A method described evaluates the pilot-to-data power ratio based on the values of the demodulated data at the output of a receiver such as a conventional maximal ratio combiner or an equalizer receiver. Since the data are not scaled to their particular decision grid, their distribution and statistics deviate from the ones defined by the known decision grid. Measuring these deviations of the statistics, the pilot-to-data power ratio and the appropriate demodulated data scaling can be accurately determined. Subsequent data processing may then commence.