It is a valid interaction and it takes place on daily basis, see for instance electron capture process.
The process H+H -> D + e^+ + neutrino
is also allowed, but occurs only on daily basis in very hot places, like the core of the sun for instance. So there is no correlation between a valid allowed subatomic process and "daily basis" (whatever that might be..)
Ok, the "weak fine structure constant" is about half the value of the electromagnetic one, it is about 1/400. So the strength of the interaction is about the same. BUT the RANGE of the weak interaction is very very very small. The exchange quanta of EM force is by massless particles, photons, but the weak interaction have particles of mass about 100times the proton mass. The interaction will depend on the mass ^(-4) of these particles.
(The matrix element is proportional to 1/(|q|^2 + M^2), where q is momentum transfer, and a process depends on the matrix element squared. When the mass is zero, you don't get infinity, that is a quite usual question one get's "if the photon then is massless, why is not that strength infinite if it depends on 1/(M^4)?")
So, this means that a process has about 10^-11 times smaller cross sections -> i.e probability to occur. So, you were correct! :-)
No we don't know why it has that value, we does not know why any constant at all has its value (the electron mass etc). These questions are asked in the grand unified theories and theories for everything (such as String Theory)
That question is nonsense, for the first, the virtual particles is not really exchanged, they are mathematical "tools/terms" which arises when you do a perturbation expansion of the interaction.
Secondly, when you calculate that, you will have an infinite number of such terms (Feynman diagrams which I assume you have heard of and seen). So there is an infinite number of virtual particles being exchange (in each instant). The nice thing about perturbation theory is that since each vertex of the particle lines in a feynman diagram depends on the coupling. (in electromagnetism, the coupling is the fine structure constant).
So, the first order diagram, where one particle is exchanged, contributes to the sum 137 times more than the second order diagram (where two particles are exchanged). And so on. So the value of the process (the cross section, the likelihood to occur) is often very very good approximated by just calculating the first order term. But the total process IS the ENTIRE sum of all such sub-processes.
Third, not even in theory you can calculate how many is exchanged within a time interval since they occur in one instant.