NFL WEEK 1

Model vs. Kalshi on all 16 openers

Week 1 markets live now

FOMC SEPT 16

Kalshi, Polymarket and futures disagree on the Fed

Decision Wed 2pm ET

NFL EDGES

Live game edges — model vs. Kalshi, ranked

Gridiron Edge · 5-min refresh

COMBO EDGE

Correlation-aware Kalshi combos — free

SILVER EDGE

Options vs. Kalshi on weekly silver brackets

FREE + PRO

Ladder Arb Scanner

Every Kalshi college football and NFL spread and total ladder, checked for strikes priced out of order. Locked arbitrage, crossed mids with the orders that capture them, and wide books worth making a market in. The whole board is free.

The monotonicity rule, and the three signals
For strikes Xlo < Xhi on one side of one game:  P(margin > Xhi) ≤ P(margin > Xlo)
LOCKED    ask(Xlo) < bid(Xhi) → buy lo at ask, sell hi at bid; edge = bid(Xhi) − ask(Xlo)
INVERTED  mid(Xhi) > mid(Xlo) + 0.5¢ → post BUY lo @ ask(Xhi)↓1¢, SELL hi @ bid(Xlo)↑1¢
WIDE      ask − bid ≥ 3¢, volume ≥ $20k, 5¢ < mid < 95¢, pre-kickoff; ranked by spread × √volume
Fee per contract = 0.07 · P · (1 − P) (taker). Net edge = gross − fees on both legs. The board sorts and tiers on NET.
STRONG LOCKED net ≥ 2¢, or INVERTED net ≥ 3¢ on ≥ $50k min-leg volume · MODERATE INVERTED net ≥ 1¢ on ≥ $20k · SPECULATIVE WIDE rows and any INVERTED under the volume floor
A pair is a pair only if it is the same game, the same side and the same series. Never across games, across sides, or across spread and total. Moneyline ladders on KXNCAAFGAME / KXNFLGAME check the two-sided variant: ask(A) + ask(B) < $1.00 − fees.

Quick Answer

the Ladder Arb Scanner reads every Kalshi college football and NFL spread and total ladder and flags strikes priced out of order — a higher strike quoted above a lower one on the same side, which is impossible by arithmetic. LOCKED rows are true arbitrage (rare); INVERTED rows are crossed midpoints a resting order pair captures; WIDE rows are books worth making a market in. Measured Sept 2026, CFB spread ladders are internally inconsistent 5.8% of the time versus 0.3% for NFL. The full board is free; Pro adds the resting orders, net-of-fee edge and ¼-Kelly size on each row.

The book has these two prices backwards

That is the whole product. Kalshi lists a spread or a total as a stack of strikes, and the stack has to slope one way: it can never be more likely that a team wins by more than 30 than by more than 27. When the order book prices it that way anyway, nobody has to be right about the game — the two prices disagree with each other. The scanner finds those, says which of three kinds each one is, and links straight to the contract.

NFL brings the traders. CFB pays them. The NFL board is tight and carefully made; the college board on a Saturday has 144 spread ladders and thin maker coverage on most of them. The sport toggle defaults to whichever has more live pre-kickoff signals.

Loading the board…

Go Pro · 26% off your first payment

Every edge the engines find, the moment they find it

Pro opens every engine on the site — the cross-market arb scanner, the Mispricing Scanner, Thee Oracle, and the silver, gold, oil and bitcoin edge models that price commodity contracts in real time.

Claim 26% off →

$111 for the first year, or $11.09 for the first month. Code applies automatically. Ends Dec 31.

Every signal graded in public, losers left on the board. See the record →

Related Tools

Related Reads & Boards

How the scanner works

Every scan pulls the open pre-kickoff markets in six Kalshi series — college football and NFL game lines, spreads and totals — and groups them into ladders: one ladder per game, per side, per series. Within each ladder the strikes are sorted and every adjacent pair is checked against the one rule a ladder cannot break: a higher strike can never be worth more than a lower strike on the same side. Prices are read once at the fetch boundary; a Kalshi leg with no book is treated as absent, never as a zero. Games that have kicked off are dropped entirely — an in-play ladder is one venue lagging itself, not a signal.

Three things can be wrong with a ladder and the board reports them separately. LOCKED is the ask at the low strike sitting under the bid at the high one — true arbitrage, rare, the headline when it fires. INVERTED is crossed midpoints: not takeable at market, but a resting order pair captures it, and the exact prices to post are computed for every row. WIDE is a two-sided book quoted 3 cents or more apart on real volume — not an error, a market nobody is making yet. Fees are computed on every leg and the tiering runs on the net figure, because a 2-cent inversion at the wrong price is fee-negative and a scanner that publishes those loses the one reader who checks.

