I need help understanding, in that trend specifically, as to when the length factor dominates, and when the electronegativity factor does?
There is nothing much to understand here. All you need to understand is that this is basically an experimental observation.
We are trying to use qualitative arguments (higher bond length/more electronegativity) to explain the actually calculated values. When we observed a contradiction in our answer for fluoromethane and chloromethane, we corrected our reasoning, by saying that the bond length dominated over the electronegativity difference in this case. And that is true, because this correction is based on the calculated values only.
So, we first try to predict the order by our own qualitative arguments (logically reasoned, of course). If they work, great! If they don't, we look into the factors that we originally deemed to be minor instead. We calculate their values, and see what effect they had. Then, we justify those minor factors by saying that they dominated over the other in the particular case.
Another such interesting case is the order of the basicity of ammonia derivatives in water. It is $\ce{NEt2>NEt3>NEt1}$ (for the ethyl derivatives) and $\ce{NMe2>NMe1>NMe3}$ (for the methyl derivatives). Here, three factors are operating simultaneously! The solvation effect, the steric hindrance of alkyl groups, and the positive inductive effect of alkyl groups - all either reinforcing or hindering each other. In such edge cases, it becomes extremely tough to predict any order solely based on these factors, and hence, computated values are much preferred.
If you prefer to read the theory of those experiments (be warned it's scary), here is one such paper discussing the detailed calculation of the dipole moments of various fluoromethane derivatives.
