Boom or Bust: My fantasy football season with the help of Watson
By Betsy Schaefer | 5 minute read | September 25, 2019
As a native Ohioan, I have college football in my blood. I was raised an Ohio State fan, and the fact that I’ve lived in the suburbs of New York City for years will never change that allegiance. But this fall, it won’t only be on Saturdays that I have one eye on the TV.
Why? Because this year, I’ve got fantasy football fever. I’ll be playing in an ESPN Fantasy Football league for the very first time, and, with a little help from my colleague Ben Gardner, who’s given me a crash course in how fantasy football works, my players have been drafted and are already dominating on the field.
But Ben’s not the only one helping me out this season. I’m also getting input from IBM Watson, the leading AI for enterprise, delivered via the ESPN Fantasy Football app.
Fantasy Football Insights with Watson scours millions of unstructured data sources — like news feeds and analyst and fan chatter — to deliver contextually relevant insights on players. Armed with this information, I can decide who should play and who might need to ride the bench in a given week. Fantasy Football Insights with Watson is revolutionizing the way people experience fantasy football, and it’s a fun and exciting example of how the right AI can augment human intelligence to help us make better decisions.
In years previous past, ESPN would provide a projected score for each player to help users set their lineups. But what if I’m torn between two players with similar projected scores? That’s where Watson comes in. With its ability to digest unstructured data and produce relevant, actionable insights, Watson can predict the likelihood a player will “boom” or “bust” in a given matchup. So for example, if I’m up against a tough opponent and I need to take a risk, Watson’s insights will help me decide which of my players has the best prospects of having a big week.
Alternatively, if I’m feeling good about my game week and I want to play more conservatively, Watson can tell me which player has the highest “floor” – and is therefore least likely to be a bust in his matchup.
The 2003 book Moneyball depicts the true story of a forward-thinking baseball general manager named Billy Beane who hired a numbers-crunching Yale grad to perform statistical analysis on players. This enabled him to spot trends in player-performance that his eyes and intuition wouldn’t have been able to spot and other franchises were missing. He was able to build teams made up of players who were high value in proportion to their low cost on the market. The right understanding of data, coupled with the right human knowledge and experience, helped Beane lead the Oakland A’s to an unexpectedly successful run, including a 20-game winning streak in one season.
Fantasy Football Insights with Watson represents the next evolution of this trend, only now team managers have access not only to performance data, but also to unstructured data, allowing them to complement their football smarts with insights from data. This technology is powered by machine learning techniques driven by IBM Watson Discovery and augmented by anti-bias techniques in IBM Watson OpenScale, both running on the IBM Cloud. Where the average fantasy football player takes in fewer than four sources of data before making decisions, Watson is trained on millions of unstructured data sources — from articles, podcasts, video commentary, and beyond.
IBM Master Inventor Aaron Baughman and a team of researchers trained Watson to analyze these sources using the latest in natural language processing. Watson evaluates media sentiment toward each player, providing a range of potential upside and downside on ESPN’s existing player analysis. Then it assigns the “boom” or “bust” designation based on the player’s likelihood of exceeding or falling short of his projected scoring range. Users can then compare player data to help them balance risk and reward across their rosters.
All of this data is presented in the app with simplicity. Busy users or those new to fantasy football will appreciate the app’s clear recommendations. While dedicated fantasy football managers and true data geeks have been able to dive into the statistics Watson uses to draw conclusions, until now, managers haven’t been able to mine insights from unstructured data and have instead had to rely largely on their own intuition or best guesses. Watson contextualizes mountains of data in an unbiased way, changing the game for pros and new players alike.
Watson also captures the sentiment of player commentary in stories across the internet, and gives a percentage breakdown on which are positive and which are negative. This gives insights into how others view player trends and matchups, providing another illuminating data point to include in the decision-making process.
I may be a football superfan, but I’m also a busy working mom, so Fantasy Football Insights with Watson is the perfect tool for me. It enables me to make smarter decisions, and when I have time, allows me an easy way to geek out on all the football data I want. Ultimately, I’m the decision-maker. I have access to this wealth of information, but I bring my own insight and knowledge to the table, which I can use in deciding how I want to act on the tool’s recommendations.
Fantasy football makes me think of the ways we structure and build our teams in business, slotting highly skilled and specialized people in roles that suit their abilities and interests. It’s what I love about business, and what I love about team sports. Access to data that helps me make those decisions makes it even more fun.
ESPN Fantasy Football Insights with Watson is, in many ways, a glimpse into the future of human decision making. In the age of AI, machine intelligence will serve as a vital supplement to human intelligence. This technology promises to spread through every industry, from medicine, to law, to transportation — virtually every human endeavor you can imagine. Over 10 million people are using the app, but the true promise of this technology will continue to grow. How exciting it is to be part of this revolution at the ground floor.