The QQPK bot scandal has produced a remarkable spread of numbers: 724, 893, 2,587 and about 12,000. Each figure comes from a different source, a different method and a different level of certainty. For players trying to understand how safe online poker is, the dispute is a masterclass in why headline numbers need context.
We covered the operator's response when QQPK banned 893 accounts and set up a $300,000 compensation fund. This follow-up explains why the counts diverge, what the compensation rules actually say, and what the episode means for anyone who plays on offshore or Asia-focused sites.
How the Story Unfolded
According to PokerListings, players on X, TwoPlusTwo and Reddit alleged in October that about 2,587 accounts on QQPK, an Asia-focused poker room, were bots. Their 14-page document analysed 4.18 million hand histories and pointed to similar statistical profiles, round-number 3-bet sizes, consistent limp/3-bet patterns, and all-in and calculation results that were implausibly accurate.
QQPK confirmed receiving the data and froze the flagged accounts on October 10. By October 13, it had confirmed 724 accounts for unauthorised automated play. Its October 15 report described a bot syndicate, criticised the data for selection bias and sampling limits, and said it had reviewed 2,913 accounts, identified 893 bots, and banned them. It also confiscated $64,627 and set up a $300,000 compensation fund.
Separately, investigator Vojtech Cervinka published his own analysis on October 15. He estimated that about 70 percent of the collected hands, roughly 17.5 million, were played by bots, and he suspected 11,469 nicknames of botting, for about 12,000 bots in total. PokerListings noted at publication that he had not received a response to his question about a reward for his work.
Four Numbers, Four Methods
It helps to lay the figures side by side.
| Figure | Source | What it represents |
|---|---|---|
| 2,587 | Community researchers | Accounts flagged in a 14-page statistical analysis of 4.18 million hands |
| 724 | QQPK (Oct 13) | Accounts confirmed for unauthorised automated play |
| 893 | QQPK (Oct 15) | Bots identified and banned after reviewing 2,913 accounts |
| ~12,000 | Vojtech Cervinka | Suspected nicknames, estimated from his own data (11,469 nicknames) |
These are not necessarily contradictory. A community report flags suspicious accounts using patterns; an operator needs stronger evidence before banning, because a wrongly banned legitimate player can result in complaints or legal issues. An independent analyst working from a large hand sample may identify a broader network than the operator can prove. PokerListings itself noted that the article does not reconcile the counts, and we cannot either.
What we can say is that the direction of travel is consistent: every figure is much larger than zero, the operator has acknowledged a syndicate, and the investigation remains contested.
Selection Bias: QQPK's Main Criticism
QQPK's report criticised the original data for selection bias and sampling limits. This is a legitimate technical point. If researchers collect hand histories from specific tables, stakes or times of day, their sample may over-represent the places where bots operate. Extrapolating from it to the whole site is risky. Conversely, an operator reviewing its own full database has access to more information, such as IP addresses, device fingerprints and timing data, that outsiders cannot see.
That said, an operator also has a conflict of interest. Admitting a bot problem damages trust, and an interest in minimising the headline number is understandable. This is why independent analysis, such as Cervinka's, is valuable. Neither side's figures should be accepted without scrutiny. Our guide to how poker sites catch bots explains the types of evidence involved.
The Compensation Rules
QQPK's $300,000 compensation fund has defined eligibility. To qualify, a player must have:
- Played tournaments in the past three months.
- Shared a table with two or more violating accounts in the same hand.
- Lost money in those tournaments, with buy-ins exceeding prizes.
Payouts in USD were scheduled for October 20, 2026.
Consider what this means. The requirement to share a table with at least two violating accounts in the same hand is demanding; many affected players may only have faced one bot at a time. The requirement to have lost money overall excludes winners, even if bots reduced their winnings. And the three-month window may leave out earlier victims. On the other hand, a $300,000 fund is a rare example of an operator offering any compensation. In the history of online poker, many bot busts ended with accounts closed and no refunds at all.
If the fund covers a limited number of players, the per-player payout depends on how many qualify. QQPK has not, in the material we reviewed, published a per-player figure. If you believe you qualify, check your account messages and the operator's announcement.
Why This Matters Beyond QQPK
Bots are not a niche problem. We have documented crackdowns at WPT Global, which banned 575 tournament bots, and studies of real-time assistance and bot enforcement across the industry. Every major site battles automation, and the arms race favours attackers who can use cheap computing power to run many accounts.
For players, the lesson is that no site is bot-free. Asia-focused rooms may face particular pressure because of lower operating costs and a large market, but the risk exists everywhere. What differs is how seriously a site invests in detection, transparency and response.
How to Protect Yourself
Practical steps for online players:
- Choose licensed or reputable sites. Our safe poker sites guide ranks operators by security, payout history and transparency.
- Watch for odd patterns. Opponents who 3-bet with identical sizes, never tank and act with perfect regularity may be automated. Report suspicious accounts through the site's support channel.
- Avoid sites that do not communicate. An operator that publishes a bot report, even a contested one, is more trustworthy than one that stays silent.
- Keep records. Hand histories and screenshots help if you ever need to claim compensation.
- Diversify. Do not keep your entire bankroll in one room. See our bankroll management page and payment methods guide for safe ways to move funds.
Players on crypto-focused rooms such as BC Poker should note that provably fair shuffles address deck integrity but not player-side automation; bots remain a separate concern. Our crypto poker guide explains the difference.
What Comes Next
The unresolved question is how many bots there really were and how much money they won. If Cervinka is right that roughly 70 percent of the hands he studied were played by bots, the damage would be larger than the compensation fund suggests. If QQPK is right that its review found 893 bots, the issue has been dealt with. The truth may lie between the two.
Look for three developments: additional detail on the compensation payouts, any reply to Cervinka's methodology, and whether other sites adopt similar disclosure practices. Transparent reporting, even when it exposes uncomfortable facts, is the only way to rebuild trust.
The Takeaway
The QQPK affair shows how difficult it is to measure a hidden problem. Four numbers, four methods and no agreed answer. For players, the practical conclusion is to treat bot claims seriously, favour sites that disclose and compensate, and maintain a healthy scepticism toward any single figure, whether it comes from an operator or an independent analyst.