Argument with buddy about RNGs

[pkrNinja]

My friend has recently seen a show on travel channel with some schmuck who discovered that the RNGs in slots were not really random and actually acted according to a pattern and exploiting them. This guy I suppose claimed that it is impossible to for a computer to really pick a random number.
So from this my friend decides that it is impossible for a poker site to really deal random cards. We have been arguing about it but I really dont have sufficient learning about the subject to offer a case.

All of my hands in PT align with the law of averages so I am not concerned about the casinos being shady, but I just want to know whether the RNGs do make truely random numbers.

Any input would be appreciated

[Innocentius]

Yes, it is impossible for a computer to generate truly random numbers on its own (just using software, no extra hardware). This is why software RNGs are often called pseudo-random number generators. They can, however, come very close to being truly random.

The poker sites, as well as others who depend on having truly random RNGs, use computers together with outside sources of entropy. One method is to have a geiger counter measure the rate of decay of a radioactive substance, and use the time between decays of atoms as input for the RNG. In this way, truly random numbers can be generated. There are companies who specialize in this, and sell "randomness" to others.

[pkrNinja]

Thanks for the reply. I tried to look it up on my own online and found some stuff similar to what you were saying, but I dont understand the language since I know nothing about RNGs. But thanks for the input. Do all poker sites (or the large ones anyways) generate random numbers this way?

[Innocentius]

Yes, I think all the big sites do it this way, either with radioactivity or with thermal noice. If you are interested in how it works, many sites (at least Party and Stars) have info on their web pages about how their RNGs work.

[Sciolist]

I'm lazy, so copy paste:

[ QUOTE ]
We take input for our shuffle from two sources. The first is a thermal noise generator developed by Intel. This measures minute changes in the background temperature, and is theoretically impossible to predict. The second is from mouse movements of all the players on the tables. This would be possible to predict if you firstly knew the algorithm that we use there, and secondly if you knew exactly where the mouse was for every single of the thousands (or tends of thousands) of players on the site. In short, the shuffle is impossible to predict in any way. We have had it audited by two companies, the results of which you can find here - pokerstars.com/rng_audit.html


[/ QUOTE ]

Oh, and pokerstars.com/security.html

*

Edit/MH: Note: Sciolist works for PS.

[BruinEric]

Always "play offense" as well in an argument...so make sure bring up a couple bonus points:

-Large commercial deck shuffling machines used in LIVE PLAY have some sort of Random Number Generator as part of the electronics. Does he believe this to be a problem?

-Ask him to explain how a deck shuffled with a cut and five riffle shuffles is relevantly better for player than an entropy-generated RNG seed with a robust (appropriate # of bits) shuffling algorithm.

[LoveDub]

[ QUOTE ]
Always "play offense" as well in an argument...so make sure bring up a couple bonus points:

-Large commercial deck shuffling machines used in LIVE PLAY have some sort of Random Number Generator as part of the electronics. Does he believe this to be a problem?

-Ask him to explain how a deck shuffled with a cut and five riffle shuffles is relevantly better for player than an entropy-generated RNG seed with a robust (appropriate # of bits) shuffling algorithm.

[/ QUOTE ]

And I'm certain that RNG in mechanical card shufflers use a RNG algorithm without external entropy sources (perhaps the timestamp only).