Laughing at Attempt to Predict 2013 NFL Draft With Social Media Algorithm
Since I laughed at iProspect‘s attempt to correctly predict the top 10 picks of the 2013 NFL Draft with a social media algorithm, along with Mashable’s misguided review of its success in 2012, I figured it only right to revisit this after the smoke had cleared. Only fair.
“But before you scoff at how accurate the methodology could be, consider this: iProspect ran a test using mentions of last year’s top prospects in the weeks leading up to the 2012 NFL Draft,” the Mashable review read earlier this week. “Those social predictions, the company says, correctly placed six of the top 10 picks, with the remaining four projections each being off by just one place.”
I can go on about how predestined the top five was last year and how this six out of 10 is little accomplishment, but I’ll stick to how well the algorithm did this year: zero out of 10.
Yup, couldn’t even hit one.
Score one for the NFL draft’s unpredictability:
1) Luke Joeckel, Kansas City Chiefs
2) Geno Smith, Jacksonville Jaguars
3) Dee Milliner, Oakland Raiders
4) Dion Jordan, Philadelphia Eagles
5) Eric Fisher, Detroit Lions
6) Lane Johnson, Cleveland Browns — wrong (and hilariously misguided)
7) Chance Warmack, Arizona Cardinals
8) Jarvis Jones, Buffalo Bills
9) Barkevious Mingo, New York Jets
10) Star Lotulelei, Tennessee Titans
A commenter on the Mashable review made an interesting but flawed counterpoint, “Hardly rocket science. Sports fans repeat on social media what they hear on talk radio and on ESPN. So it’s second-hand information from analysts who earn millions of dollars to be right about this stuff.”
Only problem here is that talk radio, ESPN and the many, many, many other outlets that offer expert opinions hardly came to a consensus on the top 10. Teams drafting up top are less inclined to negotiate contracts well before the draft due to the rookie wage limitations implemented in 2010, which throws a giant wrench in this whole prediction thing anyways.
Now, I wouldn’t go back and make fun of anyone’s mock draft. There’s a general acceptance that this stuff is just too hard to predict. I also like the idea of using social media metrics in a wide range of sports analysis.
However, the insistence that populist fan opinion can predict something like the NFL draft, along with the smug reaction to last year’s misleading results, just wreaks of a group of social media editors and marketers stepping out of their lane.