It appears that many of the questions raised by the original poster have been mostly answered?
Not really - what about the bear?
The strangle backup for a bowline can be tied on the: near leg of the loop, far leg of the loop, or the standing part. Dan pointed out that the far leg may be more secure against accidental ring loads than the near leg? Are there any obvious objections as to strangling the S-Part (the loop is cleaner this way)?
Any time you use such an untensioned backup measure, there is not much to resist flogging forces that would pop things open, especially if the line has some stiffness to it. This is why these types of backup measures are often the first to fail and then provide little help beyond the extra rope length once they come undone.
Would it mean that having a limited choice between the BL with one of the Strangle backups and the Scott’s Locked version you would choose the latter for greater (flogging bear) security?
Since I choose my own knots, I would decline either option. If someone was asking which had better flogging security, that would be a matter for testing, preferably with the most unruly rope the person anticipates using. Unfortunately, the initial tightness or looseness of the strangle backups would likely make a significant difference in results.
:-) The bear is trained to chew up all the other knots, except the two in question - with which she just fiddles. Not tying a knot at all is verboten and there is no more time for research.
ahem…
3 years of use and counting. Flogging does nothing to my rock of a knot. The rock only loosens whenever the bowline collar can move at all (that should be almost always) and with two simple moves (bend collar and pull slack, bend girth hitch loopback and pull slack). Ring loading does nothing to the knot. The untensioned backup slippage is a few mm on a 2 KN fall and cyclical loading changes nothing with that. It is set once it is set. The end bound through that girth hitch simply isn’t moving.
Today, I experimented with a Blake’s Hitch (5-2, 5-3, 6-3…man, gym ropes are old and full of skin and metal) with the (super extra long) tail end of my knot as a Blake’s Hitch backup and part of an ascender / top-rope-solo system using the top-rope self-belay side “tail” with a Blake’s hitch to (either side) of the top rope as the foot step ascender.
—side note, forcing myself to hang on my butterfly emergency stopper and then find a way out forced my (belay side tail of rope on an already unweighted top-rope-solo belay) use of a rope that should not generally be traveling up (ATC Pilot, hold the rope down, please)… self rescue from a pure hanging stance—
To be fair, it took me 10 seconds longer to untie my tie-in than it usually does. After ring loading my weight on a hold for fun and then loading the tail’s Blake’s Hitch maybe 20 times, then hanging on it on and off for 30min, it untied as well as I could expect of any knot. De-loading the collar was the hardest it has ever been, due to the tail pull. I’ll leave the collar looser, next time, to see if the tail locks down on the end bound move earlier when given a little extra collar slack. I tend to tighten the collar in normal use, and did so, today. Collar slack should be maintained with my “pseudo”-bowline (and all bowlines) (either actually put it through its paces or leave it alone, Mark) even when the tail needs to be loaded. It never tensions without tail load, so tightening it has never been an issue.
(please put it through its paces, Mark 
mcjtom, the bear doesn’t tear this one apart. If I had to choose? Neither. Mine.