Let me try to explain why one can see full voltage or almost full voltage from the com to ground, when com is open. Refer to the following diagram. I have represented each load as a resistance.
It is important to note that the voltmeter draws almost no current. Consider first the switched load. If it is switched on, there is a direct path from hot to the com side. Since there is no current drawn by the voltmeter, there is no voltage drop across that load resistance. Thus, one sees the full voltage when measuring to ground. Now assume the switched load is off. Why can one still see a voltage in most cases? Well there is always some leakage which may be a very high resistance and might be measured. But more commonly in modern equipment, there is going to be some real, but very small load, such as an LED (whether seen or not), maybe a defrost timer, etc. In that case, the no-current voltmeter will see the voltage. That is why I suggested applying a load across the points being measured.
Now, are there instances in which an open com may result in no voltage reading from com to ground? I am sure there are, but considering the complexity of the wiring in our houses and the number of potential non-switched loads on the com lines down-stream of the break, it might be uncommon. Also, given the close proximity of the com wire to the hot wire, coupling to the field around the hot line to com (think of a radio antenna) could give a voltage reading. Just look at how the value on an AC voltmeter will tend not to be zero when one lead is grounded and you touch the other or just leave it in free air.