This is a multi-dimensional question. Lots of info in other thread:
Blocks Sharing Loads. i don't know how to answer such an intro.
In that we have several factors compounding each other. There are the base/nominal forces as predictables taking place on the straight tensioned run of the Standing Tension part in it's Natural/minimal loss force/ state and loaded weight force. Any bend/deformation of the line conducting force can leverage force by angle, and weaken the line be deformity. Some of these are akin to hinging stresses. The lines though are flexibles, and only conduct force to carry load on the inline(cosine) axis, and only on tensioned directions(unlike wood).
We have a given rope strength, and a given load force. Then we have increases in load force multipliers and reduction of rope strength efficiencies; all at once't. See this as a pivotal change, whereby the increase in loading is set against a decrease in rope efficiency. This is like putting more load on a weaker line. The quicker/smaller distance a given change takes place over, the more sudden and impacting the change is.
Any bend in the Standing Tension part(s); give higher leveraged force on the line at that point. Thus, a reduction in the remaining rope efficiency. Also, there will be other changes impacting. As the line goes around it's host load, or as it is bent by a choke of it's own non-supporting leg(s) we have frictions. These frictions can cause also less stretching, than the loaded and leveraged parts. These both are changes (change in rope tension and stretch) happening suddenly; making them impacting changes (even as we are looking at a static system).
But then, we have weakness in the line at the tight bight around the line or host, krab etc. This is more like hinging, both wood and rope are fibres... Rope, works under tension, but in the tight bight, the rope side on the inside of the turn creating bight compresses, the stiffer and larger a rope is gives more compression. This compressed area, can't be tensioned, so it is not carrying the load force. Only the outer part of the bight is carrying said tension. Even the distance inbetween is just leveraged distance, not carrying much load (like as in a hinge as it bends). Like a hinge a rope's tight bight has to form 3 non-inline points, but rope can't support on compression like wood fibre. So rope is a cosine and tree is cosine and sine considerations. So, the rope efficiency/strength is reduced below 100% at this point. Certainly a lot going on, even with these basics.
Each situation, material, braid, thickness, knot type, teepee height etc. contributes something to the equation. Thus, the same material etc. in a flat sling on a curve (that isn't moving which would tear up a flat 'line'; in fact a rope is round for easier handling and moving around items) is stronger, because it will bend easier and has less 'height' on the bent dimension(2 compounding factors). How tight the teepee/choke is set, gives how deformed and tight the line is that raises the line tension).
Different knots will yield different, especially dressded by different peoples, and in different materials etc. See a single turn noose as just bending the Standing as is stated above. Giving a sudden change in tension (by deforming) and stretch. But a superior dbl.Noose/ scaffold/"fisherman's"(for some) does the same, but commits the deformation/change and stretch over a longer distance, so is less dramatic/impacting, so is 'stronger' or less loading and weakening.
Now, the difference in them in tying is taking a half hitch construction and making it into a round turn/anchor hitch, and using any round turn is usually better. But notice, if we take a round turn around the krab, then give dble.noose the effect is lost, like the the remaining rope tension is reduced, and so the noose just bends the line; doesn't have the tensions to empower the round turn to buffer the loading effects by how it grips the standing with residual tension in line making it as it bends the Standing too(?).
And also, realize the difference in how a knot lacing that serves / T-Bonz to the host mount perpendicularly isn't leveraged like one meeting pairallell, and how the latter is leveraged a different way (like timber for pulling straight away from a rail perpendicularly, and killick for an inline pull on same or to drag etc.
many combinations possible, what worx in braid doesn't in fishline, manilla had hairs to grips securely and buffer some shox(so many knots that were trusted loss favour when synthetics came along and became cheaper, stronger, longer lasting, lighter etc..).
Still, 1 can dress differently,teepee higher, angle better, deform less, run different etc. a seemingly matching system than another that is more surely crafted at each turn.
This can't be complete, hope it answers sum of those questions that burn in may of us. But, ya should really ask others like knude and Mahk...
