By Matt Kelley (@Fantasy_Mansion)
Special to Yahoo Sports
Analytics movements continue to gain momentum across most sports. While baseball and basketball front office executives, fans, and fantasy gamers pushed into the outer reaches of advanced stats and metrics galaxy in 2017, football teams were still generally afraid to fly.
NFL teams often dismiss the value of sports analytics publicly, though many teams, including the New England Patriots, employ a math-savvy back office team to run the numbers on everything from salary cap ramifications to ticket prices. However, most NFL teams are reticent to discuss the role that evidenced-based analysis plays in player personnel decisions and in-game tactics. Given the resounding dismissal of “analytics” by the majority of NFL coaches, it remains safe to assume that math plays a negligible role in the on-field performances every football Sunday. SMH.
The Cleveland Browns instilled hope in a fast-tracked football analytics revolution when the team became the first NFL franchise to openly embrace sports analytics throughout the organization. Cleveland promoted Sashi Brown to executive vice president of football operations in 2016. Brown is a well-regarded proponent of a mathematics-based approach to front office decision making, but after less than two years in the role, he was fired (granted with mishaps under his watch). The Browns’ recent front office purge and seemingly backing of Hue Jackson serves as an ominous warning flair to the analytics community: Sports quants are not welcome at the highest levels of NFL front offices. Back to the shadows.
Math in Fantasy Football
While NFL franchises under-utilized analytical football analysis, fantasy enthusiasts leveraged advanced stats and metrics more than ever. For example, user traffic increased as the fantasy season went on at PlayerProfiler.com.
If Browns owner Jim Haslam abandoned analytics, why should fantasy owners trust mathematics-driven conclusions? Beyond increased interest, did PlayerProfiler’s evidenced-based analysis yield positive fantasy results?
Kamara > McCaffrey: CORRECT
In early August, Christian McCaffrey began his assault on fantasy draft boards. His average draft position (ADP) in Yahoo leagues ultimately climbed into the top-50 overall. Meanwhile, Alvin Kamara, a bigger, more explosive satellite back, wallowed in the double-digits rounds.
The advanced stats and metrics suggested Kamara would out-score McCaffrey all along, because situation is critical for running backs. Throughout his Saints career, Drew Brees has targeted running backs at the highest rate among his NFL contemporaries. Kamara was set up to exceed expectations by virtue of operating in a New Orleans offensive system that emphasizes running back touches and enhances running back efficiency.
|Stats & Metrics||Christian McCaffrey||Alvin Kamara|
|Fantasy Points Per Game||15.1||19.5|
Cam Newton, on the other hand, targeted running backs less than any current NFL quarterback throughout his career. Despite historic tendencies working against him, McCaffrey received a significantly higher Opportunity Share than Kamara this season, and it did not matter.
The more explosive Kamara better exploited higher quality touches in high-leverage situations. Kamara led the league in numerous running back efficiency metrics and became 2017’s signature league-winning, late-round RB pick.
McKinnon > Murray: CORRECT
In the wake of Dalvin Cook’s season-ending knee injury, in a not-so-shocking twist, PlayerProfiler.com’s workout metrics affirmed that athleticism matters in athletics.
Juxtaposing Jerick McKinnon and Latavius Murray’s advanced metrics illuminated an upside gulf. While most fantasy analysts pinpointed Murray as the back to own in Minnesota, more quantitative football enthusiasts heeded a passionate contrarian plea. Based on projected role, historic efficiency, and most importantly, athleticism, McKinnon was the back to own in Minnesota. Sure enough, McKinnon has since dominated Murray in both the efficiency and productivity departments when it came to PPR leagues.
Funchess > Benjamin: CORRECT
Once upon a time, NFL Draft analysts suggested that this year’s late-round smash hit receiver, Devin Funchess, should to convert to tight end. The analytics disagreed.
The most predictive stats on Funchess’ advance prospect profile suggested he was poised to usurp Kelvin Benjamin as the Panthers’ primary receiver this summer.
