The pitch on TikTok and X sounds simple - bet on the news, get rich.
The math says no. About 69% of Polymarket traders lose money. Most winners aren't even human. The big platforms are now a fight where pros and bots take cash from a much larger crowd.
Polymarket and Kalshi let users bet real cash on events. Think elections, jobs reports, sports, even weather.
The pitch is simple. Read the news better than the next person and win.
The data tells a different story. The top 1% of traders take three out of every four dollars in profit. And 14 of the 20 best wallets on Polymarket are run by bots. Those bots aren't side hobbies.
They are trading systems built by hedge funds and crypto firms. These pros saw prediction markets as just another order book to game.
Bots don't sleep, don't doubt, and don't chase a hunch. They watch hundreds of markets at once.
They trade on tiny gaps before any human can react. That edge adds up across thousands of small trades.
Most traders lose to it without knowing why. A new study found that only 3% of traders move prices the right way.
That tiny group plus the bots wins. The other 97% feed them the cash.
The losses are real. Over 100,000 wallets each dropped at least $1,000 since January 2025. Their losses, $131 million in all, flowed to a small group of pro traders.
Researchers ran a check on the top winners. They tested each one against a fresh set of events.
Only 12% beat a basic mark. About 60% of the so-called winners turned into losers on the second test.
In English: most people who think they are good at this just got hot for a stretch. That has stakes beyond gambling.
The press, hedge funds, and even regulators now use prediction-market odds as forecasts. But those odds are shaped by a tiny group of pros and bots.
Prediction markets ran on a clean idea. Crowds with cash on the line beat any single expert.
That worked when the platforms were small. Most traders back then were just informed fans.
Big money changed it. Once the contracts grew big enough to trade like stocks, pro firms moved in.
The same edges that beat retail in stocks now beat them in election odds. The crowd is still in the trade, just on the losing side.
Prediction markets are still growing fast. The CFTC and Congress are now weighing how to treat them as real money tools.
The volume on Polymarket and Kalshi has surged in the past year. So has the share of trading run by automated systems on both sides of the market.
The bigger question for the rest is whether the bots ever leave room. So far, the numbers say no.
Researchers can already answer that one for the public.
