🔒 There's a pick for this game. All-Access shows it in full, free this season.
2026 · Week 5
StatGrades AI Picks
A call on every game against the spread and the total, made from the lines at every sportsbook, both teams' play-by-play efficiency, injuries, the weather, rest and travel, and a fresh search of the latest reporting, then graded against the final score. Every pick is labelled with what made it, and shows what it read.
These picks are for information and entertainment, not financial advice, and no pick is a sure thing. StatGrades doesn't take bets or link you to a sportsbook account. Odds are shown for information and can change at any moment; check with the book before you bet. 21+ only where legal. Gambling problem? Call 1-800-GAMBLER.
From when the lines open after the weekend. The AI's first read, re-checked as the line and the news move, so it can still switch sides. Not final.
From 6 PM ET on the Wednesday before the game. This is the pick, at its line, and the one the record grades. After that it changes only if key injury news lands, and the card says what changed.
StatGrades AI record
Every pick becomes official at 6 PM ET on the Wednesday before its game and is graded at its official line against the final score. After that only key injury news changes it, and the card says what changed. About 52.4% against the spread breaks even at standard -110 prices. The published record starts with the Sunday, September 27 slate.
Top calls
6-7-1
46% of decided picks
Against the spread
4-11-1
27% of decided picks
Over-under
8-8
50% of decided picks
Straight up
9-7
56% of decided picks
StatGrades Formula record
Every game. A simple ratings formula, not AI: last season's results blended with this season's play-by-play efficiency, plus home field, starting-QB injuries and wind, compared with the market line. Frozen at kickoff and graded like any pick. About 52.4% against the spread breaks even at -110. The published record starts with the Sunday, September 27 slate.
Top calls
2-2
50% of decided picks
Against the spread
9-7
56% of decided picks
Over-under
7-9
44% of decided picks
Straight up
9-7
56% of decided picks
Against the closing number
Whether the market moved toward a pick after it was published, measured against the number the game actually closed at. A win-loss record takes seasons to mean anything at sixteen games a week; this says something after a few weeks, because it compares the pick with the market's own final opinion rather than with one bounce of the ball. Positive is good.
Spread
−0.03 pts
market moved our way on 1 of 16
Total
−0.03 pts
market moved our way on 2 of 16
The AI grades itself
After every game the AI reads its own locked pick next to the play-by-play and says whether the game went the way its reasoning said it would. A win can be lucky and a loss can be the right read; a record can't tell those apart, so it says which, about the wins as well as the losses. Each game's review is on its card below.
Right, for its reasons
7
of 12 winning calls
Right, but lucky
5
of 12 winning calls
Wrong, read held up
3
of 19 losing calls
Wrong read
16
of 19 losing calls
The record is open. Every pick is free with All-Access this season.
All-Access shows every pick in full: the Best Bet of the Week, and the side, the total, the predicted score and the reasoning for all 15 games. It's $12.99 a month, and free with an account for the rest of the season.
🔒 There's a pick for this game. All-Access shows it in full, free this season.
🔒 There's a pick for this game. All-Access shows it in full, free this season.
🔒 There's a pick for this game. All-Access shows it in full, free this season.
🔒 There's a pick for this game. All-Access shows it in full, free this season.
🔒 There's a pick for this game. All-Access shows it in full, free this season.
🔒 There's a pick for this game. All-Access shows it in full, free this season.
🔒 There's a pick for this game. All-Access shows it in full, free this season.
🔒 There's a pick for this game. All-Access shows it in full, free this season.
🔒 There's a pick for this game. All-Access shows it in full, free this season.
🔒 There's a pick for this game. All-Access shows it in full, free this season.
🔒 There's a pick for this game. All-Access shows it in full, free this season.
🔒 There's a pick for this game. All-Access shows it in full, free this season.
🔒 There's a pick for this game. All-Access shows it in full, free this season.
No pick yet: picks are made once the lines are posted.
The StatGrades Formula rates every team in points against an average opponent on a neutral field. It starts from half of last season's scoring margin and shifts toward this season's play-by-play efficiency and opponent-adjusted margins as games are played. It adds home field (about 1.8 points), takes points off a team whose starting quarterback is listed Out or Doubtful, and trims the total in strong wind. Its scoring baseline is the week's average market total. When the market has moved 2 or more points since it opened, which usually means news it hasn't seen yet, it keeps only a quarter of its disagreement on that game. It takes the side of whatever gap remains. On each card the Gap is that difference in points against the books' live number, and its top call is whichever of its two gaps stands out more against its usual differences with the market that week. It gives no confidence rating: a big gap is a measurement, not a promise.
StatGrades AI reads the same things plus every book's line and where it opened, the full injury report with our player grades, betting trends, the kickoff weather forecast, each team's rest and travel, and a fresh search of the 32 team sites and established outlets for news the market may not have priced. Then it writes its own call and reasoning, and lists the articles it read under the pick. Its Confidence is its own, in words: a coin-flip lean (1 of 5), a lean (2), a solid edge (3), a strong edge (4) when several independent reasons agree, and rare (5). Most games get a 1 or 2; we don't inflate. The Best Bets of the Week have a bar: the AI's needs at least a solid edge (3 of 5) and the Formula's a gap of at least 3 points; otherwise that side passes.
Who wins on each card is the picker's own chance of each team winning outright and its projected score; the thinner bar under it is the sportsbooks' chance from their moneylines, with their cut removed. The Best Bets of the Week freeze theirs when they lock.