0x6256b4016256…6256b404

ConfirmedInstitutional Markets483 vB139 sat/vB2 min decode

Prediction Markets Mostly Matched Polls in the Midterms — With Three Misses

Kalshi and Polymarket mostly tracked polls in the midterms, but missed Wisconsin's 1st, Alabama's 2nd and the Iowa Senate seat — raising fresh questions about their forecasting edge.

Did prediction markets call the U.S. midterm election?: State of Crypto
WitnessDid prediction markets call the U.S. midterm election?: State of CryptoFirmbee / pixabay

Outputs

  1. Prediction markets matched polls in most tracked midterm races but missed at least three contests

  2. In Alabama's 2nd District, Shomari Figures won despite polls and markets showing Rhett Marques leading

  3. In Iowa's Senate race, Josh Turek beat Ashley Hinson after markets favored Hinson and polls sat at 50/50

  4. Wisconsin's 1st District race was rated a dead heat by both sources before Bryan Steil lost to Mitchell Berman

  5. The markets-vs-polls debate has run since at least the 2020 presidential election

Prediction markets such as Kalshi and Polymarket largely mirrored polling — and the eventual results — across most of the key races CoinDesk tracked in this year's U.S. midterm elections, but they missed at least three contests outright, according to the newsletter's post-election review of polls, market pricing and final outcomes.

The misses matter. In Wisconsin's First House District, Republican Rep. Bryan Steil lost to Democrat Mitchell Berman. Both polls and prediction markets had rated that race a dead heat as late as the Friday before Election Day.

In Alabama's Second House District, the error was sharper. Incumbent Rep. Shomari Figures defeated challenger Rhett Marques — even though both polls and prediction markets had shown Marques leading in the redrawn district. Neither information source caught the actual result.

Iowa delivered the third divergence. State Rep. Josh Turek defeated Rep. Ashley Hinson, flipping the Senate seat. Prediction markets had given Hinson the edge; averaged polls had put the race at a dead 50/50.

What does this change for election forecasting?

The debate over whether prediction markets beat polling models has run since at least 2020, when CoinDesk's predecessor newsletter tracked polls and early prediction platforms ahead of that year's presidential contest. Four election cycles later, the honest verdict is that the markets have carved out a durable niche — while leaving their core value proposition unresolved.

The "blue wave" outcome revived the argument on X, formerly Twitter, over whether platforms like Kalshi and Polymarket were more accurate than pollsters. On the CoinDesk-tracked races, the answer is: mostly the same, occasionally worse, never clearly better.

There's also a chicken-or-egg problem the midterms didn't settle. Did prediction markets follow polling sentiment, or did they genuinely predict where sentiment was heading? The Wisconsin, Alabama and Iowa results suggest the markets absorbed polling inputs rather than seeing around them — when the polls were wrong, the markets were wrong with them.

A useful frame, per the newsletter's author: read prediction markets less as a gauge of polling sentiment and more as a gauge of confidence in a given outcome. That reframing doesn't settle the accuracy question. But it clarifies what the prices actually encode — collective conviction, not independent foresight. Even a reporter who has followed these markets for years reports they have not replaced polling data or other tools in a personal workflow.

Whether that niche serves, in the author's phrasing, "degenerate gamblers" or legitimate market insight remains an open question — one the platforms' own track record has yet to answer decisively.

What comes next?

The unresolved accuracy question now heads into another presidential cycle with prediction markets operating under fuller regulatory footing in the U.S. than in 2020, meaning the 2026 midterms and 2028 contest will offer the first clean, multi-platform test of whether event contracts can outperform polls rather than shadow them.

via cryptonews.net (Source)

More from Marcus Bennett

Marcus Bennett

Show full bio

Senior reporter covering money, funding and market shifts around crypto and web3 industry at Mempool Brief.

1,192 articles