Week 2 Upset Watch
Football Outsiders reveals most likely upset of Week 2 in the NFL
Upset Watch appears every Thursday for ESPN Insider, as Football Outsiders uses a proprietary formula to forecast the expected point spread of each game based on current DVOA ratings (explained here) and, early in the season, our DVOA projections. Each week we highlight the most likely upset on a game with a line over three points, plus an additional game in which a significant underdog has a strong chance to cover.
Here's the problem with a feature called Upset Watch: Occasionally, you'll have a week in which your systems don't project any upsets. That's the surprising scenario we face in Week 2 of the 2013 season. Even though our algorithms tend to project closer "expected lines" in the first two weeks -- based on the idea that our knowledge of how good teams are isn't as strong as it will be later in the season -- we still ended up this week with all 16 favorites. That includes games with a line below three points, which we wouldn't write about in this column, anyway. In fact, this week's Football Outsiders premium picks picked just three out of 16 underdogs to cover, let alone win outright.
So we can't give you any games this week in which we feel an upset is more than 50 percent likely, but we can tell you the game in which an upset is most likely, and how the underdog can make that upset happen. Based on Football Outsiders stats, the most likely upsets of the week are:
• Dallas (+3) at Kansas City
• Miami (+3) at Indianapolis
• Minnesota (+6) at Chicago
• San Francisco (+3) at Seattle
• Tampa Bay (+3.5) vs. New Orleans
And the closest of those games is probably the one you would think is the closest: the NFL's newest, hottest rivalry.
To read more of which upsets are most likely to happen in NFL Week 2, become an Insider today.
MORE NFL HEADLINES
- Report: NFL eyes Pats locker room attendant
- Browner's targeting plans sadden Seahawks
- Pro Bowlers: Arizona missing vibes of Hawaii
- Carroll: Sherman comments just his opinion