NFL Confidence Pool Picks, Week 2
Every Week 2 game ranked by point value, 16 points on the most confident pick down to 1 on the least, with the spread and kickoff behind each one.
| Pts | Pick | Matchup | Spread | Kickoff | Win % | Market | Our rating |
|---|---|---|---|---|---|---|---|
| 16 | SFSan Francisco 49ers | vs MIAMiami Dolphins | -10.5 | Sun 4:25p | 83.0% | 83.7% | 66.5% |
| 15 | SEASeattle Seahawks | at ARIArizona Cardinalsdiv | -10.0 | Sun 4:25p | 79.8% | 81.2% | 57.2% |
| 14 | LALos Angeles Rams | vs NYGNew York Giants | -9.5 | Mon 8:15p | 79.2% | 79.6% | 73.6% |
| 13 | LACLos Angeles Chargers | vs LVLas Vegas Raidersdiv | -8.5 | Sun 4:05p | 74.8% | 75.1% | 71.8% |
| 12 | BALBaltimore Ravens | vs NONew Orleans Saints | -7.5 | Sun 1:00p | 73.4% | 74.8% | 61.2% |
| 11 | KCKansas City Chiefs | vs INDIndianapolis Colts | -6.5 | Sun 8:20p | 70.6% | 72.6% | 56.0% |
| 10 | TBTampa Bay Buccaneers | vs CLECleveland Browns | -6.5 | Sun 1:00p | 69.7% | 70.5% | 64.4% |
| 9 | GBGreen Bay Packers | at NYJNew York Jets | -5.5 | Sun 1:00p | 67.3% | 68.1% | 63.4% |
| 8 | PHIPhiladelphia Eagles | at TENTennessee Titans | -5.5 | Sun 1:00p | 66.1% | 66.9% | 62.2% |
| 7 | NENew England Patriots | vs PITPittsburgh Steelers | -5.0 | Sun 1:00p | 65.5% | 65.8% | 64.1% |
| 6 | DALDallas Cowboys | vs WASWashington Commandersdiv | -4.5 | Sun 4:25p | 63.4% | 64.5% | 58.9% |
| 5 | HOUHouston Texans | vs CINCincinnati Bengals | -2.5 | Sun 1:00p | 61.7% | 57.2% | 72.6% |
| 4 | CHIChicago Bears | vs MINMinnesota Vikingsdiv | -3.0 | Sun 1:00p | 59.9% | 60.0% | 59.8% |
| 3 | BUFBuffalo Bills | vs DETDetroit Lions | -3.0 | Thu 8:15p | 59.2% | 59.3% | 58.9% |
| 2 | DENDenver Broncos | vs JAXJacksonville Jaguars | -3.0 | Sun 4:05p | 57.7% | 58.3% | 56.1% |
| 1 | ATLAtlanta Falcons | vs CARCarolina Panthersdiv | -1.5 | Sun 1:00p | 56.0% | 53.3% | 61.2% |
Standard confidence scoring: the point value if the pick wins, nothing if it loses or ties. Games are ordered by the betting market's implied win probability with the vig removed, adjusted toward our own power rating, which is opponent-adjusted expected points added per play built from play-by-play. How far it moves depends on how sure the market is: on a ten-point favourite the market is almost impossible to beat and our rating barely registers, while on a near coin flip it takes up to forty percent of the call. Where the two disagree by ten points or more the rating is highlighted. We do not claim to beat the market. Ordering by our rating alone earned 69.2% and 63.5% of available points across 2024 and 2025 against the market's 75.9% and 70.9%, and over five seasons the picks where it overrules the favourite have been close to a coin flip, so read those as a different opinion rather than a better one. Spreads are the median across books, snapped to the half point, from the pick's own perspective. The sheet refreshes while lines move, then freezes permanently at the week's first kickoff. Past weeks are shown exactly as they stood at lock, never restated.
This is the current week. Confidence pool picks always shows the latest sheet. The survivor pool grid covers all 18 weeks.
Confidence pool questions
How does an NFL confidence pool work?
You pick a winner in every game on the slate, then rank those picks by how sure you are. With sixteen games you assign 16 points to your most confident pick down to 1 point for your least, using each value exactly once. You collect the point value when a pick wins and nothing when it loses, so the ranking matters as much as the picks.
How are confidence pool points scored?
Standard scoring awards the assigned point value on a win and zero on a loss or a tie. A sixteen-game slate puts 136 points on the board each week. Because a wrong 16-point pick costs sixteen times what a wrong 1-point pick does, ordering the board well is usually worth more than any single pick.
Where do these 2026 confidence pool picks come from?
Games are ranked by the betting market’s implied win probability with the vig removed, adjusted toward our own power rating. That rating is opponent-adjusted expected points added per play, built from play-by-play, and it is independent of any sportsbook. How much it moves a game depends on how certain the market is: almost nothing on a lopsided line, up to forty percent of the call on a near coin flip. The most likely winner gets the most points and the closest game gets the fewest.
Why does the market lead instead of your own rating?
Because we tested it. Ordering a board by our power rating alone earned 69.2% and 63.5% of available confidence points across the 2024 and 2025 seasons, against 75.9% and 70.9% for the market, with the rating’s settings fit on earlier seasons so it was never shown the answers. The betting market is very hard to beat at picking winners, so it decides the shape of the board. Our rating gets the most say exactly where the market is least certain, which is where a disagreement is worth acting on.
When do the picks lock?
The sheet resyncs every fifteen minutes while lines move, then freezes permanently at the week 2 first kickoff. After that it cannot change, which is enforced in the database rather than in the site, so a past week always shows exactly what was published before those games were played.
Do these picks ever take an underdog?
Yes, but not often, and never as a gimmick. An underdog appears when our power rating disagrees with the market by enough to outweigh it, which happens on roughly nine games a season and can land at any point value. We do not force one onto the board each week or park it in the cheapest slot. Being straight about it: across five seasons those picks have been close to an even split, so they are a genuine second opinion rather than a proven edge.
Should your 16-point pick always be the biggest favorite?
Under standard scoring, yes: expected points are the point value multiplied by the chance the pick wins, so the safest game should carry the most points. Getting clever at the top of the board costs more than it gains. The room to differentiate is at the bottom, where a near coin flip carrying one or two points barely changes your expected score either way.
Does the point spread matter in a confidence pool?
Only as a signal, never for scoring. Confidence pools are decided straight up, so a team winning by one point scores the same as a team winning by thirty. The spread is shown here because it is the quickest read on how large a favorite each pick is.
Want the raw data behind these reports?
25 seasons of NFL data in Excel format, updated weekly.