the draw

[MikeA]

Several years ago, a VP writer whose name I shall not mention wrote that after receving a certain sequence on the draw he was curious about what would have happened if his hold decision had been different. So he called an attendant over, and they "backed up" the machine and let him see what could have been. Is this, in fact, possible?
 
I would doubt that would be possible.  Not if each card drawn was selected by an separate query to the RNG.  If the cards were "pre-selected" (i.e. the first two scenarios Shadowman identified) then it might be possible since the 5 cards to be potentially drawn were selected at the time of the deal.  However, I've not myself seen such an action by an attendant or anyone else.

[shadowman]

Several years ago, a VP writer whose name I shall not mention wrote that after receving a certain sequence on the draw he was curious about what would have happened if his hold decision had been different. So he called an attendant over, and they "backed up" the machine and let him see what could have been. Is this, in fact, possible?
 
For many years VP machines have kept a history of the last so many hands (deal and draw). I can't remember the exact number. Assuming these were older machines an attendent could go back and see the dealt hand and the final hand. From this they could ascertain whether a "hit" would have occurred.
 
I've seen the history used a few times when players complain that a "held" card was dropped by the machine. I have used this myself especially when it occurs on a dealt winner. Once I was seated next to a man who was dealt quad 6's played DB. When he hit draw one of the 6's was dropped. He immediately called over an attendent and they paid him for the quad. I had recently moved off the machine because it was dropping held cards so he had a witness ... which I suspect may have been the only reason they paid him. Some casinos are better than others.
 

[MikeA]

I can understand the aspect of being able to see what was dealt and what was held and what the results were from a history of hands even on current day machines (ie, gamemaster).  What I can't see is how you could play the "what if" game on these machines. 
 
Now, possibly on those older machines where the hand potential was determined by a "pre-draw" that happened at the time of the initial deal, you could maybe "look" at what each card was that was either "on the stack" or "behind" the initial cards.  With that information, you could certainly see what your results would have been had you changed your hold.

[faygo]

[QUOTE=rascal]
Several years ago, a VP writer whose name I shall not mention wrote that after receving a certain sequence on the draw he was curious about what would have happened if his hold decision had been different. So he called an attendant over, and they "backed up" the machine and let him see what could have been. Is this, in fact, possible?
 
For many years VP machines have kept a history of the last so many hands (deal and draw). I can't remember the exact number. Assuming these were older machines an attendent could go back and see the dealt hand and the final hand. From this they could ascertain whether a "hit" would have occurred.
 
I've seen the history used a few times when players complain that a "held" card was dropped by the machine. I have used this myself especially when it occurs on a dealt winner. Once I was seated next to a man who was dealt quad 6's played DB. When he hit draw one of the 6's was dropped. He immediately called over an attendent and they paid him for the quad. I had recently moved off the machine because it was dropping held cards so he had a witness ... which I suspect may have been the only reason they paid him. Some casinos are better than others.
 [/QUOTE]
 
I had a slot supervisor open a slant top for me to check on a dropped card.(Sticky Button) He showed me how he could go back 10 hands. He said this is used most often when someone claims they hit a good hand and wasn.t paid the proper amount.
 

[MikeA]

Excellent bit of information to know. 
 
I'll bet the Attendants have extremely embarrased some players who might have been trying to con the casino by utilizing this feature of the machines!

[New2vp]

The method you just mentioned is the one used by IGT years ago. It is very efficient but does not provide an even distribution of cards. Since the RNG usually works off a power of 2 and a 52 card deck is not an exact multiple you end up with more cards selected at the lower end of the table when the RNG number is divided by 52. Pretty insignificant but none the less a small difference.
 
I have no idea what IGT does now. Using the RNG to shuffle a virtual deck is a reasonable and fairly easy technique. It is used by Dan Paymar in his Optimum Play software tool. It has a more even distribution as well.
 
Most writers speak to a virtual deck to make it easier to understand. It's something we can all relate to.
  It is actually fairly easy and 99.99%+ efficient to use an RNG to get the probability of each card to be exactly 1/52 even with a binary or powers-of-2 system.
 
