The most important aspect of all of this is the Delta T. Heat always moves from areas of greater concentration to areas of lower concentration. Having two coils immersed vs. one coil immersed, volume and temperature of incoming water being equal, two coils would be superior due to the increased surface area. Plate chillers work so well because of the increase in SA, and counterflow chillers work so well because of the large volumes passing by each other at every moment, that is the total amount of heat removed increases due to the large volume. Think of it this way - if you dropped a red hot bolt into a pitcher of water at 33°, it would warm the water a lot. But if you dropped the same bolt into a swimming pool at 65° there would be little change - water has a pretty high heat capacity.
For flows,
q = hA(Ts − Tb)
h is the heat transfer coefficient. This is larger for turbulent flows than for laminar flows, but for comparing differences in systems, it would be nearly constant since we're only talking copper tubing in direct contact with wort. It could effectively be removed from the equation.
A is the surface area of heat transfer (for copper tubing it would be 2 * pi * r * length)
Ts is the surface temp and Tb is the temperature of the liquid not near the exchange.
Obviously keeping the wort moving (i.e. homogeneous temperature) is ideal to avoid a heat gradient where it will be cool around the coils and warmer far away form the coils, thus reducing efficiency. Also, if I were to use dual immersion chillers, I would design one to flow from the bottom up and the other from the top down, again to maintain as high a Delta T between liquids as possible.
For my money, the prechiller is really worth it. I'm not an engineer, but I do play one on the internet.