In fact, the first two factors are equivalent: 10 3\-messages are equivalent to one 310 message, and in general, $$n$ $k$\-messages are equivalent to one $n^k$\-message\. If the individual n\-messages are dependent on each other, then different nk messages have different likelihoods: For example, if message 3 never follows message 2, then in the combined message, "32" never appears as a substring\.
Is what follows the colon meant to be justification for what precedes it? I'm not following.