Hi xarax,
A first thought that comes to mind is that, contrary to what happens with the Shakehands bend, about which both the two versions with crossed tails seem to be better than the version with the tails that pass parallel, the first version of the crossed tail Hunter's that you show above seems be weakest with respect to the "standard" version of the Hunter's bend, which seems to be in a middle way regard to the valence of the three versions of this bend,and where the second version that you show perhaps is at the highest level of stability and worthiness.
I admit that the setting that you show is what makes it more justice to the potential stability of the first version, but I find that (perhaps because I'm not very practical to handle this version) there is to "work" a little bit to get this setting , and that there is obtained almost exclusively by pulling the tails,that,otherwise, when are crossed in this manner, almost seem to work in favor of a "disconnection" between the two overhands of which is formed the knot at a higher level of as is the case for the two other versions.
As for the possible problem of jamming (which in fact may afflict mostly the "standard" version), should be resolved by taking the knot by the "collars",and try to "rotate" them towards you when you look at the knot in favor of the front(EDIT:top view(why top /bottom and not front/rear in this case?))(I agree) view.
Bye!
P.S.This has nothing to do, but just for the record I like to point out that the X1 version has the same topology of ABoK1408/9:it should be almost automatically visible if, when the knot is not yet tight, one pulls one of the two standing parts of this bend, and simultaneously pulls the tail of the other standing part .