Do Kalshi and Polymarket price NFL games differently from sportsbooks?

Not in Week 1 of 2026. Across the fourteen games logged on 5 September, the largest gap between either market and the de-vigged sportsbook consensus was 2.2 percentage points, and the median gap was 1. All three sources named the same favourite in every game.

Which source moved the most since open?

Polymarket. Its Chargers contract went from 45% to 81%, Jaguars 51% to 78%, Steelers 46% to 63% and Rams 50% to 64%. Those opening prints were thin placeholders; each converged to the sportsbook figure. Kalshi’s reported opening prices were within four points of the current consensus.

Is a prediction market price no vig?

The contract price carries no built-in hold, so a 64 cent contract implies 64%. Both exchanges charge trading fees, which are a cost of the same kind as vig. The comparison in this log removes vig from the books but does not deduct exchange fees.

How will the log be scored?

After the games, each source gets a Brier score across the fourteen games, plus the favourite hit rate. A new dated page will carry the results; this page will not be edited.

Why are only 14 of 16 games listed?

Two Week 1 matchups were not on the source index page when the snapshot was taken. The log is not back-filled.

Yes. The CSV is CC BY 4.0. Link to this page.

NFL Week 1: Kalshi, Polymarket and the books, logged before kickoff

Data log · Published September 5, 2026 · 18+ · Not betting advice · By BetG8 Editorial · Prediction Markets

NFL Week 1: Kalshi vs Polymarket vs the books, logged before kickoff

The 2026 NFL season opens on Wednesday 9 September. Before a single snap, this page records what three different kinds of market think about every Week 1 game: the sportsbook consensus with the vig stripped out, the Kalshi contract price, and the Polymarket contract price. All fourteen games available on the source index were read in one pass at 5 September 2026, 11:50 UTC. Nothing here has been adjusted after the fact, and the file will not be edited once the games are played. It will be scored instead.

The point of logging first and judging later is simple. Every week someone claims prediction markets are “sharper” than sportsbooks, or the reverse, and every week the claim is made after the result is known. A log with a timestamp removes the hindsight. Next week the same fourteen rows get a result column and a Brier score for each of the three sources, and the month-end version rolls the weeks up.

The finding, before any game is played: the three sources agree on every game. The largest gap between any market and the de-vigged book consensus is 2.2 points (Jets at Titans), the median gap is 1.0 points, and six games sit under a single point. Where Polymarket looks like it disagreed at the open — Rams 50%, Steelers 46%, Jaguars 51%, Chargers 45% — those were thin opening prints that converged 14 to 36 points toward the books. Kalshi’s opening prices moved by four points at most. Download the log (CSV, 14 rows).

The log: every Week 1 game, three prices

Probabilities are for the favourite named in the column. “Books” is the median of 14 sportsbook moneylines with the vig removed, as reported by the source page. Kalshi and Polymarket are the contract prices at the same snapshot, which carry no vig by construction. “Max gap” is the largest absolute difference, in percentage points, between any of the three. Open-to-now columns are shown only where the source page reported a since-open move.

Game	Kickoff	Consensus spread	Favourite	Books (no vig)	Kalshi	Polymarket	Max gap	Kalshi open→now	Polymarket open→now
Patriots at Seahawks	Wed 9 Sep, 8:20 PM ET	Seahawks -3.5 14/14 books	Seahawks	63%	62%	63%	1	—	—
49ers at Rams	Thu 10 Sep, 8:35 PM ET	Rams -3.5 13/14 books	Rams	64%	64%	64%	0.1	62→64 (+2)	50→64 (+14)
Falcons at Steelers	Sun 13 Sep, 1:00 PM ET	Steelers -3.5 9/13 books	Steelers	62%	63%	63%	1.4	59→63 (+4)	46→63 (+17)
Ravens at Colts	Sun 13 Sep, 1:00 PM ET	Ravens -3.5 14/14 books	Ravens	62%	63%	62%	1.3	—	—
Bills at Texans	Sun 13 Sep, 1:00 PM ET	Bills -1.5 10/14 books	Bills	51%	50%	50%	0.7	—	—
Bears at Panthers	Sun 13 Sep, 1:00 PM ET	Bears -2.5 12/14 books	Bears	58%	59%	60%	1.7	—	—
Buccaneers at Bengals	Sun 13 Sep, 1:00 PM ET	Bengals -3.5 13/14 books	Bengals	65%	65%	66%	1.4	63→65 (+2)	62→66 (+4)
Browns at Jaguars	Sun 13 Sep, 1:00 PM ET	Jaguars -8 8/14 books	Jaguars	77%	78%	78%	1.3	76→78 (+2)	51→78 (+27)
Saints at Lions	Sun 13 Sep, 1:00 PM ET	Lions -7 11/14 books	Lions	73%	74%	74%	0.6	—	—
Jets at Titans	Sun 13 Sep, 1:00 PM ET	Titans -1.5 9/14 books	Titans	54%	52%	53%	2.2	—	—
Cardinals at Chargers	Sun 13 Sep, 4:25 PM ET	Chargers -10 7/14 books	Chargers	82%	81%	81%	0.5	80→81 (+1)	45→81 (+36)
Packers at Vikings	Sun 13 Sep, 4:25 PM ET	Vikings -1.5 11/14 books	Vikings	52%	52%	53%	0.9	51→52 (+1)	50→53 (+3)
Dolphins at Raiders	Sun 13 Sep, 4:25 PM ET	Raiders -3.5 14/14 books	Raiders	63%	62%	62%	0.6	—	—
Commanders at Eagles	Sun 13 Sep, 4:25 PM ET	Eagles -5.5 11/14 books	Eagles	67%	68%	68%	1	66→68 (+2)	64→68 (+4)

Two Week 1 matchups were not on the source index at snapshot time and are therefore not in the log. They will not be added retroactively; they will appear in the Week 2 log if the source carries them.

