"At a certain point it felt like we were looking down a road of impossibility. Today, it's quite the opposite."
(PRWEB) October 12, 2017
Back in February of 1996, the world was stunned when 'Deep Blue' defeated chess master Garry Kasparov in game one of a six game series. Kasparov later won three and drew two in the series, defeating Deep Blue 4-2. But it didn't change the fact that our world had now seen machine beat man at least once at his own game. Deep Blue was heavily upgraded in 1997 and played a rematch with Kasparov, beating him and becoming the first computer to defeat a reigning world champion under standard chess tournament time controls.
Now it is 2017. Deep Blue is considered a relic by today's standards. Computer science has grown beyond imagination since 1997. A San Francisco based company specializing in data modeling and statistical analysis in everything from the medical field to financial markets seems to have cracked what many said was an impossibility: the statistical hairball that is sports wagering. "This is not really a field that you'd typically see us even get involved with," says Stephen Hanner of the newly formed Blue Streak Sports. "One really tries to stay away from predictive aspirations involving so many tangent possibilities, but we all agreed this was, bar-none, one of the best stress tests you could undertake."
Blue Streak Sports (a modestly small subsidiary of BAT Technologies) then had to employ a host of programmers and at least two dozen data entry professionals to tackle the odds and build a system that would, in essence, beat the Vegas sportsbooks. Craig Demarc, lead systems architect, stated it was "one of the most challenging projects of my career"--no small claim considering his 30 year career in data systems engineering and architecture. Demarc has lectured around the world on systems development and is now under exclusive contract with Blue Streak. Over the next 6 years the team set out to create data points on every variable in professional football. "At a certain point we knew we were looking down a road of impossibility," says Demarc, "It was a statistical nightmare. You could spend the next 25 years figuring out rule changes, player variables, injuries, and referee bias, then look back and see that your data pretense is already out of date."
Hanner then had an epiphany that changed the approach. "I realized...it doesn't have to even be about the sport. As crazy as that sounds, what we began to focus on were studies of sports handicappers. These are the people who know the ebb and flow of wins and losses better than anyone. What we found was that by continuing to expand our data pool of examined results to include hundreds of handicappers from sports services to popular YouTubers, someone was always on a streak." In fact, what the team realized was that while you couldn't begin to run computational processes on the outcome of games, you could run them on the smaller, finite variants that encompass sports handicappers. "We're not handicapping sports, we're handicapping people." Hanner says with a Steve Jobs kind of smirk that says there's a lot more to it than he's letting on--or that he plans to turn us all into cyborgs.
"Yes, we can predict what handicappers are going to go on a streak. That's what we came out the other side with," Demarc states. "You can identify these markers before that streak begins. We're pushing a little over a 70% success rate right at the moment. That's the highest percentage we've hit in the past 2 years." Hanner says that when they proved the theory and began the analysis, it was like a cakewalk. "We knew this had to be as good as it gets. We mock wagered using our data model on every game for 5 years using a moderate dollar amount. At this point we'd all be multimillionaires."
And now, that's exactly Hanner's plan. By launching Blue Streak Sports (http://www.bluestreaksports.com) the fledgling subsidiary is set to surpass its parent company’s profits by a surprisingly large margin. "No one really expected this project to go anywhere past a stress test concept. It's just become a monster."
So how far can a machine get in predicting win streaks? "The sky's the limit," says Demarc. "As our process becomes more streamlined, as the technology becomes more advanced, past projects have seen us get more accurate as we're able to get far more granular with our data sets. I don't ever see 100%, but we're not stopping at 70%."
The computer is now predicting us. You knew it was bound to happen.