Video Poker basics article

[doublediamonddlx]

Not too insightful....
http://www.mesquitelocalnews.com/viewnews.php?newsid=4320&id=2

Random Number Generators In Control Of Video Poker
While it might be a common conception that the game of Craps is the king of the table games, there is another game that is far and away the king of the entire casino.

Video poker enjoys such wide popularity that in most casinos the video poker machines outnumber all other games combined.

And on any given night most of these machines will be occupied with players hammering away at the keys, hoping and praying for the elusive jackpot.

It is a safe bet that most people are fully aware of the rules of video poker and various game playing strategies abound.

There are those who will be constantly “playing for the Royal� while others are happy just to try to build their pot slowly and methodically.

But what about the actual functionality of these machines and their varying pay tables? [[AD-3-Right]]

Ted Maupin, director of slots for the Eureka Hotel and Casino, offered some sage insight into how these machines tick and what can a player look for on those pay tables to increase their odds of winning.

“The pay tables are the first thing the savvy players should look at in order to find the machine with the best returns,� Maupin explained.

These pay tables are generally shown right above the playing area and they are very easy to read.

The first thing to look for is the payout for a Royal Flush.

More often than not these pay tables reveal that a Royal Flush hit with max coins bet will pay 4,000 credits, unless of course it is a machine tied to a progressive jackpot.

But there are some machines around that have a much lower payout for the Royal.

They are rare, but they can be found.

The advantage to playing these machines is that because of the lower payout on the Royal, the machines will actually pay the smaller jackpots more often.

And the Royal itself will actually hit slightly more often.

What about those smaller jackpots?

This is another part of the pay table that the savvy player will look for.

At every chance a player needs to be playing what is known as a "9/6" machine.

This refers to the nine-to-one payout on Full Houses and the six-to-one payout on Flushes.

Many machines sprinkled throughout the casino will have an 8/5 payout, eight-to-one on the Full House and five-to-one on the Flush.

While they are a lesser payout, that does not translate to a “looser� machine.

It simply means more profit for the casino and less return for the player.

These pay tables can vary from game to game on the same machine, so always be aware of the game that is dialed in and that game's particular payout.

But what makes the machine work, where do the cards come from and are they being manipulated in some way to beat the player?

The answer to that last question is yes and no.

Video Poker is a game of chance played in a casino and there simply are no games anywhere in that casino that do not give an edge to the house.

But are the cards being manipulated by management and can these machines simply be “tightened-up� on a whim?

According to Maupin the answer is an unqualified no.

“The machines are not tied together for any reason other than the player tracking system,� Maupin said.

He went on to explain a bit of the inner workings of a Video Poker Machine.

Hands on a video poker machine are dealt using a Random Number Generator, and while the software is programmed to give the house an edge, the player is given what amounts to “true odds� of hitting a winning hand.

In developing the software for these machines, companies like IGT and Bally’s arrived at a “par� for the payouts based on game simulation that took place over the course of 10 million hands.

With the knowledge of these hands imbedded in the software, programmers are now able to give as lifelike an experience as would be enjoyed in a real card room.

In other words, you have the same chance of hitting a Royal (or any other winning hand) as you would with an actual deck of cards.

These number generators are constantly shuffling cards while the game is in an idle mode.

Once the play button is hit the shuffling stops immediately and 10 cards are dealt.

The first five are shown while the second five are held in reserve.

At no point during that particular hand do those hidden five cards change.

Many players mistakenly believe that the draw cards are dealt based on the actions of the player.

This is simply not true and any evidence to that misconception is purely coincidental.

While the machine is in an idle mode, waiting for the play button to be hit, these number generators shuffle the cards at a rate of about 6,000 hands per minute.

That means that, in theory, with over 42,000 possible hand combinations, the Royal Flush could be dealt at any second every eight minutes, proving that hitting that indefinable Royal is nothing more than an extreme amount of “lucky timing.â€�  {EDIT note... so should we start slowing down our play...LOL}

So get yourself a players card, find a 9/6 jacks or better machine, preferably one that is tied to a progressive jackpot, and you will now have your best chance at winning with the king of the casino.

