NFL handbook — spread / total / props / live workflow
NFL is the most-bet, most-modeled, and most-efficient market in American sports — the hardest league to find raw mispricing in, and the easiest one to leak money in by being sloppy about numbers and timing. This session is the operating manual: what the key numbers are worth in EV, when in the week each edge appears, and why spread, total, props, and live are four separate workflows that happen to share a scoreboard.
What this session covers
Three things, in order. The key-number structure of NFL scoring and the arithmetic that makes a half point around 3 worth several times what a half point elsewhere is worth. The weekly rhythm of the market, from Sunday-night openers through Sunday-morning inactives. Then the four workflows, each with its own timing and its own failure mode.
Key numbers, and what the half point is worth
NFL scoring is lumpy. Touchdowns, field goals, and extra points concentrate final margins on a handful of values. Historically, roughly 14-15% of NFL games are decided by exactly 3 points and roughly 9% by exactly 7. Margins of 10 and 14 form the next tier. Almost everything else is thin.
Here is what that does to a price. Take a game genuinely priced at 3, and say the favorite wins by exactly 3 nine percent of the time, with the remaining 91% split evenly. At -3 and -110, that 9% is a push, and EV per unit is 0.455 × 0.909 − 0.455 = −4.14% — just the vig.
Now cross the number. At -3.5 and the same -110, the 9% becomes a loss: 0.455 × 0.909 − 0.545 = −13.14%. You gave up 9 percentage points of EV for half a point.
That relationship is exact and it generalizes: crossing a number costs you its push mass. Run the same arithmetic at 7 with 6% push mass and you lose 6 points of EV. At a dead number like 5, with maybe 3.5% push mass, you lose 3.5.
Priced as odds instead of EV, if -3 is available at -110, then -3.5 has to pay about +111 to be the same bet — roughly 21 cents of price. The same half point across 7 is worth about 14 cents. Across 5, about 8. When a book prices the half point at 10 cents on both, one of those numbers is a gift and the other is a trap.
The practical rule that falls out: the number matters more than the price, inside reason. Taking -3 at -115 beats taking -3.5 at -105 by more than 5 percentage points of EV on the same game.
The weekly rhythm
Edges in NFL are timing edges, and the week has a shape.
Sunday night into Monday. Look-ahead lines and openers post at low limits — the least-informed price of the week, and also the point at which your own information is thinnest.
Tuesday through Thursday. Limits rise, syndicate money arrives, and the market does most of its price discovery. A model-built opinion is usually worth the most here.
Wednesday through Friday. Practice reports. Wednesday is soft signal, Thursday firms it up, and Friday's game-status designations are the first hard information. Books price the designation quickly; they price the implication more slowly, especially two or three players down a depth chart.
Sunday morning. Inactives land 90 minutes before kickoff — the fastest news window of the week, and the one where slow books lag furthest behind fast ones.
Weather, late. Wind moves totals more than temperature and far more than precipitation. Forecasts firm up inside 48 hours, and outdoor totals often do not fully catch up until Sunday morning.
Four markets, four workflows
Spread. A number-hunting exercise, not a price-hunting one. Your job is to be on the right side of 3 and 7 and to get there before the market crosses.
Total. Weather and pace, in that order, with the same key-number logic applying weakly at 41, 44, and 51. Totals move on information later in the week than spreads do, which makes them the better market for anyone who cannot watch a screen midweek.
Props. A different game entirely. Prop lines are modeled less carefully, repriced more slowly, and reflect depth-chart changes with a real lag. Proptimizer computes a no-vig probability for every priced prop across 100+ books and surfaces edge % against the best available price.
Live. Highest hold, fastest movement, most model overreaction, and where discipline fails first. That is a full lesson of its own.
What the live session adds
- A live walk of that week's board, calling half-point value out loud on real numbers before checking what the books are charging for it
- Two worked timing examples: an opener taken Monday versus the same position taken Saturday, scored on CLV rather than result
- Screen-share of the Sunday-morning inactives routine, from news drop to price check to sizing decision
- Open Q&A, usually heaviest on when it is worth paying up to stay on 3
Session status: Replays post to this page. The next live date is announced in Discord and in the Sharp Report.