Microsoft Research Blog

The Microsoft Research blog shares stories of collaborations with computer scientists at academic and scientific institutions to advance technical innovations in computing, as well as related events, scholarships, and fellowships.

CIKM: Sports bookie wins machine learning competition

October 21, 2015 | Posted by Microsoft Research Blog

CIKM-2015-Hung-Doan

Editorial note: The CIKM 2015 Machine Learning Competition asked participants to predict the outcomes of every match in the Australian 2015 football league season. The competition received more than 90 submissions from the machine learning community. Sports bookmaker Hung Doan won the competition using just a spreadsheet. Below is his story, told in his own words.

 

Hi, my name is Hung, and I’m the winner of the CIKM 2015 Machine Learning Competition.

I was born in an Indonesian refugee camp, where my parents stayed as part of their journey to escape the war in Vietnam, making their way to Australia.

Growing up in Sydney and excelling in school, I finished high school with a score of 96.7 and went on to study for a bachelor’s degree in computer engineering at Sydney University. Although I completed my degree, I never felt comfortable nor passionate about my chosen field of study, and at the time I saw it only as a natural progression following high school, the next step that leads you to a “brighter future.” This was an idea instilled in me from a young age by my migrant parents, who after all their hardships only wanted the best for their children and believed it would only come through the highest education one can achieve.

Upon leaving university I found it quite difficult to obtain employment in my chosen career path. I eventually accepted a job as a help desk technician in an attempt to gain a foothold in the IT industry. But after a year, my parents had the opportunity to open up a bakery, so I quit my job to help them manage it and work as the overnight baker. Trading the effortlessness of being in the office for something very physical, having long hours, starting very early in the morning, and no longer making any use of my degree, I did, however, feel a sense of freedom: I accepted the fact that I may not find any job within my qualifications, and decided to enjoy learning and improving ways to make bread.

Four years later, my dad became gravely ill and passed on. Grieving heavily, my mum felt that she couldn’t continue to run the bakery, as being there was a constant reminder of how much she missed my dad. So, we sold the bakery at a loss, which lead me to declare bankruptcy at the age of 26.

Unemployed, financially depleted and in need of work, I took on the role of a driver delivering groceries to make ends meet. After 4 years of driving I realized that such a labor-intensive job couldn’t be sustained as a long term career, as you are physically limited by your health, strength and endurance.

CIKM presentationSo began my research into other less-strenuous means of making a living. I began observing and studying the mathematics behind sports betting odds and advantages held by bookmakers in order to turn a guaranteed profit depending on the odds they set. Wanting a career change into an industry I always had a curiosity for, I applied for a role as a sports trader with a leading bookmaker and was quickly rejected due to a lack of experience.

A couple of months passed, and then one day I received a call from the head of trading in Australia to come in for a quick chat.

When I came in, I was told that my cover letter was well written and something stood out about the fact that I was driving a truck but willing to learn and do the research about something I was interested in. I was offered a job as a trader’s assistant, and within 6 months I was promoted to become a soccer trader for what is one of the biggest bookmakers in the world.

In my yearning to learn new things I stumbled upon the CIKM website and Sportsbet’s machine learning competition. I decided that I should give it a go to see how well I would perform against others in the data and machine learning industry. Never would I have expected I’d win a competition with such a high caliber of entrants.

—Hung Doan, with Christine Clifton-Thornton

For more computer science research news, visit ResearchNews.com