Is it more important for philosophers to argue well than it is to write well? Posed this way, the question sets up a false dichotomy for you cannot argue well without writing well. Logic is not identical with rhetoric, but the logical form of an argument cannot be neatly drawn apart from its rhetorical component. (Classical rhetoric has been insisting forever that we cannot separate form and content.) We define validity and soundness of an argument in formal semantic and syntactical terms; and unsurprisingly, those notions find their greatest traction when evaluating arguments expressed in formal languages. But philosophical disputation takes place using natural languages; and arguments are made in order to persuade or convince or induce other changes in the epistemic make-up of our interlocutors.
We argue with someone, somewhere, in some time and context; we argue to achieve some end, whether moral, political, economic, legal. Any evaluation of the arguments we make must take these factors into consideration; without them at hand, our evaluations are sterile and pointless. (Why, after all, do we concern ourselves with notions of epistemic justice if not for the fact that some arguments are more likely to be ‘heard’ than others?) Fallacies abound in natural language arguments; correcting them is not just a matter of paying attention to the abstract logical form of the argument ‘underlying’ the sentences we have deployed; it is a matter too, or making sure we have chosen the right words, and deployed them appropriately in the correct context. To use an example from an older post, we reject a smoker’s argument that we should stop smoking on ad-hominem grounds, but the smoker really should have known better than to try to convince someone to quit while puffing away merrily and seemingly enjoying deep lungfuls of smoke. Good argument; terrible form. The same smoker would find a more receptive audience if he spoke with some feeling about how miserable his health has become over the years thanks to his smoking habit.
(On a related note, consider that when programmers evaluate ‘good code,’ they do so on the basis of not just the effective functionality of the code in accomplishing its task, which is a purely technical notion, but also on aesthetic notions: Is the code readable? Can it be modified easily? Is it ‘beautiful’? No programmer of any worth elides these notions in evaluative assessment of written code.)
There is a larger issue at play here. Philosophers do much more than just argue; sometimes they just point in a particular direction, or make us notice something that we had not seen before, or sometimes they clothe the world in a different form. These activities have little to do with arguing ‘correctly.’ They do, however, have a great deal to do with effective communication. Writing is one such form, so is speaking.
Note: The examples of great philosophers who are considered ‘terrible’ or ‘obscure’ writers–by some folks–does not diminish the point made here. Hegel and Heidegger–with due apologies to Hegel-and-Heidegger-philes–achieved their fame not just because of the quality or depth of the arguments they offered in their works but also because they wrote from particular locations, in particular times. (Some think they made terrible arguments, of course!) The sociology of philosophy has a great deal to say about these matters; more philosophers should pay attention to it.