The impact of guessing auxiliary population attributes, as opposed to relying on actual values from a prior survey, was quantified for three unequal probability sampling methods of tree stem volume (biomass). Reasonable prior guesses (no-list sampling) yielded, in five populations and 35 combinations of population size and sample size, results at par with sampling with known auxiliary predictors (list sampling). Realized sample sizes were slightly inflated in no-list sampling with probability proportional to predictions ( PPP ). Mean absolute differences from true totals and root mean square errors in no-list-sampling schemes were only slightly above those achieved with list sampling. Stratified sampling generally outperformed PPP and systematic sampling, yet the latter is recommended due to consistency between observed and expected mean square errors and overall robustness against a systematic bias in no-list settings.