“I’m not a hater,” say people as a preface to an enraged explanation that’s meant to justify their unfiltered hatred.
Well, I’m actually not a hater. But, there happens to be one website that is objectively terrible at giving fantasy football start’em/sit’em advice, and it’s time you knew about their misguided excuses for predictions.
I realize that by making this information public, I’m throwing away a competitive advantage I hold over my fantasy football competitors. Yet, there’s a reason that I’d rather expose a fraudulent front for fickle fantasy advice, who disguise their novice assertions as expertise.
They screwed me over.
If I didn’t have beef with this website, I wouldn’t have known to look for them in the data, and I wouldn’t have written this article. But, I did. They say every beef comes with baggage, so here’s the brief synopsis.
Following a breakout performance in the wake of the Adrian Peterson scandal, I picked up the enigma that is Matt Asiata off free agency, and began mixing him into my lineup for the rest of the season (I had weak runningbacks in a twelve-man league). Each Sunday morning, I would base my game-time decisions on one expert and another collection of advisors. Matthew Berry was my go to pundit, and NFL.com was my go to website.
Last week, I detailed how incredibly average Matthew Berry is relative to other pundits. This week, it’s time to expose my other misguided mentors. NFL.com, from a statistical perspective, is about as bad as Jay Cutler in clutch-time: they’re both consistently awful. There’s actually one in the bunch whose actually performed moderately: Michael Fabiano. When we (literally) take him out of the equation, NFL.com’s predictive capabilities are even worse.
| Likelihood that NFL.com is below average | SD (from average rank/50th percentile) | SD (from avg. overall weighted score) | Percentile | Average Score | |
| All of NFL.com | 99.9866% | -3.82 | -3.87 | 19.5% | 55.71% |
| NFL.com, without Fabiano | 99.9998% | -4.73028229 | -4.61 | 6.8% | 54.76% |
[NOTE: To apply a relevant z-score, I had to assume that the pundits on NFL.com were identical, which in a literal sense, they are not. The fact that the majority of other professional pundits factor in data analysis into their predictions does, however, mean “gut-based pundits” are identical, in the sense that they simply make predictions based on their “gut” evaluations.]
Referring back to my situation from last year: the crux of the website’s poor predictions was the fact that NFL.com continuously overestimated Matt Asiata’s value. From a statistical point of view, the likely issue is that these analysts didn’t factor in the large amount of risk he (figuratively) carried: one week he’d run for three touchdowns, the next week he’d run for thirty yards and three fumbles. Always, following a phenomenal week from the Vikings’ backup running back, NFL.com would place him within the top twenty fantasy running backs. Every Sunday I’d listen to their advice, and every week that I lost, I’d spend Monday blaming my bad luck. The pattern should’ve been obvious, which I hope it is for you.
NFL.com is to fantasy football advice as the NFL is to suspensions: in both cases, the people speaking for the NFL have terrible judgment and make millions of dollars because they supposedly have great judgment.
I love the NFL, I just think their website needs some personnel improvements. Still, don’t call me a hater. That guy from NFL Redzone who yells at the audience about how you can watch every single game with the octo-box, that’s my kind of analysis. Even if their fantasy advice is subpar, the NFL’s still doing something right.
NOTE: I assessed NFL.com’s performance in the same way that I determined Matthew Berry’s extreme mediocrity: through a z-score to assess whether the hypothesis was statistically significant, and a percentile ranking to directly compare the “experts” against their peers.
Sconnie Sports Talk