How the math works

Every pick risks exactly 1 unit at the odds posted when the line went up. A final total above the line wins an Over; below wins an Under; exactly on the line is a push and the unit is refunded.

Win rate ignores pushes: wins ÷ (wins + losses). ROI is units won per unit risked on decided picks.

American odds imply a probability. For negative odds −A it is A ÷ (A + 100); for positive odds +B it is 100 ÷ (B + 100).

That is the win rate needed to break even. Removing the vig from a two-sided market (dividing each side by their sum) gives the book's fair probabilities; at −110/−110 that is 50/50 and the book's hold is 4.55%.

95% range is the Wilson score interval, which stays honest at small sample sizes where the textbook p ± 1.96·SE formula breaks.

Sharp probability is Bayesian: start from a uniform prior on a player's true win rate, update with their wins and losses to get a Beta(W+1, L+1) posterior, and report the probability that it sits above break-even. It's computed exactly with the regularized incomplete beta function.

Majority vote takes whichever side more people picked. Ties sit out. By Condorcet's jury theorem, if voters are independent and each is right more than half the time, the majority gets more accurate as the group grows — but correlated voters (friends who talk it over) add less than they seem to. That's why picks stay hidden until lock.

Skill-weighted uses the optimal rule for independent yes/no voters (Nitzan–Paroush): each vote counts by the log-odds of that person being right.

The 10 phantom wins and losses shrink small samples toward 50%, so a hot 4–0 start barely moves the needle while 60–40 over a hundred picks does. Anyone below 50% gets a negative weight — the model fades them. Weights are rebuilt walk-forward using only results settled before each game started, so the crowd's historical record has no look-ahead.

Final margins in the NFL aren't smooth: field goals and touchdowns make games end by 3 and 7 far more often than by 2 or 8. The Number doesn't assume a bell curve — for the spread you enter, it takes every regular-season game since 1999 that closed at that number (6,999 games from nflverse) and reads off what happened.

Win probability also shows the classic normal model, Φ(spread ÷ σ), with σ = 13.19 fitted from the same games (final margin minus closing spread). Your own number works the same way: the model uses games that closed at your spread and asks how often they beat the market's line — so a half point through 3 or 7 counts for what it's really worth.

A 17-week season is about 270 games — far too few to tell a 55% picker from a lucky coin. What can be measured is whether a pick got a better number than the market settled on at kickoff. Bettors who consistently beat the closing line are the ones who profit over time; the close is the market's best estimate, and beating it by a point is worth about a point of expected margin.

The site records each game's closing spread and total (from an odds feed before kickoff, or typed in by a commissioner) and shows every player's CLV in points and how often they beat the close. In the pick'em everyone plays the same frozen number, so CLV there measures which side the market moved toward after the freeze — a cleaner read on judgement than wins and losses. Half a point through 3 or 7 is worth more than half a point elsewhere, and The Number's margin distributions are used to price it.

Everyone picks every game against the same frozen spread and puts confidence points on it: 11 games at 1 point, 4 at 2 and 1 at 3 — 22 points if you sweep. A pick scores its points when the team covers; a push scores nothing. The pot is the buy-in times the paid entries, and it all goes back out each week.

Two honesty checks sit next to every score. vs. coin is how far a week beat what random picking averages. Skill? is the exact binomial chance a coin-flipper would hit at least that many of your decided picks — lower is more impressive. And because the pot goes to the best of eight entries, the winning score among eight pure coin-flippers is usually about 15, not 11: beating that is the bar.

The Crowd is a ghost entry: the majority side on every game, ties broken by total confidence, with the biggest point values on the most lopsided votes. If it keeps up with the winner, the room's consensus is worth something.

The pot pays deeper as the field grows, on the ladder Sam wrote. Up to five entries the winner takes everything. At six, second place opens and takes whatever is left above the winner's $50. At eight the winner's cut steps up to $60. At ten, third place opens at $10, and from there every extra entry adds its buy-in to the next place down: first, then second, then third, and around again.

Two properties are worth stating because the site enforces them. The shares always sum to the pot exactly, to the cent, at any buy-in. And a place is never worth more than the one above it. Ties share every place they cover: two players tied for first in a three-place week split first and second between them, and third is still paid to the next score down.

The fair line is the median of the data rounded to the half point (half the outcomes land on each side). Over probabilities come from two sources: the historical hit rate, and a normal model fitted to the sample mean and standard deviation (with a continuity correction for whole-number totals). Expected value averages the two, and pushes drop out because the stake comes back: