Well, since you brought up tri-axial loading. The literature does seem to support that it is not a good thing.
I did my own setup trying to understand why? So I guess any load on the locking part is not good. Although, to me this way seems to spread the load more evenly? I was never good at physics.
I came up with another technique to avoid the tri-loading problem, but now I am violating the metal on metal rule...
I know it looks messy, I did not have an actual anchor for my quick setup.
Doesn't the anchor plate and carabiner violate the metal on metal rule as well? It seems that one of those triangle shape screw links would be a good solution for the basket.
If one were to follow the "rule", you would have a soft link (webbing) between your 8 and biner, and your biner and harness.
If there is something in the system that allows for absorbing the torsional movement, like rope, webbing, etc... then the "rule" will probably not apply.
Yes, the delta (tri) link is always a great alternative when you're purposefully tri-loading.
A carabiner's strength is along its spine. They're designed to be pulled along the long axis, more or less parallel to the general direction of the spine. So another abuse similar to tri-loading is diagonal loading. Remove the force component from your tri-load picture that is coming into the spine on the wide end and you're left with the one coming into the gate side. The line you're left with is diagonal across the length of the carabiner. Also an abusive loading that can greatly weaken a carabiner... Not so much physics, per se. More so an issue of using a piece of equipment outside of the design intent.
As for a "no metal-on-metal rule", I hear that on occasion from fire folks. You can add that to the crazy dogma list of "rules":
"the bowline is dangerous"
"a tensionless hitch is always better because it's stronger"
"never use rope on webbing"
"never use webbing on webbing"
"everything has to be 15:1"
"directionals on belay ropes require pulleys"
and the list goes on...
Such dogma is simply born out of a lack of understanding, and then the teaching of the lack of understanding by some instructors causes the spread of the dogma. But I see that you're already seeing the flaw in the "no metal-on-metal rule". Keep on thinking critically.
I watched the CMC video of the MPD and it does look neat, but at a $650 price tag, I am not feeling it.
Now I decided to do a price comparison.
CMC boast that the MPD replaces the RPM system, so to be fair, I used CMC prices for RPM components when making my comparison.
Note: I did not include one carabiner in this comparison as both systems require it.
MPD = $650.00
RPM System components (Using CMC pricing)
Carabiner $30.0 ea x4 = $120.00
Brake Bar $130.00 ea x1 = $130.00
Pulley $52.00 ea x1 = $52.00
Anc. Plate $50.00 ea x1 = $50.00
Misc cord $10.00 50 ft. = $10.00 (For RRH and prusiks)
Now if I were to build my system using non- CMC parts all General use rating, I could build my system for much less, in fact; I could build two RPM's for the price of 1 MPD.
For cash strapped agencies, the MPD is a luxury, as cool as it may be.
Now if CMC were to send me one to play with, I might just change my mind. :) Oh I am bad...