[en] The distribution of intensities of KLⁿ K_α-satellites is nearly binomial, with parameter p_L, the mean L-shell vacancy probability per electron. The chemical environment of an atom produces a shift, δp_L of p_L from its value for an isolated target atom. δp_L tends to increase as p_L increases, so for greatest chemical sensitivity one wants p_L as large as possible. p_L increases with Z_p but, because it must remain <1, p_L must saturate for large Z_p. Thus, little additional sensitivity is gained, for a given target and impact speed, by using heavy-ion projectiles of charge greater than some moderate value. We have made theoretical calculations of p_L to provide information on the rate of this saturation. Whereas all earlier calculations have employed a single-particle model and a simple collision approximation, we use the Hartree-Fock independent Fermi particle model, which contains Pauli correlations, and refined coupled-channels collision theory. Because of a tendency toward random phases in the scattering amplitudes, the intensity distribution is nearly binomial in most cases, but has the possibility of deviating strongly when one or two channels are dominant. We show, as a typical example, p_L as a function of Z_p and E/A for isolated Ar targets, where the tendency toward saturation appears for F⁺₉ and even for C⁺₆