The Equation for Divorce
One of the most famous examples comes from psychologist and researcher Dr. John Gottman, who brought mathematical modeling into his “Love Lab.” By observing couples and coding their interactions, Gottman and his colleagues were able to predict which pairs
would divorce with startling accuracy—often cited as over 90 percent. The key wasn't the arguments themselves, but how partners interacted. The biggest predictor of divorce was contempt, which erodes the foundation of respect. His research, conducted with mathematician James Murray, identified a “negativity threshold.” If positive interactions consistently outweighed the negative, couples thrived. But once interactions tipped into a spiral of negativity, characterized by criticism, contempt, defensiveness, and stonewalling, the relationship was mathematically on a path toward separation.
The 37% Rule: When to Settle Down
If you've ever wondered how long you should date before settling down, mathematics has a theory for that, too. It’s called the “optimal stopping theory,” or the 37% Rule. Imagine you plan to date a certain number of people in your life. The theory suggests you should spend the first 37% of your dating life just exploring your options and setting a baseline. You don't commit to anyone in this initial phase; you're just gathering data on what makes a good partner. After that 37% mark has passed, the rule says you should commit to the very next person who comes along and is better than anyone you dated in that first group. This strategy doesn't guarantee you'll find the absolute perfect person—you might miss them if they show up too early—but it mathematically maximizes your chances of picking the best possible partner from your pool of candidates.
The Science of a Stable Match
Another fascinating area is the “stable marriage problem,” which was solved by the Gale-Shapley algorithm. This algorithm provides a way to pair up two groups of people (like men and women, or medical residents and hospitals) based on their ranked preferences, creating a “stable” outcome. A stable matching means that there are no two people who would both rather be with each other than with their assigned partners. The algorithm works iteratively, with one group making proposals and the other group accepting or rejecting them based on their preference lists, until everyone has found the best possible partner they can get within a stable system. Interestingly, the group that does the proposing ends up with a more optimal outcome, getting their best possible matches. It's a powerful demonstration that a 'good enough' pairing for everyone can lead to a more stable system overall than if everyone holds out for a perfect but perhaps unattainable match.
Modeling the Ups and Downs
More recent research treats relationships as dynamical systems, much like predator-prey populations or even the start of a war. Mathematician Laurent Pujo-Menjouet and others have developed models based on three key factors: the mutual attraction between partners, the natural erosion of feelings over time, and the way each person reacts to the other's emotions. These models suggest that every relationship has a critical threshold of strength. If a couple's connection falls below this line and stays there, the math predicts a slide toward separation. The important factor isn't whether couples fight, but whether their bond is resilient enough to recover and return to equilibrium. These models aren't meant to be crystal balls, but rather tools to help identify when a relationship might be entering a danger zone, giving couples a chance to course-correct.
