The requirements are that it needs to work as a ring or small diameter spar hitch of some sort (they are probably the easiest to jam), it needs to use just a single pass through the ring or around the spar (no round turns or Backhand/Munter type wrapping to reduce tension on the hitch), it needs to be quick-release which doesn't jam that easily, and also secure enough to resist slack and cyclic loads.
Some pretty specific points. Not sure of all reasons behind qualifications,
but mr.Ashley shares several times Backhand/Munter type can be single pass to dbl.bearing, and offers to reduce chafing on small hosts as well, counting towards long term strength.
Dan Lehman uses that here in a very clean release, strong, minimal hitch.

Slip bight kinda increase stiffness of Backhand keeps it from inverting to normal center pull(?) as usual greatest loaded (alone this is a good rigging hold too, doesn't take much pressure on center to keep from inverting to normal)ropePart of Backhand, and this bight for Backhand around rope is not at bottom of host towards gravity load with this .
So not sure if that reverse strategy puts this then in the scope for your project.
Double Bearing can reduce chafe in especially small hosts, as possible target here.
The sharpest primary arc is buffered behind host frictions and arching.
Used mostly on spars ; but definitely have used on carabiner and ring too.
.
Have cyclically loaded with some rotation from TDC, not untie; but not tested in all materials.
>>to fight rotation , have buffered hits some by placing works slightly past TDC/other side than load and set there.
>>the few times did use that, and even tested like that then purposefully hard>> no slight rotation as before
>>even on carabiner or ring a fewer times, setting load pull on one side of the device and knot works rotated past TDC to top of other side of carabiner/ring type device.
.
In any case, this is a fresh one for the mental and utility toolboxes.
.
i got the idea for the rotation past TDC from here the ol'man again bailing us out i think:
ABoK Lesson# 0465/pg.076: "Sailor's Hitch"/un-named Shows what we call a Sailor's Hitch, Ashley does not name this wonder at all ! But, in may be explaining that in ?:
ABoK Lesson# 1208/pg.221: Ligature/Surgeon's knot quips: that #1209 :
"is commonly called by laymen the Surgeon?s Knot.
But surgeons do not speak of the Surgeon?s Knot
any more than a sailor would speak of a Sailor?s Knot.?
-and then doesn't name the "Sailor's Knot".... (which has most exemplary architecture to some points so much as this seems appropriately named as perhaps a proudest presentation)
.
The final nip will fall into the X provided when drawn up together.
This is positioned past TDC to a more buffered position on normal loading of hang in Ashley's pic it seems. Buffered from the TDC 2x hits. This is best nip position historically thru the text, but child's swing usage is repeated dynamic/not static !
>>perhaps change in strategy to that point, to carry to other architectures.
BUT most direct nip at worst angle of swing projected, wrapping tighter as swings back the other way then to full cycle then!
Am much a student of pendulum as a model of clean horizontal movement to gravity load.
(as opposed to spring of clean vertical to gravity load, both in frictionless models)