The Mortality Conference

Two Actuaries Calculate the Probability of Forever

You meet at the annual mortality conference in the Fairmont, where three hundred actuaries gather to discuss death rates with the enthusiasm other people reserve for sports statistics. She's presenting on urban longevity patterns, how living near green spaces adds 2.3 years to your life expectancy. You're in the back row, calculating the probability that her wedding ring is actually a wedding ring or just armor against conference creeps.

During the coffee break, you collide at the urn. She's reaching for decaf; "Regular coffee after forty is just asking for standard deviation," she says, and you know immediately that she thinks in spreadsheets, dreams in bell curves.

We are, those of us who calculate risk for a living, terrible at taking chances. We know too much about probability, about how most things fail, about the mathematical certainty of entropy. But here, between sessions on morbidity tables and pension solvency, something incalculable happens.

She finds you at lunch, sits without asking. "Your presentation on suicide rates in post-industrial cities," she says, "it was almost poetry." You talk about Montreal's numbers, how they spike in February, dip in June. She talks about Vancouver's rain, how it correlates with everything except what you'd expect.

"We're professional pessimists," she says over the conference wine, which tastes like disappointment with hints of oak. "We literally calculate how things end."

But that night, in the hotel bar, she removes the ring—"Divorced. Officially, statistically, irreversibly"—and suddenly the math changes. Two actuaries, you realize, might be the only people equipped to love despite the odds, because of the odds, inside the terrible beautiful knowledge that everything ends.

The conference closes with projections for the next century. You sit together in the back, her hand finding yours in the darkness of the PowerPoint presentation, and for once, the calculations fall away. Love, it turns out, is the only variable that refuses to fit the model.