Meet the median Polymarket trader. Over six weeks this spring she opened the app on ten different days, tapped through forty-six trades, and staked about six dollars and fifty cents each time. Money moved in and out; a little over six hundred dollars passed through the account. When the six weeks were up she was down less than two dollars. Not a whale. Not a cautionary tale. Just a person who paid almost nothing to sit at the table for a while.

That portrait comes from the Pew Research Center, which pulled the public trading records of nearly twelve thousand Polymarket accounts - all of them active on ten of the platform's biggest markets in early 2026 - and followed every trade from the seventh of May to the nineteenth of June. It is the closest thing yet to a census of who actually uses a prediction market, and it lands two days after the World Cup final left a very different impression.

The millions were real. They were also almost nobody.

We spent this past week on the handles that took seven figures out of the tournament market and the ones who bled eight. Pew's data is the counterweight. Strip out the celebrity wallets and the story of the typical account is almost boring: small stakes, frequent taps, and a result that lands within a rounding error of where it started. More than half of the traders Pew studied - fifty-eight percent - finished the six weeks having gained or lost less than a hundred dollars. The person clearing a million and the person losing one are both real. They are also both a rounding error on a base of twelve thousand.

The wins and losses that did stack up were roughly even at the top. Seven percent of accounts came out more than a thousand dollars ahead. Nine percent finished more than a thousand behind. For most of the middle, the platform was neither an ATM nor a trap. It was a churn that netted to about zero.

The part that is not boring

The uncomfortable number sits at the far end of the activity curve. About one account in nine - eleven percent - placed more than a thousand trades in the same six weeks. These were not people who checked in on the days a big market resolved. They were active on thirty-nine of the forty-two days. Who needs a day off. And frequency did not buy them an edge: the typical trader in that group finished down about a hundred and forty dollars, and a third of them lost more than a thousand. The people trading the most were, on balance, losing the most - not because the market cheated them, but because there is no volume of ten-cent probability calls that turns a coin flip into income.

That is the shape a prediction market shares with a slot lever and not with a savings account: the more you pull, the more the small edge against you compounds. The median trader treads water because she barely plays. The heaviest traders sink because they never stop.

People trade what they watch

The rest of Pew's read is human in a gentler way. Sports drew the busiest fingers - a median of sixty-nine trades an account - while crypto sat at fifty-nine and politics, the category that made Polymarket famous, trailed at thirteen. Almost a quarter of accounts touched only one subject the entire time. People are not building diversified probability portfolios. They are following the thing they already care about - the match, the coin, the race - and putting a few dollars on their read of it.

What the census actually says

Set the two stories side by side, and the honest summary of Polymarket in mid-2026 is this. A tiny number of accounts win or lose life-changing sums, and they get all the headlines, including ours. The overwhelming majority stake pocket change, trade what they watch, and end roughly where they began. And a compulsive minority trades constantly and quietly loses, which is the one part of the picture that should not be read as harmless.

A price on Polymarket is a live estimate of a probability - useful to read, honest when the window is long, worth exactly nothing as a promise. The five wallets that took the millions understood they were pricing a gap. The eleven percent who never took a day off were treating a probability like a paycheck. Between them sits the person Pew actually found: down two dollars, entertained, and not pretending it was an investment.