Methodology

1. Team power ratings

Each team gets a power rating in points versus an average team on a neutral field. A positive number means better than average; a negative number means worse. The rating is a point estimate, not a rank, so the difference between two teams can be turned into a spread.

v2, the older rating, used 2026 sportsbook win totals as its preseason starting point, reported publicly in September 2026, translated at 1.5 points per win above league average and split evenly between offense and defense. That starting point is only an input to the older rating. It is not the published v3 number.

As games are played, the v2 starting point gradually gives way to results. The results weight is:

Early games matter, but one unusual result does not rewrite a team.

2. Results and efficiency

The results portion of the older v2 rating is an even blend of two views of performance:

• Final-score margin: points scored minus points allowed, adjusted for the opponent.
• Play-by-play efficiency: opponent-adjusted expected points added, which looks at the value of individual plays rather than only the final score.

Each view supplies half of the results component. Ratings are re-centered so that the league average remains zero.

3. The player-value injury layer

Team ratings describe a healthy baseline. A separate player-value layer is applied for verified availability. The value is our estimate of the difference between the player and the replacement.

• Quarterbacks have the largest values because changing the quarterback can move a team more than most single-player absences.
• Other values depend on the player's role and the replacement. Unsettled statuses stay labeled unsettled.
• The injury layer is applied after the v3 blend below. If that ledger is not reconciled, the card says so and we do not claim a refreshed number.

4. Fair spread (v3)

Market first. v3 does not publish the old v2 number as the primary spread. It starts from the market rating, then moves part of the way toward the healthy v2 rating, then applies injuries.

• Home field is about 2 points, and 0 at a neutral site. Older cards may say the exact home-field figure they used that week. Those figures stay on the archive.
• Rest and travel are included only when we can verify them.
• We compare this number with the line on that week's named board. We do not swap in a different book's line after the fact, and we do not grade a DraftKings board as if it were Westgate.

There is no validated Best Bet probability threshold. A Best Bet is a label, not a proven win percentage.

5. Side, confidence, and Best Bet

The side and the label are different. Confidence and Best Bet do not replace the side.

• Side: the against-the-spread pick on the same board as the other opinions.
• Best Bet and confidence: how strongly that card labeled the side. They are not a probability cutoff.
• Historical passes: cards through the Thursday, Nov 5, 2026 Week 9 card could leave a game as Pass. Those passes stay on the archive. They are not rewritten into sides after the game.
• From Week 10 on: Vector publishes a side on every game. A missing side is a gap on the page, not a silent pass and not a made-up number.

A Saturday update is a separate, dated note. It does not erase the original card. Grades use the original line of record unless that Saturday note is itself the line that was live before kickoff and is shown separately.

6. Outside opinions

Other public opinions can be shown for context. They are not part of the rating, they do not create a Best Bet, and they stay on the board they were actually given. A private or paid opinion is not published here.

7. Trends

Historical trends can be mentioned as color. Trends never create a side or a Best Bet. Verified rest and travel belong in the spread, not in a trend argument added later.

8. Integrity and timing

• v2: adopted September 25, 2026, before Week 3 Sunday. It remains the older input inside v3. It is not the primary published number.
• v3: market + 0.35 × (healthy v2 − market), then injuries, with home field about 2 points and 0 at a neutral site.
• No probability claim: we do not publish a validated Best Bet win probability.
• Original cards stay: a finished pick stays graded at the line on its own board. We do not rerun a played game into a new side.
• No invented numbers: a missing line, injury, score, or card is shown as missing.
• Losses stay visible: a graded loss is a loss, including a game that has already been played.
• Entertainment and opinion only: nothing here is betting advice. 21+.