Page 1 of 2 12 LastLast
Results 1 to 20 of 21

Thread: The Video Poker RNG question... again

  1. #1
    In another thread the question of numbers generated by an RNG on a video poker machine came up... again.

    Since I am considering an investment in a glue factory (just joking here) I thought we would beat this horse to death just one more time.

    I asked again (yes for the billionth time) if we knew how many "cards" whiz by when we are thinking about pushing the draw button when faced with four to the royal:

    Originally Posted by Count Room View Post
    So if you sit there for a minute taking a deep breath deciding when to press the button for the magic 5th royal card for either a $20,000 jackpot or a frown, it's quite likely 60 billion numbers passed through the machine during that time where roughly 1.2 billion of those numbers will give you a great big smile and a sense of oneness with the casino universe.
    Actually the real question is are all of those 60 million numbers generated "equally" or do some numbers (cards) show up more... or less?

  2. #2
    The number of numbers is likely quite a bit less than mentioned. The VP machines processors are quite slow. And, the distribution of those numbers should be exactly equal over the time to complete a full cycle as the RNG is really a deterministic algorithm.

  3. #3
    thanks Arc, but "should be exactly equal over the time to complete a full cycle" is definitely different from an RNG that goes through all remaining 47 cards one by one with no repeats until all 47 are "seen."

    So in the case of four to the royal: if you were able to pin point the 47th card that goes through the RNG would you have all 47 cards "viewed" by the RNG or might one card be "viewed" more than once?

  4. #4
    Alan:

    Note that I said 60 billion, not million numbers over that one-minute period.

    Arci is saying the processors run much slower than that and I have spoken to a couple of slot techs telling me it can be as fast as nanoseconds (I tend to believe the latter).

    Honestly, none of us really know for sure because it's proprietary information with IGT. I don't even know if prying apart a reasonably recent model of a store-bought IGT machine would provide any further insights except...

    One COULD study the CPU itself and examine the transistor layout, maybe? That might give solid clues as to how fast the numbers might be processed...

    It shouldn't take that many programming instructions to create a random number. I can say with some certainty that a clock is actually used and the time is part of the algorithm. Can the time be broken down to atomic units (billionths of a second)?

    Fascinating questions. I wish I knew more.

    Have you considered interviewing an IGT spokesman about this? You definitely have the credentials and I would definitely keep my eyes peeled for such an interview!

    As for how many times your royal card will show up "behind the scenes" while you wait to press the 4-to-the-royal DRAW button? I guarantee it will show up many, many times if it's a continuous shuffling machine even if the RNG only creaks along at a mere 100,000 instructions per second.

    I also want to point out that I am a little skeptical of claims that modern machines have anything but continuous shufflers. It doesn't require much more software coding to create a continuous shuffle and it's more of a security risk for the casinos to have the RNG actually pause itself on the draw.

    Compared to fantastic computer games like "Elder Scrolls V: Skyrim", making software code for VP machines is probably farcically simple. (Yet there is a lot more money involved with the VP....)

    EDIT: For the programming geeks out there I should have said that "time" would be used as an argument in the RNG function.
    Last edited by Count Room; 01-12-2014 at 09:57 PM. Reason: Note for programming geeks

  5. #5
    Compare how much software programming went into this VP game:



    vs this computer game:



    Kind of a sad commentary on society, really. Those game developers with all the brilliant coding sophistication can only hope to make a few million dollars while such simple VP programs will witness the transactions of tens of billions of dollars.

  6. #6
    Originally Posted by Alan Mendelson View Post
    thanks Arc, but "should be exactly equal over the time to complete a full cycle" is definitely different from an RNG that goes through all remaining 47 cards one by one with no repeats until all 47 are "seen."

    So in the case of four to the royal: if you were able to pin point the 47th card that goes through the RNG would you have all 47 cards "viewed" by the RNG or might one card be "viewed" more than once?
    One card would be viewed more often as the conversion from the billions of RNG numbers to one of the 52 cards will lead to more of some cards over a short time span. However, the card seen more often could just as well be the one you need as any other.
    Last edited by arcimede$; 01-13-2014 at 07:04 AM.

  7. #7
    Count Room, years ago I downloaded a RNG off the internet and timed it on my PC. What I found was around 200K numbers over a single second on a 1.6 ghz INTEL processor. However, in looking at the processors in IGT VP machines I found they were either 12 khz or 16 khz. In other words, about a 100 times slower. This would lead to around 2,000 numbers per second or 120K over a minute of time. More than enough but not as high as I expected when I looked into it.

    I suspect the processors used today might be a little faster but I found nothing last year when I looked again.

  8. #8
    Originally Posted by arcimede$ View Post
    This would lead to around 2,000 numbers per second
    Ahhh... a manageable number. LOL

  9. #9
    Originally Posted by arcimede$ View Post
    Count Room, years ago I downloaded a RNG off the internet and timed it on my PC. What I found was around 200K numbers over a single second on a 1.6 ghz INTEL processor. However, in looking at the processors in IGT VP machines I found they were either 12 khz or 16 khz. In other words, about a 100 times slower. This would lead to around 2,000 numbers per second or 120K over a minute of time. More than enough but not as high as I expected when I looked into it.

    I suspect the processors used today might be a little faster but I found nothing last year when I looked again.
    Arci, I am going to say something to you that people rarely say to you on Alan's forum:

    "Your post could very well be correct and I will defer to your experience on this matter because I really don't know anything beyond what I've mentioned."