What the table actually says

Agreement is the story, not disagreement. Fourteen games, zero cases where a prediction market and the books are on opposite sides of 50%. Ten of the fourteen are within 1.3 points. If you came looking for a market that “knows something” the books do not, Week 1 does not offer one. The two closest games by price — Bills at Texans (51/50/50) and Packers at Vikings (52/52/53) — are also the two where all three sources sit closest together.

Polymarket’s opening prices are not opinions. The largest since-open moves in the log are all Polymarket: Chargers 45%→81% (+36), Jaguars 51%→78% (+27), Steelers 46%→63% (+17), Rams 50%→64% (+14). Every one of those contracts opened at roughly a coin flip and walked to where the books already were. That is what a market with almost no early liquidity looks like — the first print is a placeholder, not a forecast — and it is why “Polymarket moved 36 points” is a much less interesting sentence than it sounds. Kalshi’s opening prices, by contrast, started within four points of the current consensus in every game where a move was reported.

The widest gap is a two-point game nobody is excited about. Jets at Titans: books 54%, Kalshi 52%, Polymarket 53%. A 2.2-point spread between sources on a game priced at Titans −1.5 by only nine of fourteen books. The book consensus is itself loose there. Bears at Panthers is the other “wide” one at 1.7 points, with Polymarket the most confident in the favourite (60%) and the books the least (58%).

Where the vig goes. The book column is de-vigged, so it is comparable to contract prices. The raw book price is not: on the same 49ers–Rams game the best available moneyline was Rams −190 and 49ers +186 on a no-vig exchange, but −198/+164 at the mainstream books, an overround of roughly 4.3%. A Kalshi or Polymarket contract at 64¢ has no equivalent hold, though both charge trading fees that a full comparison has to include; we did that arithmetic in Kalshi vs Polymarket fees.

How this gets scored

After the Week 1 games, each row gets a result. For each source we compute the Brier score, which is (probability assigned to the winner − 1)² averaged across games, where a perfect forecaster scores 0 and a coin flip scores 0.25. We also report the plain favourite hit rate — how many of the fourteen favourites won — which is the same number for all three sources in Week 1 because they name the same favourite in every game. The scored version will be a new dated page; this one stays as published.

One week of fourteen games proves nothing about which source is sharper, and we will say so on the scored page. The gaps here are small enough that the difference in Brier score between sources will almost certainly be inside noise. That is fine. The useful output is the accumulated log, and the useful question is whether the gap between books and markets grows or shrinks as the season goes on and liquidity moves into the exchanges.

Methodology

We would rather you check this than trust it.

Source. All prices were read from the per-game odds pages at cleatz.com, which display sportsbook lines, a 14-book de-vigged median, and Kalshi and Polymarket contract prices from one snapshot at page load. Each row in the CSV carries the URL it was read from.
Snapshot. 5 September 2026, 11:50 UTC, single pass over fourteen pages in under one minute. No price was re-read or averaged.
De-vig. The book probability is the source’s “median of 14 books, no vig” figure. We did not recompute it. Different vig-removal methods (proportional, power, Shin) differ by up to about half a point on a −200 favourite; treat gaps under 0.5 as zero.
Rounding. All probabilities are whole percentages as displayed. The “max gap” is computed from the displayed integers, so a 0.1 reading reflects the source’s own sub-point figure.
Open prices. “Open” is the source’s reported price at contract open. It is shown only where the page reported a move; blank cells mean the source did not report one, not that there was no movement.
Coverage. 14 of 16 Week 1 games. Two matchups were absent from the source index at snapshot time.
Licence. The CSV is released under CC BY 4.0. Reuse it with a link to this page.

What this is and is not

It is a timestamped record of three prices per game before kickoff, and a commitment to score them. It is not a recommendation of any side, any venue, or any price. Kalshi and Polymarket are not available in every state, and their legal status is contested in more than twenty; see the 50-state tracker and the dated change log. Sportsbooks are licensed in 39 states and not the other 11. Nothing here should be read as advice. 18+.

FAQ

Do Kalshi and Polymarket price NFL games differently from sportsbooks?: Not in Week 1 of 2026. Across the fourteen games logged on 5 September, the largest gap between either market and the de-vigged sportsbook consensus was 2.2 percentage points, and the median gap was 1.0 points. All three sources named the same favourite in every game.
Which source moved the most since open?: Polymarket. Its Chargers contract went from 45% to 81%, Jaguars 51% to 78%, Steelers 46% to 63% and Rams 50% to 64%. Those opening prints were thin placeholders; each converged to the sportsbook figure. Kalshi’s reported opening prices were within four points of the current consensus.
Is a prediction market price “no vig”?: The contract price carries no built-in hold, so a 64¢ contract implies 64%. Both exchanges charge trading fees, which are a cost of the same kind as vig. The comparison in this log removes vig from the books but does not deduct exchange fees.
How will the log be scored?: After the games, each source gets a Brier score across the fourteen games, plus the favourite hit rate. A new dated page will carry the results; this page will not be edited.
Why are only 14 of 16 games listed?: Two Week 1 matchups were not on the source’s index page when the snapshot was taken. The log is not back-filled.
Can I use the data?: Yes. The CSV is CC BY 4.0. Link to this page.