I have spent long enough considering Bible translations in different languages now that I have been able to articulate the criteria by which I evaluate a translation. These are, of course, subjective to my reasons for reading the Bible — as a cultural influence and a (somewhat biased) historical document, without any particular religious affiliation.
My criteria are, in descending order of importance:
1. It must include the Deuterocanonical books.
Aside from the Jewish Tanakh, no Christian denomination — whether Catholic, Orthodox or Protestant — excluded these books until around the 19th century. Martin Luther himself asserted that they should be read, but not considered to be scripture in the same way that the other books are. Thus, they were placed in their own section of the Luther Bible, rather than among the Old Testament books where Catholics prefer them. The refusal by many Protestants to read them at all is a modern invention. In any case, since they were universally treated as scripture prior to the Reformation, they are just as important to the cultural influence of Christianity as the rest of the Bible.
Of the Bibles I own, the Good News Bible, the New Revised Standard Version and the NBV21 all include the Catholic Deuterocanonical books separately from the Old Testament, which is ideal, as it makes the separation of sources clear to the reader. They vary in the order of books within this section and the inclusion of additional books not considered canonical by the Latin Church — they all include the Prayer of Manasseh, but beyond that there is considerable variation.
An Bíobla Naofa, on the other hand, includes only the Catholic canon in the Vulgate order. This is not surprising, since — unlike the ecumenical translations mentioned above — this is a translation commissioned by the Catholic Church. This approach somewhat masks the distinction between Hebrew scripture and the later Greek additions, especially in books like Esther and Daniel with extensive insertions of entire passages, but is acceptable.
Many Protestant translations fail this test and are immediately eliminated. The New International Version is a notable such failure.
2. It should be as faithful to the original as possible, according to scholarly consensus based on available manuscripts.
This is nebulous in some cases — for example, is 1 Corinthians 7:36 about a betrothed couple, or a father and his daughter? Both are valid scholarly interpretations of the Greek, so neither one makes a translation any more or less acceptable based on this test. But there are translations that ignore scholarly consensus in favour of traditional (mis)translations, particularly where there are valid interpretations of the Greek Old Testament which do not reflect the original Hebrew.
My go-to example for this is Isaiah 7:14. If that verse contains the word "virgin", it is not an accurate translation of the Hebrew. It may be an accurate translation of the Greek, but every good translator knows you don't translate a translation if you can help it. The New International Version fails this test too, as does the King James Version. (The latter case is forgivable because modern Biblical scholarship did not yet exist in the 17th century; what is not forgivable is that such an outdated translation is still so widely relied upon.)
All four of the Bibles I own pass this test with flying colours. It is an issue I am not willing to compromise on.
3. It should be written in clear, straightforward, idiomatic and modern language, in whichever language it is translated into.
This is obviously quite subjective. Regular church-goers will have a much easier time understanding Christian jargon than the irreligious. My personal view is that while translations that try to render every word of the original Hebrew and Greek (formal equivalence) have their place as study Bibles, a good general-purpose translation is one that uses natural language to convey the same ideas that the source texts would have conjured up in early Christians (dynamic equivalence).
In practice, this is a spectrum and every translation does this to a greater or lesser degree. (Simply translating each word by itself would result in incomprehensible gibberish, while some concepts cannot be rendered into natural English without needing some background explanation — the archetypal example being the tetragrammaton.) As such, it is more of a way to choose between two translations than a test that can be applied to a single translation.
The Good News Bible passes with flying colours, and the NBV21 only slightly less so. The Bijbel in Gewone Taal does better than the NBV21, but it lacks the Deuterocanonical books, so it is already eliminated by this point. The New Revised Standard Version does not do very well here, being very literal, but that is acceptable for a study Bible. This does not really apply to An Bíobla Naofa, which is the only complete translation of the Bible into modern Irish, and therefore has nothing to be compared with.
Notably, this is where the King James Version truly proves its irrelevance. The New International Version doesn't score too badly here, but still loses out to the Good News Bible.
4. It should include ample footnotes where alternative readings are possible.
This comes last because a Bible that scores well on the other points is still an excellent reading experience, even without providing the reader with the opportunity to consider alternatives. But for a Bible that passes the other tests, this is the icing on the cake.
The Good News Bible is excellent at this. The New Revised Standard Version is also good, and the specific edition I have it in — the New Oxford Annotated Bible — includes a plethora of additional notes and essays with further detail and scholarly commentaries. The NBV21 includes ample footnotes — though not as many as the Good News Bible — while An Bíobla Naofa includes no footnotes at all (but does have introductory essays for each book).
I feel like I have a much clearer understanding of why I like the Good News Bible now, and a better appreciation of how to measure others by the same yardstick.