This is not a situation where security is going to be an issue with any of the knots selected. So, the recommendation should be based on easy of remembrance, functionality, ease of tying under stress, and other factors. The knot selection selection should be one that you know RIGHT NOW without looking it up. If you are a knot person and you have to go look up a loop, then that is not a good knot to recommend. If you didn't remember, then you can't expect somebody else to remember it, especially if the person does not study knots as much as you do.
By the way, apparently the Karash Double Loop (KDL) is a type of Bowline and is not a new invention. According to Wikipedia, the discussion is on this site somewhere. I throw that out there because some people seem to be unwilling to use a knot if it's not a type of Bowline.
Also, Google led me to published results of strength tests of various loops, including the KDL. It wasn't hard to find. I am not posting the link here because I think the results have little relevance here. Also, the KDL did quite well on the test, and so I don't want the discussion to stray off into irrelevant arguments about why the test is unreliable. Like security, strength is not much of a determining factor for recommendation, especially if you had to go look the knot up to remember it. All the knots recommended here are strong enough and secure enough. It should be pretty obvious if a knot discussed here does not have a minimal level of security or strength.