John,
You said you had never seen any authoritative reference to bank. You will find one in Appendix G of Torenbeek's "Synthesis of Subsonic Airplane Design" where he writes:
>It is generally found that the total aysmmetric drag after failure of a starboard engine is minimum for a small negative slip angle, i.e. the airplane slips in the direction of the operative engine. However there is very little drag increment in a flight with zero sideslip. In that condition the most important drag contribution is caused by the vertical tail>
He cites another paper delivered by J G Callaghan of Douglas as part of AGARD Lecture Series 67. (I was there but had forgotten it till checking the reference! - old age)
Callaghan gives a comparative chart showing the various contributions from windmilling, rudder, sideslip, aileron and spoiler. Given the date of the lecture (1974) I think we can safely say it is old technology and a contemporary MDC twin engine design.
Unfortunately his picture has no numerical Y scale, but one can get a good idea of the magnitudes by simple measurement.
For the total aysmmetric drag increment that gives:
Bank/sideslip/drag units
5/-6/5.6
2.5/-2/4.5 (optimum)
1.5/0/4.6 (zero slip)
0.5/2/5.5
0/2.9/6.8 (wings level)
-2.5/6.8/10.9[/B]
You can see from this that the difference between optimum slip and zero slip for this aircraft was pretty damn small! This might not be true for other designs - it all depends on the induced drag factors of the VT and how you use either slip or rudder to generate the corrective yaw. Wings level was definitely off optimum.