As always, good luck.
 

Enjoy,
DDD

[knagl]

there are some machines around that have a much lower payout for the Royal.

They are rare, but they can be found.

The advantage to playing these machines is that because of the lower payout on the Royal, the machines will actually pay the smaller jackpots more often.

And the Royal itself will actually hit slightly more often.

What the hell are they smoking?  That's totally inaccurate.

These number generators are constantly shuffling cards while the game is in an idle mode.

Once the play button is hit the shuffling stops immediately and 10 cards are dealt.

The first five are shown while the second five are held in reserve.

At no point during that particular hand do those hidden five cards change.

Also an inaccurate statement when it comes to modern-day machines.  When you press "deal" the first five cards are selected and displayed.  Then, the RNG continues churining out numbers and "shuffling" the remaining cards until the player presses "draw".

[CaptainHappy]

there are some machines around that have a much lower payout for the Royal.

They are rare, but they can be found.

The advantage to playing these machines is that because of the lower payout on the Royal, the machines will actually pay the smaller jackpots more often.

And the Royal itself will actually hit slightly more often.

What the hell are they smoking?  That's totally inaccurate.

  I am so happy that someone called BULLSH T to that comment!

These number generators are constantly shuffling cards while the game is in an idle mode.

Once the play button is hit the shuffling stops immediately and 10 cards are dealt.

The first five are shown while the second five are held in reserve.

At no point during that particular hand do those hidden five cards change.

Also an inaccurate statement when it comes to modern-day machines.  When you press "deal" the first five cards are selected and displayed.  Then, the RNG continues churining out numbers and "shuffling" the remaining cards until the player presses "draw".

Kevin,

I have heard both stories on the above two scenarios... Have you ever confirmed that the original was the case, and they changed it to the second scenario??? I am curious to finally put this to rest.

CH

[StatFreak]

Not too insightful....

I tend to agree.
Still, thanks for posting the article. 

there are some machines around that have a much lower payout for the Royal.

They are rare, but they can be found.

The advantage to playing these machines is that because of the lower payout on the Royal, the machines will actually pay the smaller jackpots more often.

And the Royal itself will actually hit slightly more often.

What the hell are they smoking?  That's totally inaccurate.

 

The odds of being dealt any hand do not change with the payouts because all class III video poker machines must use a standard deck of 52 cards (or 53 with a joker) and must give equal weight to each card in the RNG. Cards CANNOT be weighted like the physical stops on a stepper slot -- it's illegal.

However, the chances of getting a particular final hand ARE affected by the player's hold decisions. As an extreme example, if a player playing Jacks-or-Better always holds cards of the same suit and throws away all pairs and possible non-suited straights, he/she will get more flushes and straight flushes, but at the expense of pairs, two pairs, three-of-a-kinds, straights (slightly), full houses, and four-of-a-kinds. It's not the RNG that affects these outcomes, but player strategy.

That's why some players will do better than others when playing VP. Every paytable of every game has a "Best" strategy*. Even the same game with a different paytable will have a slightly different strategy. To continue with my above example, a Jacks-or-Better game that pays higher for a flush and less for a straight will have the player holding cards for a flush more often than one that doesn't pay as well for the flush. This is because any calculated strategy for a specific VP game instructs the player to hold and draw for the maximum EV (Expected Value), or greatest return, for each hand dealt.


*I don't use the term "perfect" strategy because different VP analysis programs will produce slightly different variations, and perfect strategy involves making decisions based, not only on the hand dealt, but on the specific cards being discarded and how their absence from the remaining deck affects the odds. That gets too complicated for almost anyone to use effectively in a casino, although some players have done so.

These number generators are constantly shuffling cards while the game is in an idle mode.

Once the play button is hit the shuffling stops immediately and 10 cards are dealt.

The first five are shown while the second five are held in reserve.

