Kalshi traders took a quarter-point Fed hike this month from 18 cents to 55 cents on Wednesday's inflation data. Daytime selling took back 20 cents of that. Evening and overnight trading, including one heavy block, took 12 cents more.
Wednesday's repricing of an October rate increase on Kalshi has nearly run its full course, and the timing of the retreat is the useful part.
The contract that settles at $1 if the Federal Reserve raises rates by 25 basis points on Oct. 28 stood at 18 cents before the consumer price report and hit 55 cents at 9:06 a.m. By about 6 a.m. Thursday it was trading at 23 to 24 cents. At 23 cents, it keeps 5 of the 37 cents it gained, about 14%.
Stage one: the trading day
Most of the giveback happened while U.S. markets were open. By 1:46 p.m. the contract had fallen to 35 cents, surrendering 20 cents, or 54% of the morning gain, before the Fed's 2 p.m. regional survey described flat consumer spending in most districts.
Stage two: the evening and overnight
The rest came in fewer, larger orders. In a three-minute window starting at 6:43 p.m., about 45,500 contracts traded on the "no" side at 30 and 31 cents, including a single print of about 24,800. No news has been tied to that burst. The price was 27 cents at 9 p.m. and 23 cents at 9:55 p.m., touched a low of 22 cents, recovered to 28 cents by 2:36 a.m. and was back at 23 to 24 cents after 6 a.m. From 35 cents to 23, that second stage removed 12 cents, about a third of the original gain.
The evening stretch followed a selloff in AI-linked debt during the U.S. session and coincided with South Korea's stock market tripping a circuit breaker. A hike becomes harder to justify if financial conditions are tightening on their own, which offers one explanation for selling at that hour. The sequence shows timing, not cause.
The gauges are converging
Three gauges with different participants now sit close together. Polymarket's equivalent market, which traded near 48.5 cents at 9 a.m. Wednesday, was at 24.5 cents at about 6:10 a.m. Thursday. CME FedWatch data, which put the October probability at roughly half on Wednesday morning, showed about 26% around midnight. With Kalshi at 23 to 24 cents, the three cluster between about 23% and 26%.
That convergence is informative. When independent markets agree, a single large order is less likely to be setting the price, and the low-to-mid 20s become a more reliable baseline for judging the next data point.
What would move it
Thursday's 8:30 a.m. producer price and retail sales reports are the next test. A hot producer price print with firm retail sales would argue that the overnight selling went too far, and a move back into the mid-30s would erase it. Soft numbers would point toward the pre-report level of 18 cents. The decision comes on Oct. 28.
Contract prices reflect what traders will pay under the contract's rules. They are not the Fed's forecast.