Updated Sep 7, 2026 · Week 1 board.
The house edge on today's board, leg by leg
Every Week 1 game in our published picks carries two spread numbers. One is our model's own predicted margin. The other is the edge, which is the distance between that margin and the market line. Seven of this week's spread picks come in under 1.1 points of edge:
| Game | Market line | House predicted value | House edge (points) | House confidence |
|---|---|---|---|---|
| Green Bay Packers @ Minnesota Vikings | 1.5 | 1.94 | 0.44 | 56.2% |
| Atlanta Falcons @ Pittsburgh Steelers | 3.5 | 2.94 | -0.56 | 57.6% |
| Washington Commanders @ Philadelphia Eagles | 5.5 | 6.15 | 0.65 | 58.7% |
| New Orleans Saints @ Detroit Lions | 7 | 6.26 | -0.74 | 59.8% |
| New York Jets @ Tennessee Titans | 1.5 | 2.26 | 0.76 | 60.1% |
| Cleveland Browns @ Jacksonville Jaguars | 7.5 | 8.44 | 0.94 | 62.1% |
| Chicago Bears @ Carolina Panthers | 2.5 | 1.44 | 1.06 | 63.4% |
An edge of 0.44 points means our predicted margin is sitting less than half a point away from a number the market already published. That is the whole size of the argument between our model and the book on that game. And that is before anybody lays a pleaser on top of it.
What break-even costs at today's real prices
Points and win probability are different units. Confidence (our model's own win chance) is a different number from that point edge, and it comes off the same published pick rather than out of the point gap. Here is the price the same leg was actually published against:
| Game | Book | Price | Break-even win rate | House confidence | Cushion over break-even |
|---|---|---|---|---|---|
| Green Bay Packers @ Minnesota Vikings | bovada | -105 | 51.2% | 56.2% | 5.0 points |
| Atlanta Falcons @ Pittsburgh Steelers | draftkings | -118 | 54.1% | 57.6% | 3.5 points |
| Washington Commanders @ Philadelphia Eagles | fanduel | -115 | 53.5% | 58.7% | 5.2 points |
| New Orleans Saints @ Detroit Lions | pinnacle | -110 | 52.4% | 59.8% | 7.4 points |
| New York Jets @ Tennessee Titans | draftkings | -115 | 53.5% | 60.1% | 6.6 points |
| Cleveland Browns @ Jacksonville Jaguars | bovada | -115 | 53.5% | 62.1% | 8.6 points |
| Chicago Bears @ Carolina Panthers | draftkings | -118 | 54.1% | 63.4% | 9.3 points |
Skip the formula and do the bar math. At -105 you need a shade over half your bets, 51.2%, just to break even. At -118 it's 54.1%. Our stated confidence clears break-even on all seven of these legs, by anywhere from 3.5 to 9.3 percentage points at the real, un-pleased price.
That cushion is a genuine number with a source behind it. It's also unaudited. It's our own estimate against a price we were quoted, and not one of these has been graded.
The tax a pleaser adds does not need to be large
A pleaser makes the same leg clear a harder number than the one it was priced against. More points for the favorite to cover. Less cushion for the dog. We hold no pleaser paytable here, so I'm not inventing a point adjustment for you.
But that adjustment doesn't have to be big to matter against a 0.44-point edge. The Green Bay-Minnesota gap is smaller than half a point. Any pleaser move bigger than that has already eaten the entire disagreement between our model and the market on that game. And books price pleasers to move the number by more than a rounding error, because otherwise there would be no reason to pay you extra for one.
The 9.3-point cushion on Chicago-Carolina is a probability cushion, not a points cushion, and nobody here can convert one into the other. Nothing in front of me connects a specific point move to a specific probability change on any of these seven games. The one thing I do know is that our raw point edge, on more than half of this week's published spread picks, is thin. A thin edge does not need a steep tax to vanish.
What we are not claiming
Nothing above claims these seven edges will hold. Nothing above claims our confidence numbers are calibrated. They're our own published estimates and I am labeling them that way on purpose. I'm also not quoting a pleaser payout or a point adjustment, because I don't have one to quote.
The same question from the market-consensus side is in Pleaser Sports Betting Explained: Where Risk Compiles, which reads how far the three tracked books agree with each other on these same Week 1 numbers.
The Receipts Drawer
Seven spread picks, every one under 1.1 points of edge, five of them under a single point. Thin, and real. All seven clear break-even by at least 3.5 percentage points of confidence cushion at the price we were actually quoted. That is a tight, well-priced board.
It is not proof any of these seven wins. It is proof the model and the market are standing very close together, which is precisely when a pleaser's extra points do the most damage. So pull the published edge and the break-even cushion before you pay a pleaser tax on any of them. The case gets worse if the market line drifts toward our predicted number.
The edge shrinks, and the argument for pleasing shrinks along with it.
