The NA-UBS is a critical text including ALL known manuscripts (at the time of publication) as well as the frequency and age of various renderings.
The TR is in a sense an early attempt at a critical text, a combination of several separate texts in an effort to "ferret out" the most likely text. As a primitive critical text, it cannot compare in authenticity to the much more complete and thorough scholarship, source and evaluation that has gone into the NA-UBS.
The BMT is similar in that it is based on the majority of texts, but does not take into consideration the age of the texts used nor include all source texts available. It is, by definition, based on a biased selection of specific texts with irrational exclusion of some of the oldest (and, therefore, presumably closest to the original) texts.
The NA-UBS is clearly the most reliable from the scholarly standpoint: it evaluates ALL the available source texts using scholarly techniques (rather than sectarian bias) and derives a reading based on those scholarly evaluations with the most common alternate renderings plainly identified - and even sort of "graded" so that the translator can determine the value of those alternate renderings in comparison to the primary rendering.
From a purely scholarly perspective, there is little merit in relying on either of the other two sources when such a superlative source is available, except in cases where the intent is to translate one of those other two sources specifically to illustrate the differences between them and the most reliable text.
Jim, http://www.bible-reviews.com/