with the maximum amplification according to the modeling done based on the measured shear‐wave velocity profiles. Both VS30 and AF values are influenced by the velocity ratio between bedrock and overlying sediments and the presence of surficial thin low‐velocity layers (<2 m thick and <150 m/s), but the velocity ratio is what mostly affects the AF. At 0.2<T<0.4 s, the AFs are largely controlled by the surficial geology of a particular site. AF maxima are the highest in the hard classes, which is the inverse of the findings used in the Australian Building Code. Only for T>0.5 s do the amplification curves consistently show higher values for soft site classes and lower for hard classes.