There's biases in gambling, such as the 0 on a roulette table, but that is arguably obvious to the player.

Is it legal for online casinos to knowingly or unknowingly bias their games at a software level (random number generator etc) in their favour? Could something like this be covered in the small print?

A problem with RNGs is they all have a bias, we have no perfect system of generating random numbers, but if a bias is detected by the casino, would they legally be required to adjust or change the system they're using?

Specifically I'm interested in the UK and anywhere notable in the US, or if there's somewhere in particular gambling websites set themselves up to avoid these regulations. It would also be interesting to see if the more popular gambling websites worldwide have any common locations.

This seems to cover randomness in the UK, does it apply to card games? (under section Randomness)


  • 1
    Not really relevant to law, but there are cryptographic protocols to generate "provably fair" random numbers. They aren't guaranteed to be perfectly random, any more than any other method, but they give either player the ability to defeat biases that the other might try to intentionally introduce. See en.wikipedia.org/wiki/Commitment_scheme#Coin_flipping for a simple example. Commented Jun 17, 2017 at 15:40
  • @Nate Eldredge A commitment scheme doesn't produceprovably fair results. In he coinflip eample Bob could use a system that gives 60% heads, say and the system would not detect it. What it does isd prevent bias introduced in the reporting of random events or predictions of random events. Commented Sep 7, 2022 at 16:30

1 Answer 1


In at least one jurisdiction for at least one game type, the answer is "Yes".

Modern slot machines and internet slot games are software, and yes, payout ratios are set as a parameter in the software, as an explicit bias in the software. There are statutory minimum payouts in some jurisdictions. The exact parameterized payout ratio for any specific game does not need to be made clear to a player. The jurisdiction may test, approve, and gather and publish aggregate statistics for all casino games.

As I am most familiar with New Jersey casinos: as a recent innovation in the law, New Jersey casinos can operate internet games. They are regulated by the Casino Control Commission of New Jersey. The Commission approves all games, and gathers and publishes statistics about the aggregate payout ratio which is shared by third party sites and becomes cited in advertisements (so there is a market based component to keep casinos from setting the parameters too far out of the pack of their competitors: NJ's statutory minimum payout ratio is 83% but you can see at the second link that all NJ casinos hover around 90-91% for the 2014 year of data).



If there is a new game or new software, before it is placed in service, it must be tested by the Commission; and the code employed by the casino must be a "true copy" of the code as tested. I don't know for certain if "pseudorandom" artifacts are a sufficient reason for failure in game testing for the Commission, but maybe you can find more to your specific case here:


  • 1
    Is this the same for games that have implicit odds based on the real world, like poker for example? If the odds of drawing an ace of spades in a single deck game is 1/52, do they have to either try and match the implied odds or I'm some way disclaim that they are not exactly the same?
    – ProfDrum
    Commented Aug 13, 2015 at 6:30

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .