This article was published on February 3, 2011

NumberFire predicts the Super Bowl winner


NumberFire predicts the Super Bowl winner

I’ll admit it, I didn’t know who was playing in the Super Bowl until yesterday, which was also the day I found out that the Super Bowl is this weekend. So when a friend told me that a very cool New York City startup called numberFire was able to predict the outcome using sophisticated data-driven analysis, I was much more interested in the algorithms and software behind such a bold statement.

Not only did I find a gem of a company, but I learned its founder Nik Bonaddio runs the company on his own using the $100,000 he won on “Who Wants To Be A Millionaire?” “You might say that Regis Philbin was my first angel investor!” jokes Bonaddio.

Launched in September 2010, numberFire is a stand-alone web application that uses a transparent suite of data-driven algorithms in the same way that financial companies use modeling to help understand the market; but numberFire uses them to help people dominate their Fantasy football leagues. Fantasy football, in case you didn’t know, is a cultural American phenomenon, a game that turns Football into an (even more) heated rivalry amongst friends, co-workers, and families. An estimated 20 million people play fantasy football, with total spending (including ad revenue) totaling $2 billion-$3 billion.

“I’m not a huge betting guy, but we made a small test case to see if we could outsmart Vegas as well. We’re 9-1 against the spread so far, and 8-2 against the over/under; basically, we’d be banned from Vegas if we ever showed our faces there. We had a great moment last week where we nailed the Pittsburgh/Jets game down cold; we projected 24-18, and it was 24-19,” says numberFire Founder, Nik Bonaddio.

The algorithm works on the concept of similarity. “We try to find things have happened in the past that are similar to those occuring in the present. For example, in the Super Bowl, the Steelers are representing the AFC. The Steelers play a very particular brand of football – heavy on running, blitz schemes, that sort of thing – and their statistics tells us that. The statistics also tell us about all of the teams historically who played a similar style of football, such as the 1995 Cowboys or 1985 Bears. If we can make a connection in the present to things that have already happened, we can learn from it and have it give us a better understanding of what is likely to occur,” says Bonaddio.

An analogy I like to use a lot is the common cold. Everyone’s had a cold before, so they know exactly what’s going to happen when they get it. Sports, we think, can be the same way.

NumberFire is on point 70% of the time by taking a unique and analytic approach to the game. The site’s traffic has grown 100% month by month since launch, despite zero user acquisition, PR or marketing costs.

Sports are fundamentally about numbers and statistics, yet the “experts” who analyze them do so in a manner that is totally qualitative, which is no different from a financial analyst recommending a stock based on the color of their logo. numberFire addresses this problem by applying a customized, data-driven algorithm to football, giving our users highly scientific and accurate projections, which in turn allows them to make smarter decisions and dominate their league. numberFire Founder, Nik Bonaddio

“Despite the growth of the sector, tools and services in the space have remained remarkably static,” explains Bonaddio, “Industry leaders such as ESPN and Yahoo! provide advice, but their suggestions are very inaccurate and have no scientific basis. Data-based services such as Bloomberg Decision Marker and PredictionMachine say that they use numbers but are not transparent, giving the user no insight into the numbers themselves and no reason concrete reason to trust their projections.”

Numberfire incorporates a mix of subscription revenues, sponsorship revenue, and content distribution revenue. Subscription revenues are derived from building out new tools to complement the projection engine, incorporating the same data-driven strategy to tackle different aspects of fantasy football, such as trades, roster management, and waivers. Bonaddio plans to partner with major media outlets and contributes twice weekly to Sports Illustrated.

CBM: So, who’s going to win the Super Bowl?

NB: This is a hard question for me to answer, because the algorithm is in total disagreement with my heart. I’m a Steelers fan – I lived there for 23 years – so my heart is definitely saying the Steelers. However, the algorithm likes the Packers on the basis of history, making the case that Green Bay’s superior passing offense and turnover-driving defense will overcome the steady and consistent Steelers. I should note however that the projection is extremely close – the official guess is Green Bay 22, Pittsburgh 20.

Check out this visual presentation of his prediction:

If you happen to be heading to Dallas this weekend, check out SeatGeek, which allows fans to shop comparatively across the top secondary markets like StubHub, TicketsNow and Ticketnetwork to discover the best seats at Super Bowl XLV using interactive seating maps. The sold out event is commanding higher ticket prices than any Super Bowl in recent years, with listings currently averaging $4,654. You can see all available Super Bowl tickets in realtime here.

If you need a place to stay in Dallas, I recommend checking out iStopOver.

And one last tid-bit, check out my favorite new commercial which will be played this Sunday during the Super Bowl.

Wicked right?

Get the TNW newsletter

Get the most important tech news in your inbox each week.