At no point during that particular hand do those hidden five cards change.

Also an inaccurate statement when it comes to modern-day machines.  When you press "deal" the first five cards are selected and displayed.  Then, the RNG continues churining out numbers and "shuffling" the remaining cards until the player presses "draw".

  I'm not sure in this case. Making certain would require analyzing the assembly code in the game chip. However, the choice of implementation shouldn't make a difference to the player in the long run. I wonder if there are any legal issues that would come into play here? (no pun intended -- okay, okay! I always  intend to pun. )

At every chance a player needs to be playing what is known as a "9/6" machine.

This refers to the nine-to-one payout on Full Houses and the six-to-one payout on Flushes.

Many machines sprinkled throughout the casino will have an 8/5 payout, eight-to-one on the Full House and five-to-one on the Flush.

While they are a lesser payout, that does not translate to a “looser” machine.   <-- HUH? Why would ANYONE think that a lower payout would translate to a looser machine? It DOES translate to a tighter machine! SF

It simply means more profit for the casino and less return for the player.  <-- Right. Someone should have edited the previous sentence...

These pay tables can vary from game to game on the same machine, so always be aware of the game that is dialed in and that game's particular payout.

The problem with these statements is that Jacks or Better is almost extinct on casino floors, and there are now so many video poker variations that telling someone to simply look for the "9/6" machines today is irresponsible.

[knagl]

I have heard both stories on the above two scenarios... Have you ever confirmed that the original was the case, and they changed it to the second scenario??? I am curious to finally put this to rest.

My information is from a 2001 interview published in the American Casino Guide.  I just looked in the 2010 guide and it's still in there.  With no intention to violate copyright, here's the meat of that section of the book, written by Steve Bourie (who is the main author of the book):

...what about the inside [of a video poker machine]?  Well, I had a few questions about that so I visited International Game Technology, which is the world's largest manufacturer of video poker machines (as well as slot machines), in January 2001 and spoke to their chief software engineer, James Vasquez.  Here's what Jim had to say in answer to some questions about how his company's machines work:

...

When the cards are dealt is it done on a serial basis where it's similar to cards coming off the top of a deck?  Or, parallel where there are five cards dealt face up and one card is unseen underneath each of the initial five cards?

Vasquez: It's serial and the five later cards aren't determined until there is more player interaction at the time of the draw.


They aren't determined at the time of the deal?

Vasquez: No.  They're determined at the time of the draw.  That varies with the jurisdictional regulation actually.  Some lottery jurisdictions tell you that you have to draw all 10 at once.  Different jurisdictions write into their rules how they want it done, specifically on poker, because it's a simpler game and they understand it.  They say they either want all 10 done at once, or however they want.


How is it done in Nevada?  All ten at once, or five and five?

IGT: In Nevada it's five and five.


The talk with Jim Vasquez confirmed that in most regulated jurisdictions video poker machines use a Random Number Generator to shuffle a 52-card deck and then choose five cards to display to the player.  ...  Then, when the deal button is pushed, the next group of cards is chosen and dealt to the player.

Source: American Casino Guide (...which I strongly suggest anyone visiting Las Vegas purchase a copy or two -- the coupons alone in the back of the book pay for the book many times over.)  He did go on to mention that IGT isn't the only company that makes video poker games, and that other companies may do their dealing of cards differently.

I'm not sure in this case. Making certain would require analyzing the assembly code in the game chip. However, the choice of implementation shouldn't make a difference to the player in the long run.

Agreed.  I always thought the PE+ dealt all 10 at once, but using the emulator I was able to "turn off" the machine mid-hand, save its state, "turn on" the machine and complete the hand and note the replacement cards.  I then "turned off" the machine again, moved a copy of the nvram file from the previous mid-hand state into the emulator's file area, and "turned on" the machine again.  As expected, the game resumed mid-hand again, but unexpected was that my replacement draw cards were different!  While that doesn't totally prove that the draw cards are determined later on a PE+, it does prove that a power cycle alters the draw cards received if a game is reset mid-hand.  Stolistic may be able to tell us for sure how the PE+ does it.  The 2001 interview I quoted above was likely talking about the Game King platform.

