The number that stopped US mid-scroll

76.8% of everything that's ever landed on the Dollar Bets board showed up for exactly one day and then vanished. Six hundred and thirty-nine markets, out of 832 unique ones we've tracked since April 25, got their fifteen minutes and ghosted. The board is a goldfish with excellent taste.

This is our own first-party data, 162 days of daily snapshots of our curated top-10(ish) market board, archived since late April and run back through our own files. 832 unique markets have rotated through 1,608 total board slots.

The headline numbers

Over seven full months plus a slice of November, the board swung from a quiet 58 slots in April up to 300-slot months in May, June, and October. Median daily volume on featured markets ranged from a sleepy $2,601 in early November to a muscular $25,285 in June. Average payout per dollar bet drifted between $16.93 and $39.65 depending on the month.

The longshots that made US laugh out loud

Ten markets hit our board paying out at $500 or more per dollar, and the top tier, six separate markets paying a flat $2,000, all lasted exactly one day each. Jonathan McColumn's run at Georgia's GOP Senate nomination got filed as "a lottery ticket filing for office." WTI crude hitting $175 became "someone's short-selling the sun." Crypto showed up twice in the $2,000 tier, including Ethereum moving a specific direction in a specific one-hour window, "a dollar on whether the number goes left or right."

The markets that wouldn't leave

  • NASA landing a person on the Moon before 2028, 81 days, first spotted April 26, still live November 3: "one small bet for man"
  • Starship carrying humans to Mars before 2030, 30 days
  • A humanoid robot walking on Mars before any human, 24 days, "the Terminator but make it NASA's problem"
  • California's 8.0 earthquake market, 24 days, "shaking the liberals to their very core"
  • Taylor Swift getting a papal audience, 22 days, "Bejeweled, Vatican edit"

Notice the pattern: space, earthquakes, and Taylor Swift are apparently the three load-bearing pillars of this board's longevity. Everything else is a day trip.

One-day wonders, or: the churn problem

639 of 832 markets, again, 76.8%, never saw a second sunrise on our board. Prediction markets tied to single games, tweets, or news cycles get scanned, featured, and replaced fast. Rippling IPOing before Deel managed 19 days, which by this board's standards makes it basically a tenured professor.

The wildest odds swings

The single biggest price move belongs to "Trump drops more UFO files this month", yes-price swung 62 points in six tracked days. Sports had its own fireworks: the single-season receiving yards record moved 46 points in three days, and the single-season sack record moved 44 points in three days. Climate markets were the slow burn, "Will 2026 be the hottest year ever?" moved 37 points, but stretched over 15 days.

Month over month: what the board actually looked like

April was a 58-slot trial run dominated by Science and Technology and Sports. By May and June the board exploded to 300 slots a month, with Other and Science and Technology trading the top category spot almost every month after. Volume peaked in June at a median $25,285 and mean payout of $39.65, with June, July, September, and October all topping out at $2,000 max payouts. November, three days in, is running thin at 30 slots and $2,601 median volume.

About this data

Our board is a curated daily top-10(ish) selection, not the entire universe of Kalshi and Polymarket markets, this measures what made our board, not "the market" writ large. There are snapshot gaps (12 days after July 25, 3 more after August 11, 14 days from late August into September 10), so some markets' true days-on-board could run longer than recorded. Category tags are our scanner's own labels, useful for trends, imperfect in the weeds.

Frequently asked questions

Is this every prediction market that exists?

No, it's every market that made our daily curated board across 162 snapshot days. Think of it as a sample of what we found interesting enough to feature, not a full market census.

Why do so many markets only last one day?

Most of what lands on the board is tied to a specific event, game, or news cycle that resolves fast, so our curation naturally rotates fresh markets in constantly, 76.8% of everything we've tracked only ever showed up once.

What's the longest-running market on the board?

"Will NASA land a person on the Moon before 2028?", 81 days on the board and counting as of November 3, making it the board's senior citizen by a wide margin.

Are these payout numbers still accurate?

No, payouts shown here are historical snapshots from when each market was featured; prices move, and none of this is an indication of what a market pays today. Check the live board for current odds before anything else.