All the atoms in a molecule attract each other until they reach their equilibrium positions at which point the repulsive forces between the atoms just balance the attractive forces and there are no resultant forces acting on any of the atoms in the molecule. Thus, we can consider that in the equilibrium geometry the atoms in a molecule are arranged as compactly as possible. This is the basis of the ligand close packing (LCP) model according to which three or four monatomic ligands X, such as F, Cl or O (formally =O or O(-)) pack as closely as possible around a small central atom such as a boron or carbon atom giving a truly close-packed equilateral triangular AX(3) molecule or a tetrahedral AX(4) molecules. Such monatomic ligands can, to a good approximation, be described as having a spherical shape with a single ligand radius r(X). In contrast, ligands with donor atoms with lone pairs such as the oxygen atom in an OX group have a less symmetrical electron density requiring two ligand radii, r(O(lp)) in the lone pair direction, and r(O(b)) in the bonding direction, where r(O(lp)) < r(O(b)) for an approximate description. On this basis we propose an explanation for the "methyl tilt", in methanol and many related molecules, and in conjunction with the concept of compactness, a model for explaining the relative energies of the conformations of molecules containing OH and OMe ligands, including molecules that exhibit the anomeric effect. We compare our model for the anomeric effect with the widely accepted "hyperconjugation" model. We also discuss the relationship between the concept of compactness and the concept of hardness.