The problem with these statements is that Jacks or Better is almost extinct on casino floors, and there are now so many video poker variations that telling someone to simply look for the "9/6" machines today is irresponsible.

Agreed.  I've seen casinos have "9/6" machines with 1 for 1 (oh yeah, it's nit-picky, but he said 9 to 1 in the article -- it's 9 for 1) on the two-pair payout.  Hardly a full pay machine.  I've also lately seen casinos tinkering with other parts of the paytable, including one case where they eliminated the bonus for the Royal, making it 1250 coins for a 5-coin bet, rather than the expected 4000 coins.  That must have been one of those machines that hits the Royal more frequently and pays out smaller jackpots frequently, too.     Nowadays, players really need to look at every single pay on the paytable before playing to know what they're getting into.  Most just don't care, though.  To a point, that makes me happy as it increases the casino hold on the good games which allows the casinos to keep them on the floor so that a knowledgeable player (me) can play them and get a good return!

[StatFreak]

I'm not sure in this case. Making certain would require analyzing the assembly code in the game chip. However, the choice of implementation shouldn't make a difference to the player in the long run.

Agreed.  I always thought the PE+ dealt all 10 at once, but using the emulator I was able to "turn off" the machine mid-hand, save its state, "turn on" the machine and complete the hand and note the replacement cards.  I then "turned off" the machine again, moved a copy of the nvram file from the previous mid-hand state into the emulator's file area, and "turned on" the machine again.  As expected, the game resumed mid-hand again, but unexpected was that my replacement draw cards were different!  While that doesn't totally prove that the draw cards are determined later on a PE+, it does prove that a power cycle alters the draw cards received if a game is reset mid-hand.  Stolistic may be able to tell us for sure how the PE+ does it.  The 2001 interview I quoted above was likely talking about the Game King platform.

Interesting. That's the kind of test I would be likely to do, and it certainly points to the code determining the draw cards after the initial draw. If this is how it's done today (re: GK) I'd like to hear the designers' reasons for such an implementation. Not to be redundant, but selecting all 10 cards at the onset using a single PIA, or selecting the five draw cards at a later time with a second PIA shouldn't affect the long-term outcome in any way, unless there are issues with the RNG.

Stolistic, what's your take on this?

...
I've seen casinos have "9/6" machines with 1 for 1 (oh yeah, it's nit-picky, but he said 9 to 1 in the article -- it's 9 for 1) on the two-pair payout.
...

You're not being nit-picky. Ask any savvy craps player. There's a big difference between 5 FOR 1 odds and 5 TO 1 odds. In some cases, the casino uses that wording to hide their edge.

There are 6 ways to roll a seven out of 36 possible combinations with two dice, so the odds of rolling a seven are 1 in 6. If the casino paid 5 to 1 odds, the bet would pay back 100%. The "Seven" bet is generally listed as paying "5 FOR 1" on the layout to deliberately mislead the player into thinking that the payout is "fair". It looks better than printing "4 TO 1" on the layout, which is the same as "5 FOR 1".

At 5 for 1, the house edge is a nasty 16.67%, or in slot payback terms, the bet pays back 83.33%. They have a 16.67% edge over the player because they print the word "FOR" instead of the word "TO" on the layout using the same numbers. That's not picking nits, it's picking pockets!

The problem with these statements is that Jacks or Better is almost extinct on casino floors, and there are now so many video poker variations that telling someone to simply look for the "9/6" machines today is irresponsible.

