Is this understanding of quantum entanglement correct? The way I understood it is as follows:
Let the superstate be represented with a ‘spinning’ coin, with heads and tails representing the two possible outcomes of collapsing. The probability mass function behind the two outcomes needs two real values for a full description. The ‘spinning’ of the coin represents that the binary outcome value (heads or tails) is not yet known. The spinning stops only when the measurement is made, returning one of the two outcomes in random.
When we have two spinning coins, there is also a probability mass function that describes the probabilities of drawing the four possible outcomes. This would need four real values for a full description.
When two spinning coins are entangled, the probability mass function of the four possible outcomes take extreme values for some of the outcomes. For example, outcomes 00 and 11 will have p = .5 each while the other two are completely ruled out from being drawn.
From here, I have two competing ways of understanding entanglement:
a) when the two coins are entangled, slapping one of the coins to get an outcome (getting the measurement), instantly stops the other coin from spinning ignoring distance between the coins. The outcome of the other coin is consistent with the outcome of the first coin as prescribed by the probability mass function.
b) when the two coins are entangled, slapping one of the coins to get an outcome, does not stop the other coin from spinning. But when the other coin is slapped it will give only one possible outcome consistent with the outcome of the first coin.
It looks like a) is the case supported from experiments, but I am not yet completely sure if I am getting this right. Is my understanding correct?
 A: If you are building an analogy like this, make sure that no physically observable effect is caused by the collapse of the wavefunction, because it is a subjective event that reflects the knowledge of the observer for which the wavefunction has collapsed.
Whether or not the remote coin keeps spinning is something physically measurable, but in reality there is no measurement that could tell a third observer if the wavefunction has collapsed for you or not.
So in the coint analogy, the other coin has to keep spinning, because it cannot be instantly visible at a distance that you made a measurement. On the other hand from your point of view both coins are in exactly the same state: stopped, collapsed to the state that you measured. 
So there is an apparent paradox that it cannot be decided which is the correct description of the object. Is the other coin in a superposition, or is it collapsed? The resolution is that there is no one universally correct description of this. Different observers can have different wavefunctions describing something that happens, and the wavefunctions are different, because they represent the different knowledge that each observer has. 
For more discussion look up Wigner's friend thought experiment, where Wigner's friend made a measurement, but in Wigner's description the state is still in the superposition.
