Coinbase’s (COIN) Brian Armstrong made the prediction markets look silly. Bill Ackman made them look real

Brian Armstrong’s last-minute crypto shout-out turned a $4,000 prediction market into a punchline. With a few words, the Coinbase boss paid out all bets on “Bitcoin”, “Ethereum” and “Web3” at once.

It was a thin market, with the top winner taking home just $111, according to Poymarket Analytics.

But if Armstrong’s impromptu word salad showed how ridiculous prediction markets can get, New York City’s mayoral market, with $22 million in open interest, shows just how serious they’ve become. Moving the odds there by just 10 percentage points would now cost about $1 million in concentrated purchasing power.

That’s because Polymarket’s open interest reflects real money sitting in a liquidity pool, not just the sum of theoretical stakes waiting to be settled. Each trade interacts with an automated price curve backed by collateral, meaning odds change gradually rather than being set by direct match between traders.

The biggest positions show how deep that pool runs: whales like “dubdubdub2” and “asfgh” each has over $2 million backing Zohran Mamdani, according to Polymarket Analytics, while most of the major “No” traders are already sitting on heavy losses and have little room to add capital.

Because of this, any new “out of the money” bet, such as buying Andrew Cuomo at long odds or selling Mamdani close to 95%, is quickly absorbed by the market maker’s curve, which adjusts prices based on supply and demand.

To move the odds by 10 percentage points, a trader has to push through millions of dollars in opposing orders, about $1 million in concentrated buying or selling, before the curve starts to change meaningfully. The size and structure of the market makes small attempts at manipulation almost immediately diluted by existing capital.

The latest opinion polls support the market’s view. A Fox News poll shows Mamdani leading Cuomo by 16 points, while an Emerson College poll puts his lead at 25, proving his 95% odds reflect voter sentiment rather than manipulation.

Perhaps there was also a misunderstanding that polls measure what voters say they will do, while prediction markets measure how confident traders are that those voters will actually do so, so a 50% candidate vote does not necessarily have a 50% chance of winning.

Ackman’s criticism missed what traders already knew: If Mamdani’s 95% odds were really inflated, anyone could have taken advantage of the mispricing.