How IBM’s Watson Improved the ESPN Fantasy Football Experience

Share
ibm-watson-espn-fantasy-football
Photo credit: ESPN/IBM

(*MKTG is a Proud Partner of FOS)

More than 10 million people played fantasy football using the ESPN Fantasy app this past season.

Normally, users spend many hours over the course of a season researching how likely a player is to perform well any given week. Consuming every piece of fantasy football-related content during the season is not something any person has the time or energy to do.

For this reason, IBM and ESPN launched a groundbreaking initiative leveraging IBM’s globally renowned artificial intelligence system, Watson

“One of the key benefits of AI is being able to consume huge volumes of information in many different formats,” states Noah Syken, VP of sports and entertainment partnerships at IBM. “So that was particularly compelling to us. How do we actually scoop all that information up, understand what is and isn’t valuable, and be able to present it to fantasy players in a way that’s really compelling? We feel like, with the boom/bust feature of this in particular, we accomplished that.” 

Watson analyzed millions of documents, including news articles, blog posts, videos and podcasts every week during the football season. This analysis yielded insight that ESPN presented to users through the ESPN Fantasy Football app and website, projecting, among other things, how likely a given player was to score on the high or low end of their ESPN-predicted range.

In other words, while fantasy apps have historically provided projected point totals for a given player for a given week, Watson adds a layer of data focused on the probability of a player scoring more than their projected point total (“boom”) or less than their projected point total (“bust”). Watson also leveraged ongoing analysis to project a player’s likelihood to play in a given week as well as the current and trending media sentiment about players.

“We’ve been talking about what meaningful ways we can bring IBM technology to ESPN platforms for a while,” says Elizabeth O’Brien, sports and entertainment partnerships program director at IBM. “Fantasy football seems almost custom-made for the technology that we developed. It has enormous popularity, it’s all digital, and it relies on massive amounts of information and data. So, we pulled together a construct for how AI could help fantasy owners while complementing the expertise of ESPN experts.”

The ESPN Fantasy Insights with Watson capability draws upon the latest in machine-learning techniques and deep neural networks to analyze “dark data” that traditional software cannot process. An expert team of IBM IX data scientists and developers collaborated to train IBM Watson running on the IBM Cloud to deliver meaningful insights on a daily basis in a way any fantasy owner can understand.

Implementing this into the app itself was no easy task. Aaron Baughman, principal data scientist on the IBM IX team, led the coding for Watson’s integration into the ESPN Fantasy app. The team rose to the unique challenge of making the analysis from all the news articles, videos and podcasts presentable in a useful and user friendly manner.

“To have a machine read and comprehend all the information, we had to go through different iterations of training. We had to do a lot of software development, but we also had to ensure that the system could comprehend all the different types of information and we used different types of AI techniques in order to do it through a novel pipeline.”

Baughman goes on to talk more about presenting the information to the user.

“We wanted to make sure that the insights could really help people. We used the evidential approach to put up stats and outputs from our AI system along with charts, graphs, and easily concealable numbers. For people that wanted to dig a little deeper and know why the AI predicted what it did, we would link to relevant articles/content from ESPN. We always wanted our story to be complementary to ESPN.”

When it came to marketing this offering from IBM and ESPN, the two took a unique approach with the help of agency MKTG. They assembled a group of influencers that included actor Jerry Ferrara of “Entourage” fame, ESPN fantasy football expert Field Yates, NFL veteran turned ESPN commentator Charles Woodson, sports journalist Bonnie Bernstein, co-host of ESPN’s “The Fantasy Show” Daniel Dopp, MIT Math PhD student and former NFL player John Urschel, and ESPN injury analyst Stephania Bell. The group entered into a fantasy football league together, which provided for some opportunities for social content both on IBM’s channels and on the participants’ social channels.

Over the course of the 2018 football season, which lasts roughly four months, IBM, and agencies MKTG, DDG, and Ogilvy worked tirelessly to create a substantial amount of editorial, social and advertising content, adding a healthy dose of personality to the campaign while educating consumers on the benefits of Watson. This was no easy feat, as MKTG Senior Manager of Digital Strategy Logan James will tell you.

“This was a massive program,” James says. “We had the influencers focus on provoking replies and comments with their audiences in order to stimulate conversation about the Watson tool. In the four-month long season, the influencers collectively posted over 250 times during the campaign. As you can imagine, there was a lot of content needed to support the campaign so it was important to optimize it for each influencer week-to-week based on how previous posts performed and how they were performing in the actual league.”

For all parties involved, this made perfect sense as MKTG SVP Gavin Blawie can attest.

“IBM understands the value of influencers and has for years,” Blawie says. “IBM has long been a proponent of adding credible third-party voices to help complement their official brand narratives. It really helps bring an additional social overlay to the traditional assets like television and signage that you typically get from sponsorships.”

In terms of choosing the influencers, IBM and MKTG chose names from around the world of pro football, fantasy football, analytics, and popular culture to offer several different unique points of view on utilizing Watson. All in all, the influencers enjoyed taking part in the league while gaining a whole new perspective on fantasy football from the Watson integration.

“This opened up my eyes to new ways to think about fantasy football,” says Yates. “What IBM allowed me to do was quantify risk. If I play someone with a high-bust percentage, I could also play someone with a high boom percentage and that way I felt like I was able to balance my lineup, especially in the early and middle parts of the season.”

Ferrara, a very early adopter of the fantasy football world, noted that he is always looking for a new competitive edge when it comes to the game that he has invested so many hours into over the years.

“In real-life football, data is king in a lot of ways, so it’s natural that it would be just as important in fantasy,” Ferrara said. “So when you get to see what Watson was able to do by analyzing all those articles, which no one in their right mind really has the time to do, and basically streamline all that relevant data in an easy way in the app, that was something that I relied on heavily in this league.”

IBM has been providing its sports and entertainment partners with valuable insight for years. This ESPN integration is just the latest example of the company creating new ways for partners to leverage cutting-edge tech innovation. This is the same core technology IBM brings to partners across business and industry around the world. As it continues to evolve, you can expect to see it in other areas of business and sport alike — including influencing your waiver-wire strategy for many seasons to come.

“By putting enterprise-grade AI in the hands of fantasy owners on the ESPN platform, we create value for ESPN by differentiating their platform, O’Brien adds. “We create value for fantasy owners by helping them make better decisions and we help people understand the value of Watson and IBM technology.”

(*MKTG is a Proud Partner of FOS)