Markets vs Pundits
What the money says, next to what the forecasters say — for every 2026 Senate race where both have gone on the record.
Why this board exists: markets beat the pundits at the top of the ticket in 2024 — and lost to them down ballot, which is exactly what a midterm is.
Quick Answer
this board puts the live Kalshi contract price for each 2026 Senate race next to the most recent rating a named forecaster published for that seat, converts the rating to a probability through a published ladder, and shows the gap. It is a scoreboard, not a recommendation: a gap means the market and the forecaster disagree, and the election settles which was closer. Nothing here carries a position size or an edge claim.
Frequently Asked Questions
Are prediction markets more accurate than pundits?
This page is built to answer that with receipts rather than assertion. It puts the live Kalshi contract price for a 2026 Senate race next to the most recent rating a named forecaster published for the same seat, converts the rating to a probability through a published ladder, and shows the gap. Nothing here claims the market is right — a gap only means the two disagree, and the scoreboard is settled by the election, not by us. Where the two are close, that agreement is itself the finding.
How do you turn a rating like "Leans Democratic" into a number?
Through one published ladder, shown on the page: toss-up is 50 percent, a lean is 60, likely is 75, and safe is 90, expressed toward the rated party. So "Leans Democratic" becomes a 60 percent chance the Democrat wins, and the same seat is 40 percent for the Republican. The ladder is deliberately coarse because the outlets publish a band and not a number — inventing more precision than they stated would put words in their mouth. A gap under roughly 10 points is mostly an artifact of that coarseness rather than a real disagreement.
Where do the market prices and the ratings come from?
Market prices are live Kalshi contracts for the individual Senate races, read as an implied probability — a contract at 60 cents is a 60 percent chance. Ratings come from named public forecasters, currently Sabato's Crystal Ball, with the date each rating was published and a link to the source on every row. Every row carries that link by design: a scoreboard that attributes a call to a named outlet is only as good as the proof they made it.
Why do some races have no market number?
Kalshi labels these contracts inconsistently. Some races are listed by party, and those line up directly against a rating. Most are listed by candidate name, and for an open seat with two challengers we hold no record of which party each candidate belongs to. Rather than infer a party from a name and risk labelling a candidate wrongly on a public page, those rows are left blank. A missing number is recoverable; a wrong party attribution is not.
Is this a trading signal?
No. There is no position size, no recommendation and no edge claim anywhere on this page, and none of these rows feed a tracked record. It is a comparison between two published numbers. The tool that used to sit here did claim an edge over the market, and its own settled record did not support that claim, which is precisely why it was replaced with a scoreboard that stakes nothing.
Ratings are the published work of their respective outlets, linked on every row and used here for comparison with attribution. Market prices are live Kalshi contracts and move continuously — read the price on the page, not this sentence.