Hi Mark, It was really for illustrative purposes more than what takes place over time, I admit to still learning and it's observational data, happy to have input from yourself, it's marvelous as you are one of the few to have such an understanding, so here goes.
The force of the bullet, or water dissipates, as we know energy can only be transferred so it begins to move more of the water, and more and more. Does the smaller faster orifice impart more or less energy than the slower moving but larger orifice? I think the restriction in the smaller fitting is robbing some of the energy. The fact that the water is nearer the surface means the air being less dense than water will allow as you describe allow a passage of water to be directed where you want, it's that rippling path on the surface. Does the larger less restrictive orifice moving more water, do the same, yes but a little slower, this I have observed to be actually more accurate at finding the path than the faster smaller orifice as that creates a more turbulent street so diverges more into the surrounding pool.

Most pool builders are obsessed with the circular path (Dennis demonstrated this several times) that swirling brings debris out of suspension into the dead spot in the middle of the pool where it sinks to the bottom. The same ideas of skimmer at one end and returns at the other is equally obsessed over. The directed return jet has control over leaves for a short distance only (10-12 feet) after which the flow decays and the energy previously put into the pool making it swirl takes over from the now diminished jet and the leaf can miss the skimmer, it then circulates once more and if it misses again arrives near the middle, partially water logged and beginning to sink. The pool robot salesman rubs their hands with glee as another willing customer arrives $$$.

Some pool builders do it slightly differently, returns one side and skimmers the other, the shortest distance, so much more obvious as the return jet, fast or slow will still have control over the leaf at that distance so it's easier to direct the leaf towards a skimmer (should almost always be more than one) add then that the return placed near the skimmer, facing downwards to action a horizontal rolling action creating lift and more debris is going into the skimmer without a main drain in sight.

To the other point:
If they ran say two single 1.5" lines but had a 2" to the pump from the skimmer that would be over the top unless as you said the return path was longer and more torturous, that's why we try to calculate the headloss in any pipe run. Depending on the path, number of fittings etc I would most likely run the same size pipe and separate the flows via Y's not T's to the returns but obviously each pool has it's own specific solution. As you said earlier the smaller orifice only adds 0.5 PSI, as my system runs between 0.25 and 0.75 PSI I really couldn't add 0.5 just to squirt water a little faster. LOL. I calculate the area of the pipe work and make the returns taper smoothly to the orifice dimensions that add up to total of the supplied pipe, I may vary that if I do want a directional faster flow and may also balance that out with the others.

I try to arrange the pad very accurately, pump inflow below the water line, return path above the water line but flowing down hill using the water in the pool as the minimum restriction possible, yes really obsessive, I want to try and run a system as a siphon one day. For now though, more like the HVAC just the minimal push from the pump but achieving reasonable flow. If I had a normal pool pump on there I would imagine it could cavitate especially with the pump going 0 to full RPM in under a second, I am surprised more pumps in the USA don't suffer cavitation because they are so powerful and the headloss in some cases is so high the water must be boiling inside the volute?

I the EU we have several pool companies that market monoblock filtration units in their pools, these pools have far the worst flow of any pool I work on, especially at the far end as they have a single return mounted very near the skimmer so water takes the short route. They do eventually (well sometimes) manage a bulk water movement but it's really poor and that is utilising a pump twice the power of a conventional pool (most USA pools are already using twice the power LOL) than we would in Europe but you do love add ons but even Sunny Optimism has a large pump for just a water fall but I digress. These pools still work when everything is within range so from that i deduced that we can really reduce the flow tremendously and likewise the power and still have a good pool. When the monoblock units have an algae issue then you really see how bad they are it can take more than a week to get them clear (bad course bag or pleated filters placed on the pump intakes rather than the outflow) I experimented on my own pool and even running the low power I do it still out performs the monoblock filtered pools but i have a lot more in reserve so if I need to get the water moving I can up the flow.

For commercial reasons I don't want to name the pumps I use but it's a lot more than just the pump, the whole circuit is improved where possible. I used to make many of the parts myself but as I realised that was more the icing on the cake I now adapt other parts to suit.