एक नए ऑडिट से पता चलता है कि पॉलीमार्केट (Polymarket) के मुख्य पसंदीदा विकल्पों का समर्थन करने पर 4% का नुकसान हुआ होता।

According to an October 9 scorecard from blockchain data provider Bitquery, a hypothetical approach of backing the leading candidate on prediction market Polymarket across every evaluated 2026 U.S. primary would have resulted in a 4% loss, despite these candidates securing victory 87% of the time. This outcome highlights the distinction election bettors face: determining who holds the highest probability of winning versus deciding if a contract represents a fair price.

Bitquery’s analysis revealed that candidates holding the study’s highest reference price emerged victorious in 238 out of 273 governor, House, and Senate primaries. Nevertheless, the simulated strategy of wagering $1 on each favorite at the designated reference price yielded a loss of four cents on every dollar.

The overall accuracy rate was largely driven by the most dominant favorites. Candidates valued at 90 cents or higher triumphed in 182 out of 182 races. Conversely, among favorites priced from 50 to 90 cents, 71% won despite carrying an average price tag of 77 cents.

Because an outcome share yields $1 upon winning and zero if it loses, purchasing a costly favorite generates minimal profit for a winning position, whereas a single defeat completely erases the initial purchase amount. Consequently, a series of losses can quickly overshadow a high volume of correct picks.

Allocating an identical dollar amount across every contest also acquires varying quantities of shares. Winning with a cheaper contract generates a bigger return for that specific stake than winning with an expensive one. While measuring mere prediction accuracy treats every race equally, calculating financial returns requires factoring in those varying payouts.

Reference prices limit the betting conclusion

To establish these metrics, Bitquery calculated average trades during the 24-hour window leading up to 12:00 UTC on election day. If a candidate experienced no trading activity within that timeframe, the provider utilized the final transaction recorded over the preceding 30 days. For races requiring a runoff, the runoff date served as the reference point.

These averages, alongside potentially outdated fallback transactions, function as reference prices rather than guaranteed execution bids at the cutoff time. The published framework lacks a comprehensive adjustment for slippage, spreads, or fees. Although Polymarket’s fee page currently indicates a political fee bracket spanning 0% to 1%, it does not define the exact costs applied to the historical trades featured in this dataset.

The research omitted nine markets because certain contracts settled prior to voting, others lacked usable pricing data, and one was missing a settlement log. The evaluation focused exclusively on primaries across all 50 states utilizing Polygon network trade data, omitting Polymarket’s domestic U.S. application and alternative platforms such as Kalshi.

When analyzing November prediction markets, the separation between forecasting accuracy and pricing value remains essential. Bitquery notes that the historical primary data may not directly translate to broader contests, warning that general elections naturally draw significantly more financial volume and polling attention than smaller local races.

अक्सर पूछे जाने वाले प्रश्न

• Did Polymarket favorites win their races? Yes, candidates designated as favorites won 87% of the time, accounting for 238 out of 273 evaluated primaries.
• Why did the favorite-backing strategy lose money? Buying expensive favorites leaves minimal profit on winning positions, meaning that a few losses can outweigh a high number of correct predictions.
• What data was used for the study? Bitquery relied on Polygon blockchain trade data covering primaries across all 50 states, excluding Kalshi and Polymarket’s dedicated U.S. application.