Fascinating: scientists have run a computer simulation of the dating market and have published their conclusions in a scholarly journal of mathematics. Sanmay Das (M.I.T.) and Emir Kamenica (Harvard) used a scenario from Game Theory, “The Two Bandit Problem”, to predict regret and relationship stability in dating partners.
The simulation invokes things like the Boltzman distribution (familiar if you’ve ever taken third year college chemistry, sometimes called P-Chem), applied here to dates instead of molecules!
The study concludes that relationships are most stable when “agents are patient in two different ways —if they are more likely to explore early or if they are more optimistic.”. Whatever that means. I don’t yet completely understand the article.
Three quotes from the experiment:
- "After a period of exploration, where the agents match up with many different partners and learn their preferences, agents start pairing up regularly with just one partner, and this is always the agent with the same ranking on the other side. This indicates that agents are generally successful at learning"
- "The reward of asking out a particular man depends on the probability that he will accept the offer. Thus, the reward distribution changes based on what the men are learning, introducing an externality to the search process."
- "Interestingly, even if only one side explores (that is, either men or women always pick the greedy action), populations almost always converge to stable matchings, with a slight decline in the probability of stability when only men explore (under the woman-optimal matching algorithm, women’s rankings can have a greater effect on the matching than men’s rankings)."