A Sydney mathematician managed to crack a popular dating app's formula and scored himself a girlfriend, using what he knew best.
When a Sydney actuary found himself with too much time on his hands, he turned to popular millennial dating app Coffee Meets Bagelfor help.
Caleb, 30, who works at a prominent consulting firm, turned his mathematical mind to cracking the app's code, and it didn't take him long.
"I basically found the formula," he told news.com.au with a laugh. "I've always wanted to optimise things, make them the most efficient."
And he did exactly that, figuring out what would give his dating profile maximum popularity.
He found it was nothing to do with good-looks, with the app's algorithm promoting profiles with photos in a particular style and order.
Caleb changed his profile accordingly, including a portrait photo of himself, an image with his back to the camera in a scenic location, and a standard group picture where he was right in the middle.
"I went from having a 10-20% hit rate to an 80-90% hit-rate. It was crazy," he said.
Only a month later, he met the girl of his dreams. They've been together for six months now.
Caleb is not the actuary's real name; he asked to be anonymous in case he ever returns to the dating app, worrying the platform will ban him.
However, he thinks it's unlikely he'll have to look for another girl.
"This one will be a keeper," he said.
It all started when Caleb was on his way to a ski trip in Whistler. In a comic twist, before he even set foot on the ski field, he managed to break his foot.
Doctors recommended rest and he was housebound for 12 weeks with only his moon boot for company.
"I live by myself so I was bored, (always) messaging and texting friends," he said.
Caleb downloaded Coffee Meets Bagel and occupied his time swiping left and right.
But something wasn't adding up for the mathematician. He wasn't matching with many girls. So he used what he knew best – maths – to solve the equation.
There are two ways to meet people on the service, he explained. The first way is free, where users get a limited number of daily matches.
The second method is how the company makes money. Users can view a wall of additional profiles on a "Discover" page, but in order to contact these potential matches, they need to pay.
"I realised that the Discover tab, by virtue of being the portal for generating revenue, was full of the most popular profiles on the app," Caleb said.
So he just needed his profile to become one of them.
"There was clearly a formula but I didn't have the data."
He created five fake female profiles with his desired partner characteristics such as age and ethnicity. He used those profiles to look through 200 male profiles, methodically labelling each picture.
He recorded details of each profile, including whether their pictures were in portrait or landscape, whether the guy was smiling and if it was in black and white, to name a few.
"A lot of it was trying to identify what combo of words and pictures was the most popular," he said.
After a while "it became very clear exactly what you needed".
Caleb discovered that many popular profiles shared some similar characteristics.
"There were so many photos with a majestic view, where the guy stood with his back to the camera shoulder-width apart," he said.
"I had to go out of my way to get those photos."
He went all the way to the Blue Mountains to photograph himself in nature, with his back to the camera.
Then he carefully made sure everything was in the right order. He put a portrait photo of himself first, followed by the Blue Mountains picture. A standard group picture and a shirtless photo of himself at the beach also featured.
"After I had the full array of pictures my match rate went up significantly," he said.
"I went from having very few dates through the app to having two dates a day. I screwed up and actually double-booked someone – twice."
Eventually he met his girlfriend. He was almost disappointed that all his hard work was no longer necessary. Almost.
When he told her about his data-driven attempts to score a girl, "she was in hysterics".
"She already thought I was a massive nerd but this just confirmed it," he said.
"But she was definitely appreciative of the effort I put in."