First, your timing question:

When a motor is physically set to 0 degrees, that simply means that the Hall sensor output will transition from one state to the other when the back EMF for that phase changes polarity (the so-called "zero crossing point"). For an "ideal" motor and controller, this should be the best switching point for commutation ("best" being defined as making the best usage of the flux linkage between rotor and stator).

A polyphase permanent-magnet electrical machine requires no phase advance (AKA "timing") to generate torque from a dead stop; this is the primary advantage of using a three-phase design (as compared to single-phase motors such as those used in window fans, air compressors, and so on). In fact, at low speeds, phase advance beyond 0 degrees is almost always disadvantageous to the goal of making maximum torque.

We use timing for a couple of reasons. First is to counteract a variety of effects that occur in a practical system - ESC software and hardware has latency, windings have inductance, and magnetic fields distort slightly with increasing torque. The second reason (and the most noticeable to hobby users) is that advanced timing can allow the injection of more current as the BEMF increases; you'll need to read through to figure out why this is important.

Second, your question about efficiency:

Without knowing the exact design and specifications of the motor, it's pretty much impossible to determine the efficiency at any given operating point. It's really difficult to even give any general suggestions, since relatively minor changes to any component can dramatically change the area of peak efficiency. Basically, higher speeds cause greater "iron losses" (ohmic losses in the stator iron caused by eddy currents and hysteresis losses; both get worse at higher rotational frequency), and higher torque causes greater "copper losses" (ohmic losses in the winding and controller FETs due to the greater current required to generate torque). As hard as we push the motors in hobby applications, both are considerable.

Third, your question about Lenz's "law":

Any permanent-magnet electrical motor is also acting as a generator whenever it is spinning (whether by an external force acting on the shaft, or by internal magnetic forces). The voltage that is impressed upon the winding is called "back electromotive force" or "back EMF" (sometimes also abbreviated to BEMF). It is proportional to the speed of rotation, and can be characterized by a simple expression of volts per RPM for those of us antiquated enough to still use imperial units. We call this parameter "Kv". The rest of the world would prefer to use the parameter "Ke" with units of volt-seconds per radian, which seems clumsier but actually makes life a bit easier as we'll see in a moment.

This is important for a couple of reasons. As you pointed out, this BEMF opposes the supply voltage; thus, as the motor spins faster, less effective voltage is available to force current to flow through the winding, and the output torque decreases. Eventually, no additional torque is available for acceleration, and the motor speed reaches equilibrium with the load.

In an ideal system (one with no losses), the maximum unloaded speed is indeed the battery voltage multiplied by the Kv parameter. This is never the case, since we always have losses, but hopefully you get the idea that motors with a higher Kv constant can spin faster using the same supply voltage.

Next, there is another parameter called "Kt", which expresses the relationship between torque and current. In the metric system, it has the units of newton-meter per ampere, and has the same value as Ke. If you do the math, you will find that as Kv increases, Kt decreases and vise-versa. A motor with a higher Kv value can spin faster, but requires more current to create the same shaft torque (file this one under "there is no such thing as a free lunch").

Advancing the timing has the effect of applying battery voltage to the winding before the BEMF rises, so we can extend the speed-torque envelope. The consequence is that we do not make best use of the flux linkage in the motor by applying commutation that is out of phase with the BEMF, so we are not as efficient in producing torque. Thus, more current is required, and we lose more power in the winding, and the motor runs hotter.

This all starts describes the relationship between the critical inputs and outputs of the motor - namely voltage, current, shaft speed, and shaft torque. Hopefully this is reasonably easy to understand; if not, shoot out some more questions and I'll at least attempt to answer them.