Poker analyst Vojtech Cervinka has published a list of 11,469 QQPK tournament nicknames that he said he was “virtually certain” belonged to bots, substantially exceeding the 893 accounts the operator has permanently banned in its own investigation.
Cervinka’s review covered 24 million player-hands from 3.78 million hand histories played between July 16 and Oct. 11. He estimated that bots played about 17.5 million hands, or more than 70% of that dataset, while cautioning that several hundred more suspicious nicknames lacked enough recorded hands to meet his threshold for high-confidence classification.
The published file says the selections came from unusual and highly consistent strategic patterns, rather than from account names. It also warns that the 11,469 figure is a count of unique nicknames, not necessarily distinct underlying accounts, since nickname changes can affect the total.
Naming conventions were presented as supporting, circumstantial evidence of bulk account creation. The file includes 135 names beginning with “Happy”, 245 names formed from five letters and nine digits, and a 70-name grid combining seven Thai prefixes with the same 10 poker-related suffixes. It explicitly says these patterns do not prove that an individual account was automated or that the names were generated by an LLM.
Cervinka highlighted one account, 郑鱼F, which appeared in 153,000 hands in his database. The next most active suspected account appeared in 5,000 hands. Its SharkScope profile showed 8,329 tournaments in 33 days, an average of more than 250 a day.
As a control, Cervinka applied his detection filters to a PokerStars multi-table tournament database and found no players displaying bot-like signs. After relaxing the criteria, he said the process returned “just three random fish.”
The new list is separate from an anonymous community investigation published Oct. 9, which flagged 2,587 QQPK accounts from a different sample period, July 2 through Oct. 2. That earlier review examined roughly 26 million player-hands and cited recurring statistical similarities, repeated preflop choices and consistent bet sizing among the suspected accounts.
As reported earlier this week, QQPK ultimately banned 893 accounts and created a $300,000 compensation fund after its investigation, saying 567 bans came from the community list and 326 from its internal review. The operator said a further 2,020 flagged accounts remained under monitoring and attributed the confirmed activity to a collusive syndicate sharing hole-card information at the same tables.
QQPK said its review found no manipulation of its random-number generator, backdoor access or means to view unrevealed hole cards, conclusions that have not been independently verified. It planned to publish affected-tournament details and compensation calculations by Oct. 17 before crediting eligible players.
