Separate analyses published in early 2026 have delivered a bleak verdict on prediction market fairness. Data reported by CoinTribune and Cryptonews found that 84.1% of traders on Polymarket operate at a net loss, though it has not been confirmed whether this figure applies to Polymarket specifically or to the prediction market sector more broadly, including platforms such as Kalshi. A distinct analysis published by Finance Magnates found that 14 of Polymarket's 20 most profitable wallets belong to automated bots. The findings arrive as Polymarket records its first $10 billion monthly trading volume, crossing that threshold in March 2026 and closing Q1 with $26.2 billion in total activity, up over 90% quarter-on-quarter.

The scale of the imbalance is stark. Research reported by CoinTribune, drawing on a dataset of more than 2.5 million wallets, found that only 2% of wallets have ever earned more than $1,000 cumulatively, and just 840 wallets (representing 0.033% of accounts studied) have crossed $100,000 in lifetime gains. A January 2026 SSRN preprint analyzing comparable data found that the top 1% of accounts capture 84% of all realized profits.

A separate analysis from Citizens JMP Securities found that profitable returns correlate almost entirely with trading scale: participants with more than $500,000 in lifetime activity recorded a median return of positive 2.6%, while those below $100 in lifetime volume lost a median of 26.8%. Traders in an intermediate range also posted negative median returns, reinforcing a pattern in which scale consistently predicts outcomes across the distribution.

Retail participants also fare worse here than in traditional sports betting. Research published by CoinDesk in March 2026 found that prediction market users posted a median return of negative 8% between mid-2025 and mid-March 2026, compared with negative 5% for legal sportsbook bettors over the same period. The structural reason is important: sportsbooks cap or restrict winning accounts to manage their own risk exposure. Prediction platforms do not. That means consistent winners, including bots, face no limits, and retail order flow effectively becomes the liquidity pool from which professional traders and automated systems extract value. As a Citizens JMP Securities analysis put it, retail participation in prediction markets "functions as liquidity for sophisticated traders, with outcomes determined by execution, speed and pricing rather than forecasting accuracy alone."

The bot advantage is technical and widening. A paper titled "Unravelling the Probabilistic Forest," published in August 2025, estimated that automated arbitrage traders pulled roughly $40 million from Polymarket over a single year between April 2024 and April 2025. According to figures reported in industry publications and pending verification against primary academic sources, the average window for a viable arbitrage trade has compressed from 12.3 seconds to 2.7 seconds, and 73% of arbitrage profits now go to bots executing in under 100 milliseconds.

One documented wallet, labeled 0x8dxd on the public leaderboard, converted approximately $300 into more than $437,600 in about one month by exploiting price differences between Polymarket contracts and spot prices on Binance and Coinbase.

Bots also profit by identifying logical contradictions between related contracts, by buying both YES and NO positions simultaneously when their combined price falls below $1.00 to lock in risk-free gains, and by exploiting pricing divergence between Polymarket and competing platforms such as Kalshi. That kind of sub-second execution is effectively out of reach for users on consumer devices or standard mobile connections.

A working paper from London Business School and Yale reviewed 1.72 million accounts and $13.76 billion in trading volume between 2023 and 2025 and found that just 3% of traders actually move prices toward accurate outcomes. The remaining 97% add volume and liquidity but do not, in aggregate, sharpen market accuracy. When researchers simulated each trader's bets 10,000 times with randomized directions and then tested results on separate event samples, roughly 60% of apparent "lucky winners" became net losers, suggesting that a large share of seemingly profitable trading reflects chance rather than skill.

"Just 3% of traders account for most price discovery, meaning they are the ones moving prices toward the correct outcome," the paper concluded. "The remaining 97% mostly do not."

Concerns about information asymmetry extend beyond execution speed. Researchers documented three wallets that collectively earned more than $630,000 by betting on Venezuelan President Nicolás Maduro's removal from power before a classified US operation became publicly known. The episode illustrates a structural vulnerability in permissionless markets: without gatekeeping mechanisms, non-public information can be converted into trading gains at the direct expense of counterparties who lack access to it.

The implications are particularly sharp for users in Africa and South Asia, two regions where prediction markets are expanding rapidly. Luno launched prediction markets in Nigeria and South Africa in March 2026, partnering with US-based Limitless to offer 24-hour crypto price contracts starting at 3 USDC. Luno CEO Ayotunde Alabi framed the launch as a natural fit: "Prediction Markets are a natural evolution of how our customers already engage with cryptocurrency, and there is a strong customer demand for this product." But Nigeria alone counts an estimated 168.7 million bettors, many of them accustomed to mobile sports betting platforms.

The regulatory environment adds a further layer of complexity to this expansion. In 2025, Nigeria's Lagos State Lotteries and Gaming Authority classified local prediction platform Bayse Markets as an illegal gaming operator, illustrating the regulatory grey zone in which new entrants like Luno are now operating. Not all observers frame the regional picture in purely negative terms: analysts cited in TechCabal have noted that prediction markets could serve as B2B information infrastructure for African businesses, functioning as professional signal tools rather than retail gambling products. That potential does not, however, alter the structural disadvantages facing individual retail participants.

That population is consistent with the profile of the sub-$100 lifetime volume cohort that research consistently shows loses the most, though a direct mapping between Nigerian retail bettors and that specific cohort has not been established in the underlying studies.

In South Asia, where crypto adoption grew 80% year-on-year in the first half of 2025 and transaction volumes rose from $1.4 trillion to $2.36 trillion, retail users face comparable structural barriers, though internet infrastructure varies significantly across India, Pakistan, and Bangladesh.

On short-duration contracts, the evidence consistently points in one direction: the shorter the market, the larger the bot advantage.

US regulators are beginning to respond. The BETS OFF Act, introduced in March 2026, targets contracts on war, terrorism, and assassination outcomes. A separate bill introduced in the same month, the DEATH BETS Act, addresses death- and war-outcome betting. The two pieces of legislation cover overlapping but distinct contract categories.

Comprehensive legislative responses to bot dominance, profit concentration, and retail consumer protection have yet to be developed. With Polymarket's single-day volume hitting $425 million on February 28, 2026, and platforms expanding into new markets across Africa and Asia, those questions will become harder to avoid.

Note: Several figures cited in this article originate in SSRN preprints and industry publications, including at least one primary source whose content could not be fully verified at the time of writing. All statistics should be treated as preliminary until confirmed against the original academic papers.