NFL Power Rankings & Ratings — PWR
All 32 teams ranked 1–32 by PWR, the defense-adjusted power number, with tier
PWR = Off PR + Def PR + ST PR
Each component is DAEPA (opponent-adjusted EPA per play) scaled to points per game. A PWR of +7 means the team is expected to beat an average team by about a touchdown on a neutral field.
Model as of
| # | Team | PWR | Tier |
|---|---|---|---|
| 1 | SEA | +3.9 | Playoff |
| 2 | NE | +3.9 | Playoff |
| 3 | JAX | +3.2 | Playoff |
| 4 | SF | +2.8 | Playoff |
| 5 | LAR | +2.6 | Playoff |
| 6 | BUF | +2.4 | Playoff |
| 7 | CHI | +2.2 | Playoff |
| 8 | BAL | +1.8 | Playoff |
| 9 | CIN | +1.5 | Playoff |
| 10 | GB | +1.4 | Playoff |
| 11 | DEN | +1.3 | Playoff |
| 12 | DET | +0.8 | Average |
| 13 | HOU | +0.8 | Average |
| 14 | KC | +0.4 | Average |
| 15 | MIN | +0.0 | Average |
| 16 | ATL | +0.0 | Average |
| 17 | PHI | -0.1 | Average |
| 18 | PIT | -0.5 | Average |
| 19 | CAR | -0.6 | Average |
| 20 | TB | -0.6 | Average |
| 21 | IND | -0.7 | Average |
| 22 | DAL | -0.8 | Average |
| 23 | CLE | -0.9 | Average |
| 24 | NYG | -1.0 | Average |
| 25 | WAS | -1.8 | Average |
| 26 | NYJ | -2.3 | Below Avg |
| 27 | LAC | -2.6 | Below Avg |
| 28 | NO | -3.0 | Below Avg |
| 29 | LV | -3.2 | Below Avg |
| 30 | ARI | -4.4 | Below Avg |
| 31 | MIA | -4.6 | Below Avg |
| 32 | TEN | -4.9 | Below Avg |
PWR = Off PR + Def PR + ST PR — points per game above an average team on a neutral field. Season 2026, week 6.
Use the NFL Power Ratings inside your AI agent
This tool is also live on our MCP server as nfl_power_ratings. Connect Claude, ChatGPT, Cursor, or Copilot and call it directly — six core calculators are free.
Go Pro · 26% off your first year
Every edge engine on the site, live — $150/yr
Pro opens every engine on the site — the full silver, gold, oil and bitcoin strike grids that price commodity contracts in real time, Thee Oracle, Theta Edge, and the full Kalshi-vs-Polymarket price-gap board.
$111 for the first year, then $150/yr. Code applies automatically. Ends Dec 31. Monthly is $14.99/mo.
Every signal graded in public, losers left on the board. See the record →
Related Tools
Power ratings answer the question a record can't: how good is a team once you strip out schedule luck? PWR does it with Defense-Adjusted EPA — every play credited for the quality of the opponent — so a team that piled up yards against soft defenses doesn't get overrated, and a good team stuck in a tough division doesn't get punished.
Unlike a pure ranking, every PWR maps to a market. The same number that ranks the board also prices our win-total futures, championship odds, and weekly Kalshi game edges — so "is Baltimore overrated" becomes a testable claim against the market, not a talking point.
Frequently Asked Questions
What is PWR (the power rating)?
PWR is each team’s rating in points per game above an average team on a neutral field. A team at PWR +3.5 is expected to win by about 3.5 points on a neutral field against a PWR 0 opponent. PWR = Off PR + Def PR + ST PR — the same three buckets, all measured on the same points scale so they add directly. Each is built on Defense-Adjusted EPA (DAEPA), per-play efficiency adjusted for the quality of opponents faced, over 1999–2025 play-by-play. Tiers: Elite ≥+7, Contender ≥+4, Playoff ≥+1, Average ≥−2, Below Avg ≥−5, Rebuild <−5.
Are these NFL power rankings or power ratings?
Both. The board ranks all 32 teams 1–32 — that is the power ranking — and the number beside each team is the rating (PWR) the ranking is sorted by. Most published power rankings are an ordered list with no number; here the gap between #4 and #5 is a real points figure, which is what makes it usable against a Kalshi price.
How is this different from other public power ratings?
It is the same class of opponent-adjusted efficiency rating, but every number maps to a prediction-market contract: PWR drives our win-total futures, championship odds, and Kalshi game edges. The rating is DAEPA-native — built from public play-by-play — so it passes the "available elsewhere" test and is fully publishable.
How often do the ratings update?
Weekly during the season and monthly in the off-season, as new play-by-play and roster data land. In-season the board re-derives after each slate.
Can an AI agent pull these ratings?
Yes — the same board is the nfl_power_ratings MCP tool. An agent connected to the PredictionMarketsPicks MCP server can ask for the full board, one team, or a tier and get the ranked ratings back with attribution.
Get The 7 Oracles' daily edge — subscribe free
No spam. Unsubscribe anytime.