“Buys Bitcoin markets at 20–40¢ and sells them at $1”: the bot buys both sides, and its “net profit” is before $88,000 of fees

The short version

The post, by @armouredme, was published on X on November 3, 2026 and had about 3,400 views by that evening. It says a bot run by the Polymarket account @stepahead “buys Bitcoin markets at 20-40¢ and sells them at $1”, trades only the five-minute Bitcoin Up or Down markets, and has made $108,191 in net profit across 22,829 trades. Its example is a “Down” position bought at 20.8¢ that made $5,149, a 380% gain. The explanation: at 20–25¢ the bot only needs to be right one time in four, and each win pays three to four times what it risked. The account link carries a ?via=Claude-bot tag.

The account is real and the headline numbers match Polymarket's records. Its profit history shows $107,926 of trading profit at the start of November 3 and $108,411 by the evening, and Polymarket counts 22,985 “trades”, which is about the number of markets it traded: the account made 1.64 million individual fills across 22,978 markets, 21,014 of them five-minute Bitcoin markets and nearly all the rest fifteen-minute ones. But $108,411 is profit before fees. The account has paid $88,111 in trading fees, which leaves $20,252, the figure on Polymarket's all-time leaderboard. Polymarket has also paid it $70,695 in maker and taker rebates, for an all-in total of about $91,000. Nearly all of it came in July and October; the account lost $45,343 in its first five weeks, March and April, and was idle in May, June and August.

The method is where the post goes wrong. We pulled all 45,091 of the account's closed positions, one for each side of each market. In 22,113 of its 22,978 markets, 96%, it bought both Up and Down. 86% of all the shares it bought were matched pairs, an Up share and a Down share in the same market, which pay exactly $1 between them whatever Bitcoin does. Bought together for less than $1, a pair is a locked profit; for more than $1, a locked loss. The median pair cost it 96.9¢. Markets where its pairs cost under $1 made $314,000; markets where they cost more lost $277,000. The extra shares left over on one side, the only part that is a bet on Bitcoin's direction, lost $291,171 in total, and lost money at every pair cost.

The post's own example shows it. The Down position bought at 20.8¢ on March 29 did make $5,149.05. In the same five-minute market the account also bought 6,232 Up shares at 69.3¢, and they lost $4,316.75. The market made $832.30. The screenshot in the post shows four more winning positions, at 42.2¢, 36.5¢, 37.5¢ and 52.2¢, two of them outside the post's “20-40¢”. Each has a losing opposite side in the same market: the $9,156 win sits next to a $7,322 loss, and the $5,650 win next to a $5,598 one. The account held its positions to the market's close (their results match a payoff of $1 or $0 per share), so “sells them at $1” is accurate only in that sense.

Verdict: the account and its trading volume are real, and the post's profit figure is Polymarket's own number before fees. Everything else describes a different strategy. The account is not betting on cheap sides that win one time in four; it buys both sides of nearly every market, makes money when the pair costs less than $1, and loses on the shares it ends up holding on one side. After fees, its own trading has made $20,252; most of its total profit comes from Polymarket's rebates. We have checked this template before: a similar post about another account, and a wallet that profited the opposite way, from its one-sided leftovers.