Fantasy Football Fiction: The Gamma Equation
NOTE: This article is a piece of ‘fantasy football fiction’ that I thought would be fun to write. It depicts a dystopian future of superhuman analytics that I envisioned when thinking about the next steps for powerful prediction models.
I’m a pro-analytics guy, but I often have to remind myself that having human agency (and inevitably error) in the mix makes fantasy great. I hope you enjoy.
Fantasy football was solved in 2037.
It didn’t come as much of a surprise; really, you could say that fantasy never stood a chance. We all knew that analytics was the ‘way of the future’, that projections and models would sharpen our skills and improve our rosters. What we didn’t realize was how good the models would get. If we did, we might have done things differently.
The BallersBot was the first truly generational prediction engine. Built by a couple of enterprising data scientists and named after the Fantasy Footballers (who else) it made its debut in 2025. It sucked up all available data – player performance, Vegas lines, weather forecasts, the strength of schedule, you name it – and constructed a cutting-edge, boosted deep neural network to make predictions.
At first, the BallersBot didn’t get much love. Neural networks are completely uninterpretable, which simply meant that the BallersBot couldn’t explain any of its predictions in terms understandable to our meager human minds. As such, we scoffed at most of the predictions.
We should have been paying more attention.
The BallersBot did quite well in its first year of deployment. Because of the opaque nature of the model, no one really paid attention. Its handlers, busy now with other projects, weren’t interested in further development. For better or worse, someone else was.
Former Cowboys’ quarterback Tony Romo had been on a bit of a losing streak. Tom Brady had joined the broadcast booth after winning his 8th and 9th Super Bowls (one with the Bucs, and his final ring in a reunion tour with the New England Patriots). Everyone thought Romo was the ‘next best thing’ in football commentating, but Tom Terrific simply blew him out of the water with his astute analysis, surprisingly effective dry humor, and halftime phone calls to a hyped-up Gronk.
Desperate to stay relevant after once again being outshined by the GOAT, Tony cast about for another outlet. During a casual golf game with a retired Cris Collinsworth, the elder statesmen impressed Romo with his passion for the NFL Next Gen Stats project.
“Really?” said Romo, “I thought you just did those commercials to pay the bills.”
“No, Tony”, Collinsworth chuckled, “it’s the future, really. They’ve got it all in those computers, see. Now here’s a guy….”
That night, Tony Romo did some google searching and stumbled upon BallersBot. He was blown away by its ‘rookie season’ performance. “These guys don’t realize they’re sitting on a gold mine”, he muttered to himself. After a couple of emails with the founding data scientists, he agreed to buy the rights to BallersBot for a total of $1,000,000 (that’s $1,000 in 2022 dollars, after adjusting for inflation).
Now, it might not seem like Tony Romo would be the person to take an analytics model to the next level. In a way, he wasn’t. BallersBot was already the most advanced product on the market, and Romo didn’t contribute any code to improve it. What he did provide, though, was perspective.
Romo knew as much as anyone that football was an emotional game. The stats were important, sure, but you had to account for the mental and emotional state of the players. It was hard to predict when a player was feeling nervous, or confident, or had eaten so much buttered corn on the cob that he would fumble a routine place-kicking snap. He hired a team to build the ‘human side’ of football into the model.
It turns out that this was actually an easier prospect than once imagined. Pro athletes were increasingly online, sending tweets, posting TikToks and sharing Instagram stories about their workouts. All it took was for someone to scrape, collect and otherwise organize that data and pour it into the BallersBot for the model to crunch.
After all the code had run, Tony Romo waited for Week 1 with bated breath. The model had predicted zero points scored for Davis Mills, reigning NFL MVP of the Houston Texans. It didn’t make sense but, as the first Sunday game kicked off, word came from the Texans’ camp that Davis Mills was a surprise inactive for the game. The cause: some bad noodles from the night before. Romo’s model had picked up some discoloration in Mills’ pupils during his pre-game TikTok, diagnosed the condition, and made the (eventually correct) prediction.
Tony was ecstatic. All that was left was a name, and that wasn’t even all that hard. A lifelong fan of Bruce Banner and The Hulk, Tony dubbed his model “The Gamma Equation.”