For example, many languages allow an RNG that has 2^32 = 4,294,967,296 values.  Though it is true that this number is not evenly divisible by 52, each card can be assigned exactly 82,595,524 of the RNG's values and the remaining 48 values can simply be ignored.  That is, if the RNG generates a number from 4,294,967,249 to 4,294,967,296, inclusively, just go on to the next pseudo-random number before selecting a card.  (4,294,967,296 divided by 52 results in 82,595,524 with a remainder of 48.)
 
I do not know if that is how IGT or others do it or not; but it could be done that way.  This is simply a mathematical trick that could be used in programming to get exact probabilities.  This is still pretty efficient because only 1 in about 90 million numbers would be skipped.

[New2vp]

Webman, Sorry about my inability to use the tools properly.  My 12:17 post was originally intended to be a quote from an earlier post in this thread by Shadowman.  I guess I don't know how to get the "quote boxes" that I see in many posts in the forum.   (I couldn't seem to figure out how to edit my post either).  My new words in reply to the original Shadowman post are in the 12:34 message.  If you can rearrange these three posts to get my intended effect (including eliminating this post), I would appreciate it.  Thanks. 

[Webman]

Webman, Sorry about my inability to use the tools properly.  My 12:17 post was originally intended to be a quote from an earlier post in this thread by Shadowman.  I guess I don't know how to get the "quote boxes" that I see in many posts in the forum.   (I couldn't seem to figure out how to edit my post either).  My new words in reply to the original Shadowman post are in the 12:34 message.  If you can rearrange these three posts to get my intended effect (including eliminating this post), I would appreciate it.  Thanks. 
You got it.  I should probably put together a tutorial on how to use the quote button properly.  You are not the only one who has had problems. In short, you hit the quote button and then leave the [ quote ] and [ /quote] tags alone and then type your comments after. If there are multiple quote tags, you always need a closing [ / quote ] for each opening [ quote ]
 
I suggest people use the "preview post" button the first few times when quoting to make sure their message appears correctly before posting.

[shadowman]

 It is actually fairly easy and 99.99%+ efficient to use an RNG to get the probability of each card to be exactly 1/52 even with a binary or powers-of-2 system.
 
For example, many languages allow an RNG that has 2^32 = 4,294,967,296 values.  Though it is true that this number is not evenly divisible by 52, each card can be assigned exactly 82,595,524 of the RNG's values and the remaining 48 values can simply be ignored.  That is, if the RNG generates a number from 4,294,967,249 to 4,294,967,296, inclusively, just go on to the next pseudo-random number before selecting a card.  (4,294,967,296 divided by 52 results in 82,595,524 with a remainder of 48.)
 
 
Right you are. Since a VP machine is an imbedded system, IGT can do pretty much as they like. I could see their RNG set up to automatically do the subtraction you mentioned based on the number of cards in the deck. In this way the different game programs would not need to care. The program could instantiate an RNG object with a specific deck size. From that point on all the program would do is ask for the next card.
 
However, they still might do a simulated shuffle (or simply reorder their simulated deck of cards every now and then) to help with card arrangements. Even though there are only 2.6 million unique hands, there are almost 312 million possible arrangements. That is quite close to the RNG width which is not the best possible situation. Of course, they could also build a 64 bit RNG and eliminate this concern easily.
 
I've estimated that current technology IGT machines could generate 100K hands/second. This means a 32 bit RNG would cycle every 12 hours. A 64 bit RNG would cycle every few quadrillion hours. Kind of boggles the imagination.

[babybubba]

"In this way the different game programs would not need to care."
 
Please 'instantiate' that statement. And just how much do programs care anyway?
 
What we're seeing is the same 'ol rattle & roll by shadow. He reviews posts where he has some understanding of but not entirely, he then attempts to interject his own interpretation in a way that is meant to hopefully impress the readers beyond what the original poster has accomplished, and he finalizes the dupe by adding in infrequently used words. He-Haw