Updated Sep 7, 2026 · Week 1 board.
There is no futures market this week, so we build on what is here
We checked. The board holds 16 Week 1 game lines, each with a spread, a total and a moneyline, at three books. There is no separate market for win totals, division winners, conference odds or a Super Bowl price anywhere in it. A futures piece quoting a specific future price here would be making it up. So the mechanics run on real Week 1 moneylines instead. The arithmetic does not change between a moneyline and a division-odds future. The market does.
Turning a moneyline into an implied win chance
A moneyline price is not a win chance. The book's margin is sitting inside it. For a negative price, take the price and divide it by the price with 100 added. For a positive price, take 100 and divide it by the price with 100 added. Add the two sides and the total comes in above 100%. The excess is the margin, the vig. Each side divided by that total gives a fair, book-specific chance. Two Week 1 games show the contrast.
| Game | Book | Moneyline (home/away) | Win chance, home | Win chance, away |
|---|---|---|---|---|
| New Orleans at Detroit | Pinnacle | −321 / +267 | 73.7% | 26.3% |
| New Orleans at Detroit | DraftKings | −305 / +245 | 72.2% | 27.8% |
| New Orleans at Detroit | Bovada | −350 / +275 | 74.5% | 25.5% |
| Buffalo at Houston | Pinnacle | −108 / −104 | 50.5% | 49.5% |
| Buffalo at Houston | DraftKings | −110 / −110 | 50.0% | 50.0% |
Detroit is priced as a heavy favorite at every book, 72% to 75% depending on which one you open. Buffalo at Houston is as near a coin flip as this board gets. A futures price converts the same way. A division number or a win-total over-under goes through the division a moneyline goes through. The only piece missing is a futures price to put into it.
Sizing a position: the Kelly fraction, and our own cap
Once your own win chance differs from the market's, the Kelly criterion (a stake-sizing formula) turns that difference into a bet size. It weighs the edge against the payout and hands back a share of bankroll. In words, the fraction is your win chance times the net odds, minus the chance you lose, all divided by the net odds. Our own published picks show it running. Take Buffalo at Houston. Our stated chance there is 0.581 at −102, which puts the net odds near 0.98. Full Kelly on that is roughly 15.4% of bankroll. Our recommended stake is 3.84%, near enough a quarter of it. The Green Bay-Minnesota spread pick, at a stated chance of 0.5615 and the same −102 price, lands the same way. Full Kelly is near 11.4% and our stake is 2.86%. Of the 283 published picks carrying a stake figure, 136 sit at exactly 5%. That looks like a cap at 5% of bankroll no matter what quarter-Kelly would ask for on a confident pick. We checked two picks. We state that as consistent with a quarter-Kelly rule capped at 5%, and not as a written policy.
A portfolio is many small positions, not one big one
The word portfolio is doing real work in that sentence. One 5%-of-bankroll position on one game is a bet. Spreading the same total risk across the 16 games on the Week 1 board, each with its own price and its own edge or absence of one, is what a portfolio means. Many small positions, largely independent of each other, instead of one story. We hold no correlation figure between games, so we won't put a number on the benefit. More independent positions still beats one large position at the same total stake. Our own 5% cap is the visible, sourced example of that discipline being enforced by the machinery rather than left to the bettor.
Sources
We checked two things, converting Week 1 moneylines into win chances and sizing a stake against them, both off our own published picks and timestamped below. The method is what this page is worth. No futures price exists to apply it to, and nothing here has settled. Every edge and stake figure above is a pre-game estimate. Size accordingly, and watch for a market to actually post futures prices before you treat a division or Super Bowl number as real. The full 16-game Week 1 board sits at /odds. Every edge and stake figure we publish, not only the two used above, is at /picks/nfl. A standalone converter for your own book's price sits at /desk/tools/odds-converter. The same projection-against-market method, run on one position instead of a slate, is at TE premium strategy.
