Is Online Poker Rigged? The Actual Numbers
Almost certainly not — but that is not what this page is for. Every site accused of rigging answers the same way: a certification badge and a line about a verified random number generator. If that worked, you would not be reading this. So here is the arithmetic instead, including the exact figures for the things people cite as proof, and instructions for testing any site yourself without taking anyone's word for it.
Our position, stated plainly. We are not a poker room. We do not deal real-money hands and we have no operator paying us to defend them. This page has no affiliate links on it and never will — a page that asks whether operators cheat, and then sells you one, has already answered the question about itself.
The reason it feels rigged is arithmetic, not suspicion
A live game deals about 30 hands an hour. One online table deals about 80. Four tables deal about 320 — roughly 11 times as many hands in the same evening.
Every rare, memorable, expensive event therefore happens about 11 times as often per hour of your life. Set over set, the flush that gets there on the river, the aces that lose — none of them became more likely. You simply moved to a format that deals you a decade of live poker in a year, and human memory does not divide by the denominator. Nobody remembers the four hundred hands they folded.
The specific claims, answered
These are the exact accusations that appear over and over in reviews and forum posts. Each is checkable, so here are the real numbers. The flop figures are exhaustive enumerations — every possible flop, counted — and the rest are large simulations. The script that produces them is in our repository.
Counted exactly, over all 1,081 possible turn-and-river combinations, for a player holding two of the suit when the flop brings two more. It feels like half because a missed flush draw is forgettable and a completed one costs somebody a stack.
Of all 22,100 possible flops, 55.1% contain exactly two cards of one suit, 5.2% are all one suit, and 17.2% are paired. Wet boards are not unusual — they are the normal case, and most players underestimate this badly.
When two hands see a flop and one is clearly ahead, the leader loses about one time in five. Across 397,288 simulated hands. In a session with a hundred flops, roughly twenty suckouts is the expectation, not an anomaly.
It is genuinely rare per hand. But four-tabling online deals 320 hands an hour, so a regular passes that many hands in well under a hundred hours. A one-in-twenty-thousand event is a certainty at volume — that is what volume means.
Across seven cards, most hands finish as one pair (43.84%) or two pair (23.48%). Everything from trips upwards is uncommon — but at a nine-handed table, someone holding one of them is not.
What your seven cards actually make
Best five-card hand from seven, over 400,000 deals. Bars are scaled for readability.
How to test it yourself
This is the part that matters, because it does not require trusting us either. If you suspect a site, the evidence is already on your hard drive.
- Decide first what would change your mind. Write down, before you look, the result that would convince you the site is fair. An investigation that can only confirm the thing you already believe is not one.
- Get a real sample. Tens of thousands of hands, not hundreds. Every site lets you download your hand histories.
- Count starting hands. Any specific pocket pair should arrive about once every 221 hands, and any specific suited combination about once every 331. If a site is dealing you a hand far more often than that over a large sample, that is a finding worth having.
- Check your all-ins against their equity.Over a big sample, the pots you got in as a 70% favourite should be won about 70% of the time. This is the test that most convincingly separates “I run bad” from “the deal is wrong.”
Our leak report and hand replayer both read GGPoker and Natural8 exports directly, and the variance simulator will show you what a normal run of bad luck looks like for your win rate — which is usually far worse than people expect.
The one claim we can actually prove
We will not tell you any operator is honest, because we cannot see their servers and neither can you. What we can demonstrate is our own practice table: turn off your internet connection and it keeps dealing. Every card, every opponent decision and every number the coach shows is computed on your own device, so there is no server deciding anything, and nothing that could be tilted against you. Every hand is saved to a hand history with all cards shown at showdown, so you can export a session and audit the deal.
That is a smaller claim than “we are certified.” It has the advantage of being checkable in about ten seconds.
Common questions
Is online poker rigged?
There is no credible evidence that the major regulated sites rig deals, and strong practical reasons they would not: the rake is a guaranteed profit on every pot, and the downside of being caught is the destruction of the business. But "trust us, we are certified" is not an argument, and it is worth knowing that the feeling has a real cause. Online poker deals roughly 11 times more hands per hour than a live game once you are multi-tabling, so you see roughly 11 times as many coolers and bad beats in the same evening. The events are not more likely; there are simply far more of them.
Why do I see so many flushes and straights online?
Because you see far more hands, and because the frequencies are higher than most players assume. Across seven cards, a flush appears 3% of the time and a straight 4.59% — and 55.1% of all flops already contain two cards of one suit. If you hold two of that suit when the flop brings two more, the flush completes 35% of the time, which is exact rather than estimated: it is every one of the 1081 possible turn-and-river combinations counted.
How often does the best hand on the flop actually lose?
About 19.9% of the time when two hands see a flop and one is clearly ahead. That is roughly one hand in five — so in a session with a hundred flops, being "sucked out on" twenty times is the expected result, not evidence of anything. The number comes from 397,288 simulated hands.
How often does set over set happen?
About once in every 22,727 hands where two players are dealt pocket pairs and both flop a set. It feels impossible because it is memorable and expensive — but at 320 hands an hour across four tables, a regular online player passes that many hands in under a hundred hours of play. Rare events are guaranteed to happen at volume.
How can I test whether a site is dealing fairly?
Download your own hand histories and count. Record how often you are dealt each starting hand over a large sample and compare it to the expected 1 in 221 for any specific pocket pair; check how often your all-ins hold up against the equity you had. Do it over tens of thousands of hands, not hundreds. The important discipline is deciding what result would change your mind BEFORE you look — an investigation that can only confirm what you already suspect is not an investigation.
Are play-money games rigged to make you buy chips?
This is the most common and most specific accusation in app-store reviews, and unlike the accusation against regulated real-money sites it is at least structurally plausible: a game that sells chips profits when you run out. The honest position is that you cannot verify it from the outside on a closed-source app. What you can do is prefer software that computes hands on your own device, where the deal is not coming from a server that has a financial interest in the outcome.
Count the real outs in the hand that made you suspicious.
Method: flop and flush figures are exhaustive enumerations of every possible case. Suckout, showdown and cooler figures are simulations with a fixed seed (397,288 to 500,000 hands each). The generator is scripts/generate-rigged-data.mjs in our repository, and regenerating it updates this page.