The intersection of key metrics from College Dominator Rating to Breakout Age to Speed Score comp’d Funchess to rolodex of elite receivers, including his best comparable player on PlayerProfiler.com: Brandon Marshall.
Get Marquise Goodwin: CORRECT
Analytics-minded fantasy gamers leveraged PlayerProfiler’s workout metrics to illuminate Goodwin’s extreme upside.
From Speed Score to Burst Score to Agility Score to Catch Radius, Goodwin is arguably the most athletic wide receiver in the NFL. He possessed the highest ceiling among San Francisco receivers and was finally paired with a quarterback who could get in him the ball consistently in Jimmy Garoppolo. Goodwin was the ideal waiver wire add heading into the fantasy playoffs.
Mariota > Watson: WRONG
We can’t ignore the misses.
Marcus Mariota finished 2016 as the most efficient quarterback in the NFL by some measurements. Then, the Titans invested more draft capital than any other team in the receiver position and signed free agent Eric Decker this offseason. The numbers suggested it was wheels up for Mariota this season. Yet, his 14.3 fantasy points per game currently ranks No. 22 among qualified NFL quarterbacks.
Meanwhile, Deshaun Watson shocked the football analytics community by becoming the first rookie quarterback in recent memory to finish No. 1 among NFL quarterbacks in fantasy points per game (24.3). Watson accomplished this astounding feat with truly uncommon (read: unsustainable?) efficiency.
|Air Yards Per Attempt||5.3||#1|
|Fantasy Points Per Dropback||0.67||#1|
|Receiver Target Separation||1.19||#28|
Watson’s exceptional rookie year was particularly impressive given his receivers’ general inability to separate. The Texans’ receiving corps averaged 1.19 yards of separation at target, which currently ranks No. 28 among NFL teams.
Watson also succeeded in spite of a weak throwing arm. This summer, his 49 MPH Throw Velocity (third percentile) became the crux of some of the most poorly conceived quantitative analysis in sports history. I vividly recall boldly suggested Watson would be better in the CFL. Though a top-10 quarterback with a sub-50 MPH throw velocity is unprecedented, the analysis was fundamentally flawed. Many quarterbacks do not throw at the NFL Scouting Combine, and there are less than 100 QB arm velocity values in the PlayerProfiler.com database. With a relatively small sample set in-hand, I over-emphasized the statistical significance and predictive quality of arm strength data. Sorry!
Fade Cooper Kupp: WRONG
Speaking of me being wrong, Kupp says, “Hi!” Coming out of college at age 24, I dismissed Kupp as a relatively ancient, small college compiler. Kupp laughed in my face when he became Jared Goff’s go-to receiver in the wake of Robert Woods’ shoulder injury.
Rather than focusing on Kupp’s age on draft day, I should have homed in on the two most predictive metrics on a wide receiver’s prospect profile. His 20.2 Breakout Age was above average and his 40.4-percent (80th percentile) College Dominator Rating was particularly impressive in the context of Eastern Washington’s pass-centric offense. Kupp ultimately finished his college career as the most productive receiver in NCAA history. I blew it.
Get Stefon Diggs: WRONG
While Kupp’s preseason forecast under-projected his full season production, most in-season forecasts over-projected Diggs production after he suffered a groin injury. Analytical models too often assume full health week to week and rarely discount injured players properly.
Diggs’ production collapsed from 17.3 to 11.4 fantasy points per game upon his return from a second groin injury in the past 12 months. Similarly, Leonard Fournette’s per game production dropped from 21.7 points per game to 13.3 points per game after his mid-season foot injury. When a skill position player with a well-documented injury history suffers a second injury to a previously-compromised foot, ankle, knee or groin, treat them as radioactive for the remainder of the season.
Beyond the quantitative analysis, pre- and post-injury splits also align with intuition. It just makes sense that if the foundation of a player’s speed and quickness is compromised, walk away in fantasy.