How Online Poker Game Integrity Systems Work

What Poker Integrity Systems Do, in Plain Language

Online poker integrity systems are the combination of software rules, record keeping and human review that an online poker operation uses to keep the game honest. They have two jobs. The first is to make some kinds of unfair play impossible by design, for example by making sure that one player cannot sit in two seats at the same table. The second is to notice unfair play that cannot be blocked outright, such as players secretly working together or an automated program playing instead of a person.

This matters because poker is different from a house-banked casino game. Players compete against each other, so an unfair advantage for one player is a loss for the others, and the quality of the game depends on trust between players. The technology that protects that trust is mostly invisible, which is why many players wonder how it can possibly work.

This article explains the general ideas: what the main threats are, what kinds of information a system keeps, how an alert becomes an investigation, and what regulators’ technical standards say. It deliberately stays at the level of principles. Detailed detection methods are not described, because that detail could only help someone defeat them, and no poker room, network or software is named.

The Main Threats and the Three Layers of Defence

Integrity teams usually group threats into a few families. Collusion means two or more players sharing information or coordinating play against the rest of the table. Multi-accounting means one person controlling more than one account, or sitting in more than one seat. Bots are programs that play on a person’s behalf. Real-time assistance (RTA) means using outside software during a hand to get decisions.

Defences are easiest to understand as three layers. The first layer is prevention by design: the software simply does not allow certain things, such as a second seat for the same account, and it keeps each player’s hole cards visible only to that player. The second layer is detection: the system records what happens and looks for patterns that deserve a closer look. The third layer is response: reviewers examine the evidence, and the operator can suspend accounts or sessions and close accounts when cheating is confirmed.

None of these layers works alone. Prevention cannot stop what looks like normal play, detection produces alerts rather than verdicts, and response depends on good records. The deal itself, meaning how the server shuffles and distributes cards, is a separate subject covered in how online poker software deals a hand.

Under the Hood: Records, Alerts and Review

Regulators set out what an operator must be able to do. The UK Gambling Commission’s remote technical standard RTS 11 (last updated 17 July 2023) requires measures to deter, prevent and detect collusion and cheating, and requires the system to retain records and to be able to suspend or disable accounts or sessions. Its guidance gives examples of what detection may cover: players who frequently share the same tables, players from the same address at one table, suspicious play such as chip dumping, and unusual gameplay statistics. Chip dumping means deliberately losing chips to another player to move them between accounts.

The Alcohol and Gaming Commission of Ontario (AGCO) sets similar expectations in its peer-to-peer games standards. They call for measures to deter, prevent and detect unfair behaviour, collusion and cheating, including suspected bots, and they require that hand information not be accessible to others, that a mechanism ensures a player cannot play against themselves or occupy more than one seat, that records are retained, and that the effectiveness of the measures is monitored.

In practice this means a stream of stored data feeds the detection layer: who sat where, which hands were played, what actions were taken, how often the same players met, and whether players at a table share an address. Software compares that data against expected ranges and raises an alert when something falls outside them. The key point is that an alert is a lead. Statistics can show that something is unusual, but unusual is not the same as dishonest, so a person has to review the evidence before any action is taken.

Record keeping also has a time dimension. Michigan’s administrative rules on integrity monitoring and suspicious behavior require operators to employ integrity personnel, to notify the state board of suspected cheating or collusion, and to keep records for a minimum of 5 years. Saskatchewan’s internet gaming operating standard similarly calls for logging and detecting patterns that indicate collusion or unauthorized artificial player software.

Walkthrough: From an Unusual Pattern to a Decision

The following is a hypothetical, simplified flow at a generic online poker operation, here called Site A. Real operations differ, and the figures are invented for illustration.

1. Data is recorded. As hands are played, Site A’s servers store each hand’s actions, the seats and the accounts involved, along with account and session details.
2. A monitoring rule raises an alert. Suppose two accounts have met at many more tables than chance would suggest. The system creates an alert that links the two accounts and the hands involved.
3. An analyst reviews the case. A member of the integrity team reads the hand histories and account details. Many alerts end here, because the pattern has an ordinary explanation, such as two friends who like the same stakes.
4. More evidence is gathered if needed. The analyst may widen the review to further hands or related accounts.
5. A decision is made. If the evidence shows a rule was broken, the operator can suspend the accounts or sessions and, where cheating is found, close the accounts. Under RTS 11, players should be told that accounts will be closed if cheating is found.
6. The record is kept. The case file and the supporting data are retained so the decision can be reviewed later, for example if a player complains.

Variations: What Different Standards Ask For

Rules differ by jurisdiction, and the same technology is configured differently from place to place. The table summarises a few published requirements; it is a general comparison, not a legal guide, and the law itself belongs to a different subject.

The bot question shows how much the rules can differ. In some regimes operators must tell players that bots or software-assisted players may be present, and must say whether third-party software is permitted. Where certain software is prohibited, the operator is expected to put measures in place to deter, prevent and detect its use. The details of what counts as permitted therefore depend on the operator’s published rules and on local requirements, and a player should read those rules rather than assume.

Myths and Limits

Myth: the site can see a player’s cards and decides who wins. Integrity monitoring is separate from the deal. Standards such as AGCO’s require that hand information not be accessible to others, and the monitoring described here works on records of play rather than on influencing results.

Myth: every alert means someone cheated. Unusual patterns arise naturally in a large game population. That is why an alert leads to a review, not an automatic penalty.

Myth: integrity systems catch everything. No detection system is perfect. It can miss misconduct and it can flag innocent play, which is why standards ask operators to monitor how well their measures work and to investigate complaints about cheating.

A real limit: a player cannot see integrity systems at work, so the visible signs are limited to the operator’s published rules, its notices about bots and software, and the outcome of any complaint. This article describes general principles; how any particular operation implements them is not something it can verify.

Key Facts and What a Player Can Do

A player who suspects unfair play can take legitimate steps: note the table and the time, avoid confronting other players, and send a report through the operator’s support. Standards such as RTS 11 say complaints about cheating should be investigated, so a clear report with hand details is the most useful thing a player can provide.

Related Concepts and Next Step

To see how the game runs on the server before any monitoring begins, read how online poker software deals a hand. Tournament play adds its own software for registration, blinds and table balancing, explained in how online poker tournament software works. Fair shuffling depends on the generator underneath it, which is covered in how casino RNGs work, and the limits of software that claims to predict outcomes are explained in why AI cannot predict random gambling outcomes. Together these articles show the layers of a poker platform: the deal, the tournament structure, the randomness underneath and the monitoring that sits on top. Readers who want the broader platform picture, including how accounts and game servers connect, can continue with how online gambling platforms work.

Sources & References

• UK Gambling Commission — RTS 11: Limiting collusion and cheating (last updated 17 July 2023)
• UK Gambling Commission — RTS 16: Use of third party software
• Alcohol and Gaming Commission of Ontario — Registrar’s Standards for Gaming, Peer-to-Peer Games
• Michigan Administrative Code R 432.643 — Integrity monitoring and suspicious behavior
• Saskatchewan Liquor and Gaming Authority — Internet Gaming Operating Standard (September 2022)