Many statistical analyses of the cosmic velocity held are based on the gala
xy number-weighted sampling and therefore cannot escape from the bias effec
t. We study this problem in the framework of the stochastic bias mechanism
using a simplified model and the results of a recent hydrodynamical simulat
ion. We adopt the velocity correlation functions as representative examples
and show that velocity statistics might be qualitatively affected by the p
oorly understood bias mechanism and strongly depend on the tracers of the c
osmic velocity field.