IBM and ESPN take fantasy football to the next level with Watson AI

The brains behind ESPN Fantasy Football and IBM Watson sat down to discuss whether AI can really win you a championship.

CES 2019: How IBM Watson uses AI to uncover hidden bias General Manager of IBM Watson business applications Inhi Cho Suh discusses how insufficient data, IoT devices, incorrect values, and other factors can influence machine-generated bias within AI.

Millions of ESPN Fantasy Football players are rolling into week 13 of the NFL hoping to survive injuries and bye weeks on their way to the playoffs. But starting last season, players could also count on the trusty advice of Watson, IBM's artificial intelligence platform, in times of need.

ESPN's Daniel Dopp, co-host of "The Fantasy Show with Matthew Berry" on ESPN+, sat down with IBM Master Inventor Aaron Baughman at the IBM Innovation Lab in New York City to discuss Watson's foray into the arcane science of fantasy football.

"It helps to have a compromise between the heart and the brain. We trained Watson on millions of fantasy football stories, blog posts and videos. We taught it to develop a scoring range for thousands of players with their upsides and their downsides. And we taught it to estimate the chances a player will exceed their upside or fall below the downside," Baughman said.

On the ESPN Fantasy Football app, you can now click on a player's name and find an entire section dedicated to insights from IBM Watson. The AI program gives users easy-to-read projections as well as a "player buzz" section that surveys the general sports news landscape to see whether positive or negative news is being released about a player.

It even lets you compare your player to other available players based on boom or bust projections.

"Whenever someone uses Watson, they become more informed about who to start and who to play, more than specific point scores because their perspective opens," Baughman added. "Podcasts, videos and increased context around players help, and I've found that people become fans of players that they normally wouldn't have because they see more things about a player."

SEE: Artificial intelligence: A business leader's guide (free PDF) (TechRepublic Premium)

Dopp said ESPN was eager to partner with IBM and Watson because the fantasy football audience, which he estimated to be around 10 million people, was a sizable chunk of ESPN's core readership. Fantasy football players, out of necessity, come back to the website repeatedly for any content related to the digital sport, including tips from experts or advance warning about injuries.

Readers, he said, couldn't consume enough fantasy football content, yet were eager for even more tools to help them decide on which players to use during specific weeks.

"We want to give our viewers the most accurate information. Watson enhances the product for our users and we use it on a weekly basis. It sets our users up for as much success as possible," Dopp said.

In a video released in September, Vice President of Sports & Entertainment Partnerships at IBM Noah Syken said Watson was tailor made to solve the problems fantasy football players were facing, namely the inability to spend time consuming NFL content from every team on every player. 

"Fantasy Football generated a huge volume of content, articles, blogs, videos, podcasts. We call it unstructured data or data that doesn't fit neatly into spreadsheets or database. Watson was built to analyze that kind of information and turn it into usable insights," Syken said.

Before they could even put Watson to the task of analyzing current NFL data, Baughman said they had to teach it the football basics. They fed Watson over six million documents worth of football data that was pulled from a variety of sources.

They said Watson has ingested over 90 gigabytes of unstructured text from historical fantasy football seasons and troves of information from football encyclopedias.

Baughman told the eager audience that a team of about six to eight IBM annotators, data scientists and developers taught Watson the basics of both NFL football and fantasy football using Watson Knowledge Studio.

Once it understood the game, they could start the process of teaching it to read current articles and blog posts as well as pull data from videos and podcasts. It now performs 1,000 simulations for the top 400 players every day.

Watson focuses primarily on four main topics: players who will boom, players who will bust, players who are injured and players who aren't. ESPN Fantasy Football players can now count on Watson to pull from thousands of news sources for almost two billion insights each day on the more than 400 players available each week.  

According to Baughman, fantasy football players are sophisticated enough to know that Watson's insights should be purely supplemental to a player's innate knowledge of the sport. He said he hopes people are able to use Watson's insights to win their league championship, but that users should take a more holistic view of its value. Often, he added, Watson took out the emotions that might be tied to certain players and approached it with cold hard facts.

"Historical data can often be biased. If someone is having a few good weeks it may hide that historical data may weigh negatively in certain situations," Baughman said. "Sometimes it's good to start out fresh and not have as much baggage."

Most of Watson's fantasy football functions run on the IBM Cloud and are driven on the back end by IBM Watson Discovery and IBM Watson OpenScale. Baughman was one of the brains behind other sports-related uses for Watson like The Masters for golf and the US Open for tennis. 

He said they planned to continue playing with the algorithm and adding features, like a tool that would suggest players you should trade, acquire or sign from free agency. In an interview after the session, Baughman said they also planned to use it for events like the Grammy's to predict winners and to provide real-time red carpet facts about stars.

"The ESPN-IBM partnership is a great vehicle to demonstrate the power of enterprise AI to millions of people," Syken added. 

"It's not hard to see how the same technology applies to real life. There are thousands of business scenarios where you're assessing value and making tradeoffs. This is what the future of decision making is going to look like. Man and machine working together, assessing risk and reward, working through difficult decisions."

Also see

IT leader's guide to deep learning (TechRepublic)
Telemedicine, AI, and deep learning are revolutionizing healthcare (TechRepublic download)
Managing AI and ML in the enterprise 2019: Tech leaders expect more difficulty than previous IT projects (TechRepublic Premium)
What is AI? Everything you need to know about Artificial Intelligence (ZDNet)
6 ways to delete yourself from the internet (CNET)
Artificial Intelligence: More must-read coverage (TechRepublic on Flipboard)

American football field 50 yard line Friday night Lights

Image: Getty Images/iStockphoto