Theoretically a translation should not be clearer or more explicit than the original. But this is easier said than done, especially when dealing with ancient texts: often the cultural background of implicit communication is lost to to the modern reader, so it must be explained -- assuming that it is not lost to the translator too! That's why there will always be a healthy separation between commentary and translation, even though the exact line of separation is endlessly debatable.
However there is more to overtranslation than explicitation. Overtranslation is a frequent amateur's mistake, reading more into words than their actual meaning. Often it involves the confusion of (correct or mistaken) etymology with semantics. To take one example which is not specific to the NWT (it is actually Trinitarian in origin), translating monogenès by "only-begotten" when the etymology is actually "one of its/his/her kind" and "only" sufficiently renders the meaning.
An additional problem in Bible translation is intertextuality. "Worship" for proskuneô may be contextually correct in some cases (e.g. Matthew 4), but not all (definitely not when total strangers meet Jesus for the first time, at the simplest narrative level), yet the translator cannot avoid the issue of consistency, allowing for a kind of double entendre (in the natural act of bowing down to an important man, the Christian reads more than a common gesture of respect, just as he hears more than physical healing in the phrase "your faiith has saved you"). In the narrowest context of John 8:58 "I have been" may sound as a perfectly correct translation of egô eimi, but in the largest context of the Fourth Gospel which repeatedly uses egô eimi as a Christological formula a locally awkward "I am" is the formal price to pay for the overall consistency which is theologically more important. In such cases it seems obvious that the NWT deliberately misses the forest for the trees.
From other threads here I have gathered that BeDuhn disagrees with the NWT on the introduction of "Jehovah" into the NT... I don't know if he deals with other issues that I think important, such as the rendering of the non-commutative en by the commutative "in union with," which reduces the symmetrical expressions like "you in me and me in you" to mere tautologies...