What Are Power Ratings and Why Build Your Own
Building sports betting power ratings is one of the most practical steps a serious bettor can take toward developing a systematic handicapping process. Power ratings are numerical team strength estimates that, when combined with home field adjustments, produce a predicted point spread for any matchup. Your predicted spread can then be compared to the sportsbook line to identify potential value.
The goal isn't to build a rating system that's more accurate than Pinnacle's consensus market—that's nearly impossible. The goal is to build a rating system that captures angles the market doesn't fully price and that reflects your specific handicapping philosophy.
Simple Power Rating Methodology for NFL
Start with the simplest possible approach and build from there:
Step 1: Establish a baseline. Assign every team a rating of 0 at the start of the season. From week to week, you'll update based on performance.
Step 2: Update after each game. Compare actual performance to expected performance. A simple update rule: if a team was expected to win by 7 and won by 14, their rating increases. If they were expected to cover and didn't, their rating decreases.
Step 3: Determine a home field adjustment. The standard NFL home field adjustment is approximately +3 points for the home team. This can be refined by stadium type (dome vs. outdoor, altitude, noise) but 3 points is a reasonable starting point.
Step 4: Generate predicted spreads. Subtract the visiting team's rating from the home team's rating, then add 3 for home field. The result is your predicted spread. If your model says Team A is a -6.5 favorite and the book opens at -4, that's a potential bet.
Adjusting for Rest and Travel
Raw team ratings don't account for situational factors that affect performance:
- Rest advantage: A team coming off a bye week typically performs 1–2 points better than their base rating. Apply a +1.5 adjustment for the bye team when matching up against a team playing on normal schedule.
- Short rest disadvantage: Teams playing Thursday night games on short rest show performance decrements. Adjust by −1 to −1.5 points.
- Travel distance: For extreme travel situations (East Coast team to West Coast for a 10 AM Pacific kick), apply a −0.5 to −1 point adjustment to the traveling team.
These adjustments are estimates, not precise calculations. They improve model accuracy on the margin and represent situational knowledge the market doesn't always fully price.
Building NBA Power Ratings
NBA power ratings follow a similar structure but update more frequently—daily game data provides faster feedback than weekly NFL games:
Offensive and defensive ratings: Use points per 100 possessions (offensive rating and defensive rating) as your team quality metric. These pace-adjusted metrics are available free from Basketball Reference and NBA.com.
Net rating as baseline: Net rating (offensive rating minus defensive rating) is the simplest single-number team quality proxy. League average net rating is 0 by definition, so team ratings are immediately interpretable.
Home court adjustment: Apply +2.5 to +3 points for the home team's net rating advantage.
Predicted spread: (Home team net rating + 3) − Away team net rating = predicted spread in points per game, scaled to match typical game scoring (divide by ~10 possessions per game if needed).
Calibrating Against Market Lines
Once your system produces predicted spreads, compare them to opening and closing market lines over a full season. Track where your model diverges significantly from the market and whether those divergences predict winners. If your model consistently says one direction and the market goes another—and you're right more than 55% of the time—you've identified real edge.
Be skeptical of early results. Small samples are noisy. You need 200+ predictions before trusting a win rate differential.
Building power ratings gives you a systematic framework for every bet, but the analysis only compounds in value if you're tracking your actual results. Oddible (oddible.ai) records every bet and provides performance analytics that let you evaluate your power ratings model against real outcomes. Start tracking at oddible.ai.
