Shallow Markets, Deep Risks: How Low Volume and Bots Expose Prediction Market Traders to Volatility

Despite a surge in activity, particularly during the 2024 election and last fall, the vast majority of prediction market contracts on platforms like Polymarket rarely see significant trading. A recent CNBC investigation revealed a concerning trend: roughly 70% of all closed markets on Polymarket, analyzed from 2021 through May of this year, reported less than $10,000 in total volume. This phenomenon leaves traders vulnerable to extreme price swings and financial inefficiencies.

The CNBC analysis, drawing from Polymarket's Gamma API which records notional volume on both sides of a trade, highlighted that less than 10% of closed markets ever reached between $100,000 and $1 million in reported volume. Furthermore, a staggering 45,000 markets, nearly 5% of the total, recorded no volume whatsoever.

Polymarket's competitor, Kalshi, also faces similar challenges with numerous shallow markets, according to data analyzed on the on-chain platform Dune. Unlike Polymarket, Kalshi's notional volume on Dune counts only one side of a trade.

Academics warn that these low-volume environments are far from ideal for participants. Constantin Bürgi, an economics professor at University College Dublin, explained, "Thin markets by nature imply that small investments can result in large market movements and are typically more volatile." This inherent instability can make trading unpredictable and risky.

Dartmouth College economics professor Eric Zitzewitz added that new traders face particular peril in these shallow markets. Wide spreads between buying and selling prices can significantly increase the cost of transactions, making it a more expensive endeavor than anticipated.

Seasoned Traders Seek Deeper Waters

Experienced traders, often seeking capital efficiency, also prefer markets with higher liquidity. Logan Sudeith, a former financial risk analyst now trading full-time, noted his preference for "higher volume, short term [markets]" due to their efficiency. Markets resolving within a week, especially those related to high-profile events like the U.S. presidential race, the war on Iran, or figures like Donald Trump and Elon Musk, tended to attract higher volumes on Polymarket, with some exceeding $1 million.

Zitzewitz further emphasized that traders gravitate towards markets nearing resolution and those with a substantial number of participants, suggesting a preference for clarity and collective engagement.

Bots Dominate the Shallower End

A significant factor in low-volume markets is the overwhelming presence of automated trading bots. Joshua Della Vedova, a business professor at the University of San Diego, found that over 80% of the volume in Polymarket markets under $10,000 originates from bots. His research identified wallets executing more than 50 trades daily or over 1,000 total trades as bots.

While bots are profitable across the board, earning roughly $1.2 million in shallow markets and $35.1 million in markets exceeding $10 million from November 2022 to February 2026, their presence doesn't necessarily distort fair value due to the inherent risks of thinly traded markets. Della Vedova noted that bots ultimately prefer heavier markets for their profit-per-transaction model, even if they operate across the entire spectrum.

Accuracy Debates in Thin Markets

The reliability of thin markets remains a point of contention among experts. Evercore ISI strategists, after analyzing five years of data from Polymarket and Kalshi, concluded that high-volume markets provide more reliable probabilities. They highlighted that a mere 8% of markets achieved $1 million in volume, implying that "most quoted probabilities sit in the thinly traded tail – where calibration is weakest."

Conversely, other researchers, like Yale University finance professor Theis Ingerslev Jensen, argue that accuracy is less about volume and more about the caliber of traders. Jensen and colleagues found that skilled or informed traders were the primary drivers of accuracy on Polymarket. He cautioned that "Thin markets are not automatically inaccurate, but they are less reliable," emphasizing the critical role of incentivized skilled traders.

Enduring Influence, Heightened Awareness

Despite these findings, Harry Crane, a professor of statistics at Rutgers University, believes the prevalence of shallow markets is unlikely to fundamentally alter how prediction markets are viewed by the public or Wall Street. He advises that while trading volumes must be considered, a lack of liquidity alone doesn't discredit a market's signal or render it economically useless.

Polymarket chose not to comment on CNBC's findings, and Kalshi did not respond to requests. As prediction market volumes continue their rapid growth, Crane anticipates larger markets will expand further, while low-volume markets may remain shallow. He underscores the importance of trader awareness, advising, "Protect yourself at all times. Each individual entity needs to address them on their own."

Methodology Spotlight:

CNBC gathered all closed market data from Polymarket's Gamma API, spanning 2021 to the end of May 2026, noting that this API reports notional volume on both sides of trades. This analysis was independently verified by Joshua Della Vedova of the University of San Diego, whose cross-checking with a separate 222-million-trade dataset confirmed CNBC's findings.