Assumptions from systems theory
The point made in the original column is that our selection of partners, on a case-by-case basis, in large part boils down to emotional connection (which while associated with genetic predisposition, is in large part environmental, and also unconscious and irrational). Surely most people fall in love not for the extrinsic benefits their partner can offer, but for a way they feel. The columnist takes issue with attempting to match people up based on comparison of data sets, and evaluating potential partners based on these extrinsic elements. I'd tend to agree - but then there are photographs to help establish potential for physical attraction, and what the hell is the point of online dating if not to result in real life dates in which emotional connection can be established?
Your assumption is that the way in which humans select partners is the same as the rational systems we have devised to explain statistical patterns in human behaviour. On a massive scale, these sorts of rational considerations probably account for a lot, but not on an individual basis. Our genes don't speak to us directly, nor are we predestined agents acting on their behalf. We often pick the wrong partners (on genetic criteria), or stay with partners who are damaging to us. Why? Irrational emotions. Which are far more significant to, and probably the defining force behind, the phenomenology of human relationships.