Mike Foss, writing for USA Today:

Why home-field advantage doesn’t really matter in the NFL Playoffs

During the regular season, home teams win 57% of the time.

Okay, so home field is a HUGE advantage in the regular season. The average team would win 9—10 games if they were all played at home. That’s wildcard-level play often and division-champion level play sometimes.

Come playoff time, home teams in the divisional round are 70-26 since 1990. That’s a 16% increase as compared to the regular season home team winning percentage. It’s an advantage, but not necessarily as great as many believe.

Wait, seriously? This guy says 73% odds of winning is an advantage that “doesn’t matter”? You might get close to 3:1 odds in Vegas.

[40] games were won by the visiting team in the past 100 wild card, divisional, and conference playoff matchups. Pretty much on trend with the regular season average.

… where home field is a HUGE advantage.

If you read the article you can find almost every kind of data science fallacy in the playbook. Foss cherry-picks data temporally (looking at only the last five years) and spatially (just looking at the NFC), focuses on anomalies, offers no confidence intervals, and shows a fundamental misunderstanding of basic statistics. “[73%] is a 16% increase [over 57%]” — are you kidding me?

It’s a good thing our schools don’t use USA Today in the classroom — they might “learn” something from Foss. Oh, crap, USA Today has a huge education program; my own school participated. Here’s an idea: assign students to debunk this drivel.

When I start my casino, Mike Foss is the first guy I’m inviting to play. But for now, I’ll take that bet!

@themikefoss 60% win rate over the last hundred games "doesn't really matter"? How about we go 100 $1000 bets on a 10-sided die; 1–6 I win.

— dhalperi (@dhalperi) January 8, 2015

Go Hawks!