Metcalfe's "law" is used in bitcoin to "prove" that X will grow like Y squared, for some suitable quantities X and Y. Problem is, the law does not say that Y will continue growing, that it is not being affected by MtGOX, that it will not suffer from ApplePay, that it is not determined by China, etc.
You seem to be describing interruptions to the growth of the network, but you are NOT describing the law to be invalid.... because the question becomes whether the network keeps growing or NOT and that may NOT be known while we are in the middle of the growth whether it has substantially slowed down or stopped... but after the fact, we may be able to look back and see the direction of the growth... or have better means to measure the extent of the growth... which some of those measurement systems already exist in bitcoin... but still does NOT necessarily tell us to where the trajectory is going or whether the trajectory has been materially and significantly interrupted.
There are several problems there...
For one thing, how to define the "size of the network", and how to measure it?
Total hash power is a poor measure, since over the lst couple of years the hash power is now concentrated in a few large pools. The justification given for Metcalfe's law is that each node can interact with all the other N-1 nodes, so as N increases the number of possible interactions grows as N
2, and the value of the network is guessed to be proportional to that number. But if the network, instead of adding mode nodes, is concentrating all use in a few big nodes, the interaction actually goes down. And, aniway, miners don't interact with each other.
I explained already why I believe that the blockchain traffic (whether measured in TX, BTC, or USD) is mostly bitcoins moving between addresses with the same owner; so it cannot be assumed to measure the size of the network, either. In fact, we do not have any reliable data about the bitcoin economy, except the market price, the number of coins mined, and the total hash power.
In the plots that are said to show that bitcoin follows Metcalfe's law, the last year is squeezed into a tiny area a few millimeters tall by a couple centimetres wide. But that is where most of the "weight" is. It is like plotting some property of bodies of water, from a teaspoon to the Pacific Ocean, and having the data for all the seas and oceans squeezed into that tiny sliver of the plot. Why should one assume that a property that holds for small bodies of water, from teaspoon-size to lake-size, will continue holding for ocean-size ones? (And indeed we know that real oceans have many phenomena that you don't see in lakes, like currents driven by climate differences. In particular, for all small bodies of water there are bodies that are 10x bigger; but that is not true for oceans...) If that "Bitcoin Metcalfe" plot were to be trimmed to the last 12 months and expanded to fill all the plot area, what would it show?
You are taking some arguable measurement points and arguing how these do NOT necessarily show an increase in users or transactions.... NOT very convincing to suggest that somehow bitcoin is shrinking in its importance and/or user base... especially given the fact that there are more and more businesses popping up around bitcoin.. seems like users are going to be needed to support these businesses.... and we will see, but i get the sense from the various measures on bitcoin pulse and even coinbase accounts and circle account and these other new exchanges that the bitcoin user base continues to expand at a pretty decent rate.