The efficiency of charge migration through stacked Watson-Crick base pairs is analyzed for coherent hole motion interrupted by localization on guanine (G) bases. Our analysis rests on recent experiments, which demonstrate the competition of hole hopping transitions between nearest neighbor G bases and a chemical reaction of the cation G(+) with water. In addition, it has been assumed that the presence of units with several adjacent stacked G bases on the same strand leads to the additional vibronic relaxation process (G(+)G...G) --> (GG...G)(+). The latter may also compete with the hole transfer from (G+G...G) to a single G site, depending on the relative positions of energy levels for G+ and (G+G...G). A hopping model is proposed to take the competition of these three rate steps into account. It is shown that the model includes two important limits. One corresponds to the situation where the charge relaxation inside a multiple guanine unit is faster than hopping. In this case hopping is terminated by several adjacent G bases located on the same strand, as has been observed for the GGG triple. In the opposite, slow relaxation limit the GG...G unit allows a hole to migrate further in accord with experiments on strand cleavage exploiting GG pairs. We demonstrate that for base pair sequences with only the GGG triple, the fast relaxation limit of our model yields practically the same sequence- and distance dependencies as measurements, without invoking adjustable parameters. For sequences with a certain number of repeating adenine:thymine pairs between neighboring G bases, our analysis predicts that the hole transfer efficiency varies in inverse proportion to the sequence length for short sequences, with change to slow exponential decay for longer sequences. Calculations performed within the slow relaxation limit enable us to specify parameters that provide a reasonable fit of our numerical results to the hole migration efficiency deduced from experiments with sequences containing GG pairs. The relation of the results obtained to other theoretical and experimental studies of charge transfer in DNA is discussed. We propose experiments to gain a deeper insight into complicated kinetics of charge-transfer hopping in DNA.