To me the eye style of quick remote release of midline eye most looks like this old truck driver training school method, here shown doubled, both cleanly simple.
The 2nd unused eye formed by each used eye is NOT a good secondary pull point in each.
.
i too like a sliding Prusik etc. for more range/adjustability too for eye, if pulling tree or other high 1x, cant fail/slip usually make B'Fly eye tho as a locked device at high loading. But can unreeve this non-tied eye shown and shake/dance/flirt eye out when have access to ends not so easy center on at destination/delivery, like a trucker might over a load, with that eye floating high. When initially setting up these it is good to also, set as can, soft lock, then sweat/swig, release and tighten again, as usually is more slack now. Can rinse/repeat a/n. Same geometry that Mick offers at finish after lock, used in the tensioning stage also.
.
Another correctly placed Prusik or 2 can be placed inside system as safety/rest auto-hold of purchse, if can setup to be combed when pulling correct direction and extend when tries to de-compress.
.
This setup geometry to me has 3 external main points 1x,2x,3x; so a
very flexible tool than just pure power return, but good for that too especially tweaked/compounded.
>>generally input single pull to the 1x, 2x as pivot(machine point) and then finally output to the 3x.
>>generally looking for static pivot, and then the input/outputs as dynamic, potentially in anything setup to this geometry
>>any position can take turn being any of the 3 utility functions, so can have like a car transmission of speed/distance trade for power from same 'motor' source.
Ushape deployment (like over load)w/rig on 1 leg, can give 6xEffort potential from the 3xRig etc.
>>but should be sweated/swigged to equalize forces on both sides of U, then re-tightened
>>rather than road vibrations equalizing Urig portion, so not as tight on tightened side..
.
GAME CHANGER:
Compound inputs by hanging all of body weight on single input leg and pulling up on correct leg of 2x(that doesn't go directly to load proposedly on 3x external point)gives increase to 4xEffort potential at both ends equally tho, as now a closed system , BUT, then also PLUS the original 1x,2x,3x of the bodyweight 'count' at each respective point in the balance Naturally made. Next upgrade is to stand in loop on 1x and input leg instead of arm force as the input with the dual input, taking our transmission quite far on the granny gear side. Note how now use effort PLUS the promised equal and opposite of your effort back into system, AND work those 2 matching effort amounts and bodyweight as separate available inputs thru same system to sum. Sifted function out of DdRT or dumbwaiter 2xSelf Lift; just now body weight is another input against Load, not the Load itself. Visualize if stand on scale and weight(usually unrealized to target) 150 and can orchestrate this dual pull w/o changing weight on scale then is getting just the 4x each end output, then as bodyweight comes off scale it feeds into the system to give the normal 2x,3x outputs. Simply hanging guarantees all bodyweight used, and also the autonomous equal and opposite 'free' force used/folded back into system to work against the load again. But is multiple inputs to be recognized to command then as can. The input and e/o are linked, but the bodyweight influence is separate, so can hold with effort and impact w/bodyweight, or impact with both at same time etc.
Tracing the collected forces thru the compound system, we can find these differences.
Impact can increase inputs, like sharply picking up feet to hang, and snatching correct part of rope too, then allowing frictions of inefficiency to help hold this 'purchase'(ABoK term for opposite of paying out rope to Load side of system).
.
Effort multiplier increase from this usage in any system will be :
2x greater end (like the 3x end) -2 =4x shown which will equal:
2x middler end (like the 2x end) =4x shown.
So in a 1x,4x,5x compound effort system will yield
2x 5xEnd-2=8x which will equal/cross-verify
2x 4xMid =8x.
Then add the standard 1x4x5x bodyweight multiplier too(link).
This can look like same process deftly done as matter of course, in same amount of time, to get game changer magic..
Figured out sitting in DdRT, and extending to may uses of rigid and flexible(rope) leveraged systems alike(link)
"In Tree"(in country), climber is thus in a very unique position to tighten/help tighten a rig even just over a support branch
>>AND get leg force pretighten on Load side, BEFORE FRICTIONS, then also leg force after frictions as equal opposite>>same principle
Friction points work against trying to input force, but then hold force once gained behind friction buffer
>>so as lose input efficiency to frictions, can find the also offer sweat/swig and capture points to your favour.
.
Then, after linear as tight as possible, bend/leverage as Mick says for greater tension by 90degree pull(works with webbing too, bend from flat side of web not to distort/deform weaker).
Tighter linear structure first, leverages tighter on bend 90 degrees, as like less rubbery wrench/more RIGID thus efficient leveraging force returned by tool. Like sweating/swigging a rope tighter, which can be done as input also. Want so tight can't bend to 120, where side force equals endpoints, more bend becoming more like 2/1 against side force, where side as input loses leverage/gains speed. From 120 and closed all the way to Zer0 spread would rather pull end to get more leverage as choice in the way this balance works.
i like a pair of these , then middles bent towards each other, or even better rig upside down U over load(more power), each U bent against own legs, then possible to each other too. You can bend some soft metal if in line of fire tho working this strategy at full force.
.