For McKinlay’s plan to work, he’d have to find a pat­tern in the sur­vey data—a way to rough­ly group the women accord­ing to their sim­i­lar­i­ties. The break­through came when he cod­ed up a mod­i­fied Bell Labs algo­rithm called K-Modes. First used in 1998 to ana­lyze dis­eased soy­bean crops, it takes cat­e­gor­i­cal data and clumps it like the col­ored wax swim­ming in a Lava Lamp. With some fine-tun­ing he could adjust the vis­cos­i­ty of the results, thin­ning it into a slick or coag­u­lat­ing it into a sin­gle, sol­id glob. He played with the dial and found a nat­ur­al rest­ing point where the 20,000 women clumped into sev­en sta­tis­ti­cal­ly dis­tinct clus­ters based on their ques­tions and answers. “I was ecsta­t­ic,” he says. “That was the high point of June.” […] Most unsuc­cess­ful daters con­front self-esteem issues. For McKin­lay it was worse. He had to ques­tion his cal­cu­la­tions. […] “I think that what I did is just a slight­ly more algo­rith­mic, large-scale, and machine-learn­ing-based ver­sion of what every­one does on the site,” McKin­lay says. Every­one tries to cre­ate an opti­mal profile—he just had the data to engi­neer one.

How a Math Genius Hacked OkCu­pid to Find True Love — Wired Sci­ence

I’m not sure if I should be awestruck or creeped out.

The most enjoy­able part of this sto­ry is that for all the advan­tages McKin­lay has over most daters thanks to his math prowess, once it gets to the point of live inter­ac­tion with anoth­er human he is back on equal foot­ing.

Published by

Kars Alfrink

Kars is a designer, researcher and educator focused on emerging technologies, social progress and the built environment.