Learn how to use Gridiron Edge to find betting opportunities and understand our model's predictions.
Edge is the difference between our model's prediction and the betting market line, measured in points. When our model thinks a game will have a different outcome than what sportsbooks are offering, that creates an "edge" or opportunity.
Example: If the market has Team A favored by 7 points, but our model thinks Team A should only be favored by 3 points, we have a 4-point edge on Team B (the underdog). This suggests Team B +7 points might be a good bet.
Our model calculates its own predicted point spread based on team power ratings. This Model Spread is compared against the Market Spread (what sportsbooks are offering) to find discrepancies.
Our prediction based on team strength ratings, calculated from offensive and defensive statistics. This is what we think the spread should be.
The actual betting line from sportsbooks. This is what you'd bet against. When our model disagrees significantly, there's an opportunity.
We categorize betting opportunities into three confidence tiers based on edge size:
These are our strongest recommendations. The model strongly disagrees with the market, suggesting significant value.
Good opportunities with solid model advantage. Still worth considering, but less confident than Tier A.
Lower confidence opportunities. Use with caution and consider other factors before betting.
Power Rating is a team's overall strength score combining offensive and defensive capabilities. Higher numbers indicate stronger teams.
How it's calculated: We analyze multiple statistics (yards per play, success rate, EPA, etc.) and combine them into offensive and defensive indices, which are then combined into an overall power rating.
Confidence score (0-1): Shows how reliable the rating is based on data quality. Higher confidence = more reliable predictions.
Start at the Current Slate page to see this week's games with model predictions and edge calculations.
Click any game to see detailed breakdowns including:
Use Browse Weeks to review historical data and see how model predictions performed.