After the formulas

The book can be assembled into a sequence of checks. First you need to define the observable quantity: attention, desire, duration, jealousy or relationship quality. Then the binding constraint: a shortage of people, time, information, money, trust, freedom to refuse or alignment of goals. After that you check the distribution of benefits and risks, the cost of leaving for each person, and the possibility that one person’s convenience is paid for by the other’s invisible labour. The last step is the limits of the model: the conditions under which the mechanism changes sign or stops working.

Such an analysis does not promise to be error-free. It only reduces the number of decisions made under someone else’s name: dependence stops being called love, slow disappearance stops being called compromise, controlling another person stops being called a boundary, an admission made after the choice was removed stops being called honesty, and the composition of the local pool stops being called fate. Sometimes the right label makes a union repairable. Sometimes it reveals a stable incompatibility. Sometimes it shows that keeping the relationship requires from one person a price no formula is obliged to declare reasonable.

The dating app is still waiting on page one. It asks who you want to find and offers a few sliders. You can now treat them more calmly. They describe part of compatibility, part of the available market and part of the future constraints. They do not know who you will love, they do not measure the quality of a conversation and they cannot decide which risk is worth taking. None of the models in this book knows that either. Their job is more modest and more useful: to stop an elegant formula from quietly replacing the question a person has to answer for themselves.

Main sources

Joel, S., Eastwick, P. W., Allison, C. J., et al. (2020). Machine learning uncovers the most robust self-report predictors of relationship quality across 43 longitudinal couples studies. Proceedings of the National Academy of Sciences, 117(32), 19061–19071; Bühler, J. L., Krauss, S., & Orth, U. (2022). Rank-order stability of relationship satisfaction: A meta-analysis of longitudinal studies. Psychological Bulletin, 148(9–10), 699–727.