coindesk.com
A savvy trader reportedly pulled off a high-stakes manoeuvre on the predictions platform Polymarket, pocketing $233,000 and outfoxing automated bots during quiet weekend trading. It has captured widespread attention, with some labeling it outright manipulation that demands disciplinary action.
The gambit took place late Saturday night, when trading volumes across the crypto ecosystem tend to be low, and small buy/sell orders often have an outsized impact on market prices.
CoinDesk reached out to Polymarket for a comment on the matter.
The playbook
Here's how it unfolded, according to details shared by pseudonymous Polymarket trader PredictTrader on X – a post widely circulated in industry circles.
A trader with the pseudonymous identity @a4385 began aggressively purchasing the "UP" shares at any price in the Polymarket contract that asked whether payments-focused cryptocurrency XRP's price would rise or fall between 12:45 PM ET and 1:00 PM ET on Jan. 17.
As the minutes passed, aggressive buying pushed UP shares to 70 cents, even as XRP's spot price on major exchanges weakened by 0.3% over that period.
The diverging trends saw Polymarket bots, programmed to create buy/sell orders and provide liquidity and arbitrage or exploit price inefficiencies across the two markets, automatically sell more UP shares as their prices rose, even as XRP's price dipped.
This allowed the trader to corner a whopping 77,000 UP shares at an average price of 48 cents.
What happened next gives the bet a Wolf of Wall Street-like vibe. Two minutes before the market settled, a Binance wallet reportedly linked to the trader made a major $1 million purchase of XRP, which drove the price up by about 0.5%.
This timely XRP price spike ensured the Polymarket contract settled as a win, with UP shares qualifying for redemption at $1 each, a significant profit to an average acquisition cost of 48 cents.
The trader then sold off the purchased XRP on Binance, dumping the spot price back down. The whole operation cost the trader roughly $6,200, while bots lost a year of profits overnight, according to data source PolymarketHistory.
The trader didn't stop there and replicated the tactic across multiple thin weekend markets, systematically draining bot liquidity. While some bots adapted and shut down, others weren't so lucky.
Need for smart bots
The episode underscores vulnerabilities in Polymarket's automated market-making bots, which treat every price move (tick) as the same. They are blind to critical dynamics such as volume, liquidity regimes, adversarial tactics, and shifts in incentives near settlement.
This calls for the development of context-aware (smart) bots that adapt on the fly and understand who they are trading against, at what time of the day, and when not to participate.
Chris Tremulis, global head of commodities compliance at Goldman Sachs, voiced concerns on X, saying preserving market integrity is key to institutional adoption of betting platforms.
"Prioritizing market integrity will be key for the leaders of prediction markets to achieve meaningful institutional adoption (if that’s what they want)," Tremulis said.
"Stronger rulebook enforcement, quicker investigations from exchange surveillance staff, published disciplinary outcomes, and referrals to the CFTC will go a long way. Early days," he added.