Some notable articles (one cited in Fox News last Jan 2010) attempt to provide a numerically rigorous derivation of man's (lack of) success in finding a life partner.

P. Backus of the University of Warwick claims that on a given night out in London there is a 0.0000034% chance of meeting his potential girlfriend. That makes his odds only about 100 times better than finding an alien in the Milky Way.  The result is based on the Drake Equation originally intended to estimate the number of advanced civilizations that might exist in our galaxy. 

Another author, T. Miller of the German Research Center for Artificial Intelligence, reveals a pessimistic prediction based on parametric Gaussian distributions.  Assuming he were to go on a blind date with a new girl every week, it would take him 67 years to find his mate--3 years short of typical male life expectancy.

When a Filipino physicist (name withheld) of Universiteit Van Amsterdam was asked regarding the soundness of such models, he says curve-fitting approaches are too arbitrary.  In his opinion, one should start with the explanation of social interactions, which makes it like a QFT problem.  Taking the product space, one can get a somewhat probability from the correlation functions.  But more importantly, it is possible to develop a more robust treatise on the subject. :-)