Is love predictable?

The advent of Internet dating has led rapidly to a search for better matching results, as detailed by a recent story. After all, reductionists may wonder just how many dimensions the problem can have. Consider the following:

[Researchers] decided to employ computer technology to find a few “simple, logical rules” that make up, well, the recipe for love. For help on the technical side, they turned to Michael Georgeff, director of the Australian Artificial Intelligence Institute. During his work on a NASA project at Stanford Research Institute, Georgeff had developed a methodology to teach Space Shuttle Discovery computers how to anticipate unexpected problems. Working with Thompson and Hutchinson, he applied the same principles to the design of dating software, employing many of the statistical methods common to social science research. “Say you score a 3 on the introvert scale, and a 6 on touchy-feely. Will you tend to like somebody who’s practical?” Using Georgeff’s software, Thompson and Hutchinson then developed an online quiz., the highly popular online dating site, began using’s software this year to give users a rough sense of what proportion of the dating population might be attracted to their particular array of personality traits.

The new algorithms are designed to measure not only initial attraction, but also how well the would-be couple can live in harmony. Ten thousand people a day are signing up for, which also tries to do some simple lie-detecting. According to some accounts 30 percent of on-line daters are in fact married, and often lying about that fact.

Meredith Hanrahan, at, invokes a market metaphor:

If you want to buy a car, you get a lot of information before you even test-drive,” she says. “There hasn’t been a way to do that with relationships.”

Perhaps one web-dating entrepreneur put it best:

“Everyone is high maintenance. The trick is finding the precise sort of maintenance you need.”


Comments for this post are closed