This article is part of our NFL Picks series.
I always find NFL power rankings interesting. I do not use them in any sort of isolation to place bets or make picks in the Circa Million, but they provide interesting context for wagering markets. I was curious how the data websites valued teams relative to the betting markets defined by weekly point spreads and seasonal team win totals.
For this exercise I generated two different sets of power rankings. One is a composite from six publicly-available websites -- ESPN's Football Power Index (FPI), DVOA from FTN Fantasy, nERD from numberFire, Massey-Peabody from Unabated.com, the Makinen Power Ratings from VSiN and NFL Predictive Rankings from TeamRankings. I scaled all six to get them in the same range and centered around zero for a league average team, and then used a simple mean to generate a composite. There was not enormous variation between the sites, so giving extra to weight to DVOA, for instance, would not have drastically changed the results.
The other is an implied power rating I calculated using current bettable win totals at DraftKings. I ran each team through their remaining schedule and best fit the power ratings that get them to the DraftKings win total.
Here is each team ranked by their composite power, with implied power from the betting markets and the difference between the two.
What can we learn here? Well, if the betting market underrates a team relative to the advanced stats websites, it implies that the "market" underprices them. That could manifest in either the win total being too low, weekly point spreads on the team being too low or a mix of both.
Here are the most underrated teams.
The Jets really stand out here. DraftKings pegs them for 6.5 wins with the odds juiced -150 to the over and +125 on the under, which gets the equilibrium to about 6.8 wins. The Jets sit at 3-3 heading into their bye week, so they need to just go 4-7 against a middling schedule to get there. ESPN FPI pegs them for eight wins. I love the over here and would like any point spread that prices them at this level. Following the bye, Gang Green plays my reeling Giants in Week 8. The Jets at -3 on the look-ahead line is low in my opinion, though there is more value on the season over.
As for some of these other teams, Miami's crazy-good offense has possibly overstated their level a bit, as it has come against some terrible defenses. The data websites may need a couple more weeks to fully adjust for opponents. The Colts and Vikings have injury issues, which leads to a disparity between the advanced stats that run on past data and the forward-looking betting markets which better reflect the team moving forward.
My second favorite win total over here is the Browns, as they perhaps are in the opposite situation. Deshaun Watson has missed two of their five games, but he is expected back as soon as this week. DraftKings has them at 9.5, shaded -140 to the over and +115 on the under.
Next comes the potentially overvalued teams.
The Bengals are similar to the Browns in that the backwards looking data reflects a hobbled Joe Burrow, so it makes perfect sense that the "market" sees better play going forward. The difference is that the Browns are priced in the opposite direction.
The Falcons have no such excuse, as they are relatively healthy. DraftKings prices them at 8.5 wins -- modestly juiced to the over -- and FPI says that sounds just right as they predict 8.4 wins. The advanced stats websites consider them a below-average team, but they sit at 3-3 and don't have a daunting schedule ahead, so .500 sounds realistic. The play here is perhaps that the point spreads overvalue Atlanta. They are 2.5-point underdogs at Tampa Bay this week and perhaps there is some value on the Buccaneers.
