Fantasy Football Week #1 Recap: Your Machine Learning Cheat Sheet

September 12, 2019 Evan Sakrison

Football season couldn’t come fast enough, and now you can’t wait to pummel your office buddies with a stacked roster under your mildly inappropriate team name. However, the competition is fierce this year, and you don’t want to lose to the clueless old guy who drafted the Redskins’ defense because “Josh Norman was great on Dancing With the Stars.” Lucky for you, we have a solution; one we’ve spent the past 4 years creating.

Meet your new fantasy football guru: our DarwinTM automated model building solution. It leverages a neuroevolutionary process to help you find an ideal model that’s uniquely tailored to your data sets. Gone are the days of detailed spreadsheets and praying for tornadoes to rip through the stadium whenever New England has the ball—Darwin’s got your back this season!

     ​

Sports!

How It Works

Basically, Darwin will predict how many points (standard, non-PPR) every player is going to score during the 2019 NFL season. We trained the model on historical player data from 2002-2017 as well as player age, draft position, strength of schedule, and many other variables to make predictions. We will compare Darwin’s predictions to fantasydata.com’s projections throughout the season, so be sure to tune in every Wednesday for the latest!

*Note1: We have to train the model on new data for the first few weeks, so Darwin won’t be able to make weekly predictions just yet. Until then, we will use “Expected Fantasy Production” from the Full Report, which is the projected number of points a player is expected to score during the 2019 season.
*Note2: Rookies were not included in the model building process.
*Note3: Players that did not play in 2018 were not used to create predictions in 2019 (i.e. Le’Veon Bell).
*Note4: While the model was also applied to Defense & Special Teams, we did not include projected fantasy points in the full report, and thus will not cover Defense & Special Teams in this blog series.
Download the Full Report here!

Top 5 Players of the Week in Each Position

Quarterbacks (QB)

No matter what your league’s designated Stats Guy says, anything can happen in Week One. Anything. Not even the Mayans could have predicted Andrew Luck’s sudden retirement, nor that Lamar Jackson would throw five touchdown passes in the season opener, posting the best passer rating (158.3) in Ravens history. But then again, his opponent was the Miami Dolphins defense, proving that no one special has taken their talents to South Beach since the Obama administration.

As the season progresses, we’ll learn whether Darwin’s predictions are on point. According to the full report, both Deshaun Watson and Patrick Mahomes are slated to finish as top five QBs this season, meaning you should start them every week. Lamar Jackson, on the other hand, can be found closer towards the bottom of the chart, indicating he’s a solid backup when your first-string QB has a bye week. However, given his stellar performance against the Dolphins on Sunday, keep a close eye on him in the coming weeks when he plays less pathetic defenses.

Running Backs (RB)

Once again, Week One is chock-full of surprises. As Mel’Veon Bell… er, Melvin Gordon continues to sort through potential trade options, Austin Ekeler will share the backfield with Justin Jackson. His presence paid off because he absolutely went off on the Indianapolis Colts defense. Dalvin Cook looked sharp, too, dominating the Atlanta Falcons with 111 yards and two rushing touchdowns.

Darwin pits Christian McCaffrey as the only running back in this week’s top five to put up big numbers a la Todd Gurley and Saquon Barkley this season. While Gurley and Barkley had average-to-subpar performances last weekend, they have far more fantasy value compared to everyone else in this week’s top five. It all depends on whether Dalvin Cook can stay healthy after last season’s injury woes and if Ekeler can remain a dual-threat on offense. 

Wide Receivers (WR)

While not nearly as fragile as Darren McFadden, Sammy Watkins is still prone to injuries. But miraculously, he torched the Jacksonville Jaguars on Sunday without incident, racking up 198 yards and three touchdowns. We can’t say the same for his teammate, Chiefs’ top WR option Tyreek Hill, who’s expected to miss multiple weeks due to an unexpected shoulder injury. Darwin predicted Hill would score more points than any other wide receiver this season, but sadly it couldn’t predict a random sternoclavicular joint injury on a hook route at the 49-yard line.

As for the rest of the top five this week, DeAndre Hopkins is always a must-start receiver. Darwin predicts he’ll finish fourth amongst wide receivers, assuming he doesn’t suffer the same fate as Tyreek Hill. Speaking of injuries, John Ross will only be somewhat useful until A.J. Green no longer has to waddle around in a walking boot. Phillip Dorsett may be at a worse disadvantage after the Patriots shoehorned Antonio Brown into the lineup one day before the season opener, so let’s attribute his Week One performance to good luck. If you have Watkins in your lineup, it’s probably safe to start him as he should continue to put up decent numbers while Hill remains on the sidelines.

Tight Ends (TE)

Pulling double duty on offense, tight ends aren’t wholly exciting unless they are mostly used as receivers. Otherwise, they are stone walls with cleats, joining their fellow linemen to protect the backfield from getting throttled by 250lbs of solid muscle. Of course, we can’t knock on good performances by professional athletes, but according to Darwin, Evan Engram is the only worthy player on this week’s top five. He’s a solid second-tier tight end outside of the usual suspects (i.e., Travis Kelce, Zach Ertz, etc).

In a non-PPR league, you can’t expect these guys to pull off what Sammy Watkins did this week, but that’s not to say they won’t be the deciding factor when you’re neck-and-neck with Janet from accounting.

Overall, how’s Darwin doing?

Week One was wild with a ridiculous number of outliers, so we can’t say anything with confidence yet. We’ll have to wait until we get into the groove of the season to determine whether Darwin’s top five in each position are fantasy studs or fantasy duds. As we feed Darwin new data each week, it will be able to generate weekly, well-rounded predictions. Until then, please refer to “Expected Fantasy Production” in the full report to help you make game day decisions. See you next week!

Ready to experiment with the Darwin Fantasy Football model yourself?

Sign up for the free 15-day Darwin trial now!

 

 

Previous Article
Fantasy Football Week #2 Recap: Your Machine Learning Cheat Sheet
Fantasy Football Week #2 Recap: Your Machine Learning Cheat Sheet

We hope your fantasy league's first week was spectacular. If not, don’t lose hope; even the 2010 Seattle Se...

Next Article
Quick Wins for a Short Month
Quick Wins for a Short Month

February is a month with limited time and a lot to accomplish to keep momentum going on goals of the new y...

×

Curated AI News Straight to Your Inbox Every Month. Subscribe to our Newsletter.

First Name
Last Name
Company Name
!
Thank you!
Error - something went wrong!