    How's that for a breath of fresh air?

  10. #10
    Count Room--You have been banned for seven days!!!!

  11. #11
    Originally Posted by regnis View Post
    Count Room--You have been banned for seven days!!!!
    A good deed never goes unpunished...

  12. #12
    Originally Posted by Count Room View Post
    Arci, I am going to say something to you that people rarely say to you on Alan's forum:

    "Your post could very well be correct and I will defer to your experience on this matter because I really don't know anything beyond what I've mentioned."

    How's that for a breath of fresh air?
    Not that big a deal. He wasn't ready to allow someone else to show more knowledge about any of the minor details about vp and get away with it, so all you did was diffuse the guy early. We've seen regnis frustrate the living daylights out of him several times. But you did the right and merciful thing, considering his circumstances.

  13. #13
    Rob, you surprise me. I thought you were going to tell us that when you had that VP machine for inspection and investigation that you were able to determine exactly how many "cards" went by per second. Perhaps you overlooked that during your investigation of fifth card flipovers and whatever else you claimed to prove?

  14. #14
    Originally Posted by Alan Mendelson View Post
    Rob, you surprise me. I thought you were going to tell us that when you had that VP machine for inspection and investigation that you were able to determine exactly how many "cards" went by per second. Perhaps you overlooked that during your investigation of fifth card flipovers and whatever else you claimed to prove?
    Do the math Alan....and do something you are continuously not ABLE to do: read with comprehension from your own forum. I had the machine for nearly 3 months. There was approx. 1 month of time to set up and program the test equipment suite, run some initial tests, then get it operating within the precise parameters I programmed it to, as well as intermittently stop it when it needed attention. There was a little over 2 billion hands simulated and analyzed. Figure it out if you're really that interested. I wasn't and am not.

  15. #15
    Like I said Rob... just surprised.

  16. #16
    Originally Posted by Alan Mendelson View Post
    Like I said Rob... just surprised.
    And what is it that makes you so "surprised" that I didn't use the machine to try and discover how fast the shuffle is? Who cares about that...and what difference would it make? If it were found to be as slow as molasses, would it keep you home more instead of running out to the casinos? If the machine were found to have shuffled at the speed of light, would you instead have gone on some family vacations?

  17. #17
    I'm not saying it does make a difference, but if you were recording data from a machine and checking its RNG I would think that one of the measurements you would take is how fast it works. I would think that would be one of the pieces of basic data you would want to have.

    For example, Rob, what if you tested the machine and found that it didn't go through many "cards" at all during the deal or draw processes? Then you might rightly say the machines are rigged and not fair.

    I would think that taking certain basic measurements are required when testing a machine for integrity and fairness. You didn't do that, it seems. So how can we believe any other results?

    Apparently you never even established that all 52 cards are even loaded in the game? Wouldn't that have been a hoot if you found no Jack of spades??

  18. #18
    Originally Posted by Alan Mendelson View Post
    I'm not saying it does make a difference, but if you were recording data from a machine and checking its RNG I would think that one of the measurements you would take is how fast it works. I would think that would be one of the pieces of basic data you would want to have.

    For example, Rob, what if you tested the machine and found that it didn't go through many "cards" at all during the deal or draw processes? Then you might rightly say the machines are rigged and not fair.

    I would think that taking certain basic measurements are required when testing a machine for integrity and fairness. You didn't do that, it seems. So how can we believe any other results?

    Apparently you never even established that all 52 cards are even loaded in the game? Wouldn't that have been a hoot if you found no Jack of spades??
    I even once read somewhere a long while back that VP machines can introduce a variable shuffling speed. Sometimes speeding up and other times slow. No, I have no proof or verification of this, but I can imagine several ways variable speeds can be coded into the game software.

  19. #19
    Originally Posted by Count Room View Post
    I even once read somewhere a long while back that VP machines can introduce a variable shuffling speed. Sometimes speeding up and other times slow. No, I have no proof or verification of this, but I can imagine several ways variable speeds can be coded into the game software.
    And that's why I didn't waste time looking for something that was irrelevant. There were no spec sheets that came with this machine, and what those specs were did not interest me anyway. What Alan says: "what if you tested the machine and found that it didn't go through many "cards" at all during the deal or draw processes? Then you might rightly say the machines are rigged and not fair" is meaningless, and no one would know if the machine was made that way in spec or not." So why would anyone think and how could anyone tell if the machine was fair or not if it did this? Does Alan, without doubt, know what it's supposed to be like?

    This is just as nonsensical: "I would think that taking certain basic measurements are required when testing a machine for integrity and fairness. You didn't do that, it seems. So how can we believe any other results?" I tested for a single parameter and I barely had time to do that. If he's ever set up and operated a test eq. suite he'd have a better appreciation for that. Knowing how fast a machine shuffles is not what he calls a "basic measurement" nor does it identify "fairness or integrity" when the manufacturer's specs are unknown. Then he comes up with "Apparently you never even established that all 52 cards are even loaded in the game? Wouldn't that have been a hoot if you found no Jack of spades??" How stupid a thought. I had a machine that was identified as being approved for installation on the floor at Aliante Station, and he questions if I picked one out that didn't have a full deck.
    Last edited by Rob.Singer; 01-15-2014 at 03:40 PM.

  20. #20
    It's always humorous watching Robbie try to cover up his previous lies with new lies. He gets soooooo irritated. LOL. Gosh, speaking of those tests, I wonder where all that documentation is he promised loooooong ago.

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •