New site. Forecasts become public on this site from Monday, November 16, 2026. Earlier forecasts come from our private, time-stamped logs and are marked as such.

How the ratings and forecasts work.

Each sport has one team rating and one forecast model. We publish team numbers only. We never republish another provider's numbers or any player values. Sportsbook lines appear on pages for comparison only, and never in our data files.

College basketball: Power

Power rates each team in points per 100 possessions against an average Division I team. It starts from a preseason roster estimate, then adds this season's games, with recent games weighted more and part of the shooting luck removed. Game forecasts use the two ratings, the expected pace and a 3.5-point home advantage.

Scale
Points per 100 possessions against the average Division I team
Home advantage
3.50 points. Added to the home team in forecasts. Not used at neutral sites.
Updates
Daily, about 9:05 AM PT
Data
Injuries
Our forecast does not include injuries or late lineup news.
Power
Expected margin per 100 possessions against an average Division I team: the preseason estimate plus this season's games, with recent games weighted more and part of the luck removed.
Performance
What the team did this season only: no preseason estimate, the same luck removal, every game equal, adjusted for opponents. Early in the season it rests on few games and stays near average.
Luck
Points per 100 possessions that the team's results owe to three-point and free-throw luck. Positive means lucky.

NBA: Strength

Our own team rating, built from luck-adjusted plus-minus ratings weighted by each team's projected minutes. Strength counts every player who is not out long-term. A game forecast uses the official injury report 3 hours before tip: it is the gap of the two teams' values with that report, plus 2.0 points for the home team.

Scale
Points per game against an average NBA team on a neutral court
Home advantage
2.00 points. Added to the home team in forecasts. Not used at neutral sites.
Updates
Strength: each export, from the newest game data (each morning) and the newest injury reports. Tonight: when the team's next game forecast issues, 3 hours before tip.
Data
Built from NBA.com box scores and play-by-play.
Injuries
Preview from current team strength; the official forecast 3 hours before tip uses the injury report.
Strength
Full strength: the expected margin against an average NBA team on a neutral court, with every player available except long-term absences. A long-term absence is a player who is out in each of the team's last 4 forecasts, or out after surgery, with a rupture or for the season. Our minutes model gives the minutes.
With tonight's injury report
The same rating with the official injury report and the projected minutes of the team's newest game forecast. Our game forecasts use this value.
Offense
The part of full strength from scoring. Offense + Defense = Strength.
Defense
The part of full strength from defense. Higher means fewer points allowed.

College football: Scores-only ratings

Team ratings and game forecasts come from our scores-only rating model. It updates each team's rating after every game from the final score and the game site, and it starts each season from team talent (recruiting) and returning production. All of its data comes from CollegeFootballData.com.

Scale
Points against an average FBS team on a neutral field
Home advantage
2.38 points. The model's fitted home edge. It is added to the home team, but not at neutral sites.
Updates
Weekly, Monday about 8 AM PT
Data
Data provided by CollegeFootballData.com.
Injuries
Our forecast does not include injuries or quarterback changes.
Rating
Expected margin in points against an average FBS team on a neutral field. It comes from final scores, adjusted for the opponent, with season starts from talent and returning production.

NFL: Rating

Our house rating. It combines several rating systems built from game results and play-by-play data with the expected starting quarterback, the official injury report and the roster. The rating is the expected margin in points against an average NFL team on a neutral field, at full strength.

Scale
Points against an average NFL team on a neutral field
Home advantage
1.62 points. Added to the home team in forecasts. The value changes a little by week of the season. At a neutral site the model adds 1.11 points to the listed home team (a fitted term), not 1.62.
Updates
Daily, about 6:10 AM PT, August to February
Data
Data from nflverse, CC BY 4.0.
Injuries
Our line uses the official injury report (game status) and the depth chart.
Rating
Full strength: the expected margin in points against an average team on a neutral field, with every player available except players on reserve lists (injured reserve, PUP, NFI, suspension). The quarterback first on the depth chart starts. A player on this week's injury report counts as an available regular.
With this week's injury report
The same rating with the official injury report (game status) and the depth chart for the week 10 game. Our week 10 forecasts use this value. A team on a bye uses its next game.

How we grade

For each game we keep the last forecast made before the start. After the final score, we compare its margin with the real margin, and we do the same for the closing line. The closing line is the last line we recorded before the start. We also measure if the line moved toward our number after we logged it.

A forecast is never changed after it is published. A model update adds a new entry with its own time stamp.