How much do we still not know?

[Vman96]

It was the most convincing 'evidence' that I have come across on this subject.

Thanks Billyjoe. Hope you did well on the trip slash have some nice pics for us.

But since he said nothing about AC, we still get to hear about DaBurglar's "dud rates"!

[williejoe]

shadowman----i did go back and read the latest information regarding nevada regs-----and they do seem to indicate that any "manipulation" of how the random card selection by the prng is displayed is not allowed. so, i have to agree with your position on this issue.

in regards to the issue of seeding ( or possibly switching prng algorithms ), the language ( "rng must be seeded using a sufficiently unpredictable value ) leaves a good bit of leeway to potentially alter shorter term outcomes ( hours of play ) while maintaining true long term randomness ( days or weeks of play ) . it is noted that the randomess test that the regs require for the prng examines an extremly small portion of the full random number string ). i personally believe something in this area could be going on----but i will probably never know. and if it is----i console myself by still beliving that in the long,long run ( years of play ), vp outcomes are indeed random----and our best bet is to know the correct stategy for the game you play and choose the best paytables.

[sam434343]

williejoe,its good to have outsider like yourself, who has some knowledge of chips plus checking the laws in Nevada and not baseing it all on emotions(meaning just because we loss)that machines are not cheating but you feel they are random(chips) but you cant put your finger onit that the could computer chip could cheat 5% of the time and be fair the other 95% of the time that this might be a possibility? You did not put percentages these are just my opion 5% and 95%(lol) Thax,williejoe, Sam

[williejoe]

sam434343-----if the casino could or do "manage" the seeding of the prng algorithm----i think it would be impossible to put any percentages on what the exact impact might be. while the nevada regs require that a chi-square test challange be done on the vp machine's chip on ten thousand outcomes, and that it must meet a 95% confidence level for goodness of fit-----this doesn't mean that the remaining 5% is an opportunity for the casino to cheat. it is just a fairly standard way to show that the small sample of the outcomes tested( in this case, ten thousand ) are representative of the full non-repeating number steam which is a hugh number of outcomes, about 2 to the 2000 power. if the chip meets this test requirement----then it demonstates that with 95% confident that any variance seen between the small ten thousand sample and the full number stream is due to nothing more than normal random variance inherent in the full non-repeating number stream.
i do have some questions regarding the seeding process of the machines and what impact this could have----i've run out of time for now----but will post on this issue in a day or so. maybe someone on this forum will have some enlightening answers.

[sam434343]

williejoe,My choice of word (cheating this word always get attention) )might be way overboard, but I am interested in what you are suspicious of when it come to the computer chip and what suspicions you might think are going on with chips? I understand if you don't want say on this forum or your not exactly sure what going on but you think there might be something which makes sense to me.(lol) I am confused on what I just said! (LOL) I will look to see what more you have to say(with interest)Thx. Sam

[faygo]

in regards to the issue of seeding ( or possibly switching prng algorithms ), the language ( "rng must be seeded using a sufficiently unpredictable value ) leaves a good bit of leeway to potentially alter shorter term outcomes ( hours of play ) while maintaining true long term randomness ( days or weeks of play ) .


 
If altering of short term play is done , why wouldn't short term testing find it? How would the machine know if we were in the short term or the long term and at what point would it start to decide to catch up so it met the expected outcome?

[williejoe]

faygo, the chi-square test of the 10,000 results is a test for randomess of the prng-----it is not an evaluation of that segment as it relates to what cards that segment's output represents. any given segment of the full number stream can represent "hot" or "cold" areas in regards to what hands you get. certainly the entire number steam ( 2 to the 2000 power ) is random and the much smaller test segment ( 10,000 outcomes ) is tested as random------but if the seeding is "managed" by the casino as to preferentialy play through more of these "colder" segments of the number steam ( by reseeding ) then everything is fine as it relates to randomness----but there certainly would be an issue of fairness. the full prng's number steam, while very, very long---is known and documented---and is fixed as it relates to that particular algorithm. but where we play within that number stream is determined by the seeding.   and i will be the first to say that if the machines are seeded ( and/or reseeded ) in a totally random manner then it is difficult for me to see how the machinces could be anything but fair and random----and no one should have an issue other than complain about the poor paytables.

i'm hoping someone can post here that has more information as to how these machines are seeded. the verbage in the nevada regs seem to leave a lot to the imagination.

[Vman96]

faygo, the chi-square test of the 10,000 results is a test for randomess of the prng-----it is not an evaluation of that segment as it relates to what cards that segment's output represents. any given segment of the full number stream can represent "hot" or "cold" areas in regards to what hands you get. certainly the entire number steam ( 2 to the 2000 power ) is random and the much smaller test segment ( 10,000 outcomes ) is tested as random------but if the seeding is "managed" by the casino as to preferentialy play through more of these "colder" segments of the number steam ( by reseeding ) then everything is fine as it relates to randomness----but there certainly would be an issue of fairness. the full prng's number steam, while very, very long---is known and documented---and is fixed as it relates to that particular algorithm. but where we play within that number stream is determined by the seeding.   and i will be the first to say that if the machines are seeded ( and/or reseeded ) in a totally random manner then it is difficult for me to see how the machinces could be anything but fair and random----and no one should have an issue other than complain about the poor paytables.

i'm hoping someone can post here that has more information as to how these machines are seeded. the verbage in the nevada regs seem to leave a lot to the imagination.

If machines adhere to GLI 11, which I am 99.99% certain that IGT units do at least, then the casino can't "manage" the seeding like you say. However, I don't think it is required by Nevada to adhere to GLI 11, but I am fairly confident that a vast majority of Class III gaming machines nationwide adhere to these technical standards. See section 3.3 for RNG requirements.

http://www.gaminglabs.com/downloads/GLI ... 20v2.1.pdf

[shadowman]

I doubt we will ever know when the RNG is reseeded. However, there is a noticeable pause in IGT machines after every 100 hands when they log results. It would be a perfect time to also reseeded the RNG.

[Vman96]

I doubt we will ever know when the RNG is reseeded. However, there is a noticeable pause in IGT machines after every 100 hands when they log results. It would be a perfect time to also reseeded the RNG.

I would assume they do. It's definitely suggested by GLI 11 to periodically do it.