It was a lucrative year for Romo and his team. He made headlines for his ‘ridiculous’ (and oddly specific) predictions on morning talk shows about the upcoming NFL week. His takes were so blazing hot that Skip Bayless and Stephen A. Smith entered a bidding war to bring Romo in as their new debate partner. Romo declined both, but it wasn’t all for naught. Skip and Stephen A. rekindled their chemistry during the negotiations and launched “Double Take”, the much-anticipated sequel to their original show. The “Lebron or MJ? Answer: NEITHER” marathon became the first sports program in history to garner 5 billion live viewers, a spectacular 20% of the Earth’s population.
What everyone thought was a goofy career change for Romo ultimately proved oddly prescient. Week in and week out, by hook or by crook, his oddball predictions came true. By Week 5, people were listening. By Week 9, Vegas was listening. Lines moved after Romo’s Sunday morning podcast aired, titled “Tony Talks and You Should Listen”.
After an electric Super Bowl – Tom Brady, who had returned to the NFL for a third time, tossed the winning TD to Trevor Lawrence, who had converted to the tight end position, to give the Jaguars their first ring – Tony Romo called a meeting with all of the major fantasy platforms. We don’t know exactly what happened in the room (ESPN sure isn’t talking) but Romo walked out with a private island in the Caribbean and the platforms walked out with The Gamma Equation in hand.
The rest, as they say, is history. Sleeper was the first to incorporate The Gamma Equation into its ADP, which resulted in an auto-draft so powerful that no mere mortal could defeat it. Sharp fantasy managers began noticing that following the advice of the computer gave them an edge, and the apps responded by automating every aspect of the game of fantasy. Who should you trade for? How much FAAB should you bid? Who do you start this week? Decisions were increasingly offloaded to the algorithm, and those who ceded control were rewarded.
The apps hired teams of data scientists to tweak the equation, and gradual progress was made. Those improvements began slowing, though, as the models got closer and closer to perfection; eventually, progress stalled altogether. Finally, in 2037, Benny Belichick, grandson of Bill Belichick and mathematics savant, formally proved that analytical dominance had been achieved. This seminal proof, which won the Fields Medal (awarded every four years to epic mathematical achievements) showed unequivocally that The Gamma Equation was now unbeatable in the realm of fantasy.
That brings us to modern times, where the fantasy landscape looks very different from what you’re used to. Draft day still excites, to be sure, but in a different fashion. Every manager switches their auto-draft ‘on’ and sits around watching the big board tick names off one by one. There’s cheering when you get a player you like and heckling when the model selects a player you don’t. The heckling is subdued, though: everyone understands that Gamma knows best.
In-season activity is similarly passive. Models optimize and automate every aspect of team management: sending/receiving trades, playing the waiver wire, and setting lineups. While there is an option to ‘manually override’, no one really does. That is, almost no one: there exists a small, fringe group of ‘Gamma-Deniers’, or GDs, who don’t believe Belichick’s proof and are confident they can beat Gamma. They invariably lose, which actually makes them quite popular to have as league-mates.
Tilting, and general uncertainty, have become the relics of an ancient past. There aren’t any upsets, chokes, surprises, or crazy, last-minute comebacks on Monday Night Football. Fantasy teams projected to win do, in fact, win; the script is set from Week 1 and faithfully followed all the way to the fantasy championship. Everything proceeds according to plan. The model misses nothing.
Now, don’t hear what I’m not saying. Fantasy football is still fun. While Gamma plays perfectly, even against itself, we can never be sure which super-human Gamma team will win the league. It all depends on the relative value distribution of players across the draft board in a given season; this gives a natural advantage to one of the twelve draft spots. Last year, in 2045, every league was won by Gamma teams drafting out of the 3rd position, thanks to a drop-off in talent after the (unanimous) top three players in the draftable universe.
Indeed, the draft draw – where draft positions are randomly determined – is now the most intense day of the fantasy year. Managers circle up and pick numbers out of a hat, cherishing that one remaining bit of randomness that can make the difference between a championship and a lost season. I know I’m hoping to land the 3rd spot this year. Lucky number three.
The above events are all (probably) fictional.