🚨 INVESTIGATION: Polymarket US public data reveals 100x volume inflation, 100M+ missing trade records, and systematic wash-trading patterns.
🔴 These are 4 documented cases proving these issues exist across just a short window of dates in 2026. A full audit of all public data will uncover much more.
🔴 Evidence, raw CSV files, and proof linked in the thread below 👇
1. Two markets reported 100x their actual trading volume
>> On September 23, Polymarket’s official data showed about $190.81 million traded across two tennis markets, explicitly labeled in USD (notionalTraded.value).
>> When you add up all 22,057 individual trades from the raw logs, the actual total comes out to just $1.91 million.
>> The contract counts and timestamps match, proving it is the same trading activity. Just reported at 100 times its real dollar value.
2. Almost half of July 1’s trades are completely missing
>> Polymarket’s daily summary stated that 219.14 million contracts traded on July 1.
>> However, its detailed public trade list only accounted for 118.72 million contracts.
>> That leaves 100.42 million contracts (nearly 46% of the day’s volume) completely missing from the public trade list.
>> The trade file abruptly stopped recording after 3:30:40 AM Eastern. Yet trading was still actively happening—separate market books (like England vs. DR Congo) show trades over 10 hours later at 2:06 PM Eastern.
3. Textbook wash-trading loops on the House-control market. On September 21, between 21:22 and 21:40 UTC, an identical pattern repeated 8 times in a row:
>> 62 contracts were bought at 91¢.
>> 62 contracts were sold at 90.9¢.
>> Open interest rose by 62 after the buy, then immediately dropped by 62 after the sell.
>> Later that night (22:13–22:38 UTC), another 10 trades of 62 contracts did the same thing with a synchronized Buy-Buy-Sell-Sell loop (+124 open interest, then right back down).
4. Public trackers double-counted old volume
>> A trade happened on June 11. Polymarket included that exact same old trade in files for June 12 and 13.
>> Public aggregators like DeFiLlama ingested those files to calculate daily volume and counted the old trades again.
>> We identified 2,357 duplicate records that reappeared on June 12 and 13, adding $114,187 in old trades as "new" volume (~0.23% extra volume on June 13).
🔴 Evidence, raw CSV files, and proof linked in the thread below 👇