Why the board is free

Publishing these makes CFB ladders tighter over time, and that is fine. The moat is the measurement and the memory — the per-series inversion rate, updated every scan, and the quote history that turns a fill probability from a guess into a number — not any single row. So the whole board is free: every sport, every signal, both tickers, both books, the gross edge, the volume and the kickoff. What Pro adds is the execution layer on each row: the two resting orders with the price on each leg, the net-of-fee edge and its fee breakdown, and the quarter-Kelly share of bankroll. An agent can call the same board through the MCP server as ladder_arb, with the same free/Pro split.

Where this sits among the tools

The Arb Scanner compares one contract across two venues; this compares two strikes inside one venue. The Ladder puts every NFL strike beside our model's distribution — an opinion about where the price should be. This page needs no opinion: the book contradicts itself, or it does not. Size a row with Kelly and price the round trip with the Kalshi Fee Calculator.

Frequently Asked Questions

What is ladder arbitrage on Kalshi?

Kalshi prices a spread or total as a ladder: a stack of strikes on the same side of the same game — "Alabama wins by more than 2.5", "by more than 3.5", "by more than 4.5", and so on. The probability of clearing a higher strike can never be greater than the probability of clearing a lower one, so a higher strike can never be worth more than a lower strike on the same side. Ladder arbitrage is the family of trades that exists when the order book breaks that rule: buy the lower strike, sell the higher one, and the pair is worth at least what you paid regardless of the final score. The scanner reads every Kalshi college football and NFL spread and total ladder, checks each adjacent pair for that violation, and reports three distinct signals — LOCKED, INVERTED and WIDE — never blended into one score.

Is this real arbitrage?

Sometimes, and it is labelled precisely. LOCKED means the ask at the lower strike sits below the bid at the higher strike right now, so buying one and selling the other at market locks the difference before fees — that is true arbitrage, and it is rare: on a full Saturday board in September 2026 we measured zero. INVERTED means the midpoints are crossed but the trade is not there at market; a pair of resting orders — buy the low strike at the high strike's ask, sell the high strike at the low strike's bid — captures it if both fill. If only one fills you hold a position at better than market in a ladder the book is already pricing wrong. That is a good position, not a riskless one, and the scanner says so. WIDE is not arbitrage at all: it is a two-sided book quoted 3 cents or wider on real volume, which is where a patient trader can make a market. Every row also reports the edge gross of fees, and the tiering runs on the net figure, because a 2-cent inversion can be fee-negative.

Why college football?

Because that is where the inconsistencies are. Measured live on 2026-09-12 across pre-kickoff markets only, Kalshi college football spread ladders had crossed midpoints on 5.8% of adjacent strike pairs (105 of 1,809) and NFL spread ladders on 0.3% (1 of 349). CFB ladders were also about three times as likely to be quoted 5 cents wide or worse. The reason is coverage: 144 CFB spread ladders on a Saturday, many of them FCS and Group of Five games that no market maker prices carefully, against roughly 30 tightly quoted NFL ladders a week. NFL brings the traders; CFB pays them. The scanner covers both, and the sport toggle defaults to whichever has more live pre-kickoff signals.

Are fees included?

Yes, and they decide the tier. Kalshi charges a quadratic fee per contract — 0.07 × price × (1 − price) for a taker, about 1.75 cents on a 50-cent contract and near zero at the extremes — and a ladder inversion typically sits at the extremes on one leg and in the middle on the other. Every row reports the gross edge in cents free, and the net-of-fee edge is computed for every row; the board sorts and tiers on the net figure, so a fee-negative inversion is never published as STRONG. The net number itself is a Pro field, rendered inline and locked, alongside the resting orders and the quarter-Kelly size. The fee math is the same formula the Kalshi Fee Calculator on this site uses.

How is this different from the cross-platform Arb Scanner?

The Arb Scanner compares the SAME contract on two venues — Kalshi and Polymarket — and flags the gap between them. The Ladder Arb Scanner never leaves Kalshi: it compares two DIFFERENT strikes on the same side of the same game inside one venue's order book. Cross-venue arbitrage needs both accounts funded, both resolution rules checked, and it carries the venue mismatch risk the Arb Scanner page documents. Intra-ladder arbitrage has one venue, one rulebook and one settlement — the only questions are fees and whether the resting orders fill. Both tools are also exposed to AI agents on the PredictionMarketsPicks MCP server as find_arbitrage and ladder_arb.

Get The 7 Oracles' daily edge — subscribe free

No spam. Unsubscribe anytime.