I gave this talk at Social Media Cafe (#SMC_MCR) it lasted longer than I expected but generally its about my believe in Maths and science to match people for the purposes of love. Most of its been said here before.
At the end of the slides I make reference to a couple who got together despite the year of making love crazyness but I’m sad to say they didn’t stay together… So maybe Annie/Sandra Bullock was right “relationships that start under intense circumstances, they never last”
I also make reference to the current worries/concerns over online matchmaking claims… I certainly feel that since Match.com bought OKCupid the level of the matching has certainly gotten worst, can’t quite put my finger on exactly whats wrong but I’m certainly feeling its not all fun times in online dating right now, even with the OKcupid mobile application.
Lastly I specially like Tom Morris‘s very detailed comment in reply to my question about matching people with science…
The answer to the question? Probably no, not at the moment, and if someone says that they can, definitely not. But that could change if psychology improves.
We are attracted to each other for complex, multi-faceted reasons. There’s obviously sexual attraction, but you can also be attracted to someone because you think they are a fun, interesting person… even if they aren’t someone you would naturally find physically attractive. The sexual attraction is easy enough to work out and self-report: you can sit down and write a list of characteristics you find physically attractive: gender, height, build, race, hair colour, whether they are into crazy fetishes – that stuff is all fairly easy to self-report.
But there is plenty of stuff about human psychology we don’t know yet. Matching people up based on self-reported questions only gets you so far. People aren’t necessarily honest in questions, and there are a whole stack of cognitive biases. Writing psychological survey questions is hard. You can have four questions which logically are the same, but if you phrase them slightly differently, you get completely different responses. You can put questions in the survey in different orders and get different responses, mix them in with priming questions and get different responses.
And if you were to come up with a matching algorithm, you’d have to compare it to a control. But there’s a huge number of other factors: you turn up for the date, and the music at the club or restaurant is not to your taste, or the food was a bit off, or you’ve had a shit day at work… and so you respond differently than the other person.
People don’t know what they want: you might say you want someone the same age, but you’ve never tried having a relationship with someone 7-10 years older or younger than you. Everything Eli Pariser has said about filter bubbles: that’s not just restricted to web content, but people too. If you made two algorithms, one for finding someone for sex and another for relationships, even people who just want sex would end up using the relationship one because they don’t want to seem tacky. Matching people is just difficult.
Absolutely… and I think the idea of using Augmented reality technologies in combination with dating data is a interesting solution and maybe the future of online dating?
