We tracked $4 million of Kalshi whale trades. The big money lost
1,855 Kalshi trades of $1,000 or more lost 4.9%. Copying them faster didn't help, because there was no move to catch. Plus our own bug.
On 25 August someone on Kalshi put $46,000 on one La Liga match. Trades like that print all day, and Kalshi's public tape shows every one of them, without the name. If you have ever watched one go by and thought, they must know something, this is for you.
We took the first trade of $1,000 or more in each of 1,855 settled markets across 15 kinds of market: tennis at every level, baseball games and totals, NFL, college football, WNBA, the Premier League, La Liga, UFC, and Bitcoin's 15 minute and daily markets, from 7 August to 6 September 2026. About $4.0 million in all. We scored each trade held to the end, with Kalshi's real fee.
The big money lost. $196,191 on $3,996,520, or minus 4.9 percent. Per contract, minus 2.92 cents after the fee and minus 1.27 cents before it. The whales won 67.9 percent of the time while paying 69.1 cents on average, so they won slightly less often than their own prices said they would, and the fee did the rest.
Copying them faster didn't help. We bought the whale's side at the first trade a quarter of a second, 2 seconds and 10 seconds later. Minus 2.92, minus 2.98 and minus 3.01 cents a contract. A quarter of a second after a whale trades, the price had not moved at all in 60.8 percent of markets, and on average it moved by nothing. There was no move to catch.
The Bitcoin version fails too. When Bitcoin jumps on Coinbase, does Kalshi lag? Within a quarter of a second the 15 minute price had already moved about 4 cents, and copying the move lost money on Bitcoin and on the other coins.
Bigger whales lose less, but still lose, and no kind of market showed whales winning by more than noise. Members get the full table by market and by size in the appendix.
Our mistake, in public. The first run of this study said whales made 14 cents a contract, and that speed was the whole problem. The tape stores the price each buyer paid, but our script flipped the price of every No trade, so a 99 cent bet was scored as a 1 cent bet. We fixed it and re-ran every number on this page. Our earlier copy trading video said the price moves before you can follow a big trade. Measured, it mostly doesn't: the whales are a little wrong, and the fee does the rest.
So big is not the same as informed, and speed in the range a small bot can reach doesn't change that. If you want to test a copy rule of your own, use prices real traders actually paid, and write the pass mark down before you look. We're still hunting, and this is one more door closed.
Prices from Kalshi's public trade tape; trades are anonymous. Aggregates only. Nothing here is advice.