Agreed.  I've seen casinos have "9/6" machines with 1 for 1 (oh yeah, it's nit-picky, but he said 9 to 1 in the article -- it's 9 for 1) on the two-pair payout.  Hardly a full pay machine.  I've also lately seen casinos tinkering with other parts of the paytable, including one case where they eliminated the bonus for the Royal, making it 1250 coins for a 5-coin bet, rather than the expected 4000 coins.  That must have been one of those machines that hits the Royal more frequently and pays out smaller jackpots frequently, too.     Nowadays, players really need to look at every single pay on the paytable before playing to know what they're getting into.
...

What it boils down to is that the informed VP player needs to know which games one wants to play, as well as the acceptable paytable for the games, before arriving at a casino. One should also bring along laminated cards for the specific paytable for each game unless one has practiced enough to memorize the correct plays. There are sites online that keep track of the best paytables to be found in specific casinos, which can be useful in cases where one is limited and cannot find the best paytable for the game. In such cases, one should practice and use the strategy suited for the paytable available at the casino of choice.

<ADD> It is best not to practice or play more than one or two different themes at any one time, and definitely not more than three. Doing so only gets confusing and one ends up making mistakes that could have been avoided. Deuces Wild games, in particular, have playing strategies that are radically different from typical games with no wild cards and trying to master a Deuces Wild theme and a non-wild theme simultaneously can end up costing the player money. My personal recommendation is for a player to pick one variety and stick with it.

I printed up laminated cards for Triple-Double Bonus Poker for CaptainHappy and myself when we went to the Aquarius in Laughlin. It turned out that their paytable was a bit worse that we had expected and we had to make due with slightly inaccurate cards, but at least we were better prepared than 99% of the players out there. I made up revised cards for the following year, but then found better paytables at the Edgewater.

...
Most just don't care, though. To a point, that makes me happy as it increases the casino hold on the good games which allows the casinos to keep them on the floor so that a knowledgeable player (me) can play them and get a good return!

You can say that again! Blackjack would have been removed from the casino floor years ago if it weren't for its popularity and all of the bad players out there. I thank the bad players every time that I sit down (although they don't know it ). Casinos would never have commissioned or installed video poker games that pay out over 100% if they weren't certain that the overwhelming majority of the players would fail to play perfectly and that they would still enjoy a healthy return while advertising "Over 100% Payback" on their marquees.

[reho33]

Years ago I could find 9/6 machines at Foxwoods. Now they are all gone because the casino got "hip" to the fact that people are slowly coming around and learning the strategies much better but still not perfect. My game happens to be "Jacks or Better". Now I have played on alot of 6/5 machines and unless I was playing thousands of hands, did not see much short term loss over a 9/6 machine. But I enjoy the game and I am not looking to supplement my income being a gambler. Bob Dancer is one that happens to play perfect strategy but like you said the majority of people out there won't and thats where thecasino makes it's money.

[StatFreak]

The difference between 6/5 and 9/6 in lay terms is that you won't be able to play as long with your money. Without getting into any math, each time that you get a flush or a full house, just make a mental note that you would have had an extra 5 coins (Flush) or 15 coins (FH) to play with. Keep track, and when you run out of money, take note of how many extra credits you would have still had on the machine had you played full pay.

[Stolistic]

Interesting. That's the kind of test I would be likely to do, and it certainly points to the code determining the draw cards after the initial draw. If this is how it's done today (re: GK) I'd like to hear the designers' reasons for such an implementation. Not to be redundant, but selecting all 10 cards at the onset using a single PIA, or selecting the five draw cards at a later time with a second PIA shouldn't affect the long-term outcome in any way, unless there are issues with the RNG.

Stolistic, what's your take on this?

Knagl's test was a great way to test it without looking at code.  The complete state of memory is kept in the NVRAM files upon exiting MAME (the same way the battery keeps the RAM alive on a real machine).  So if 10 cards were drawn before a simulated power-down, those 10 would remain on power-up.  I will have to look at the code specifically but I do believe the system is doing a "double dip" and drawing 5 cards then 5 cards again.  Even if I confirm the findings in code, another version of the game could use the draw 10 method instead.  There was mention above that both methods could exist depending on jurisdiction.

[CaptainHappy]

Knagl,

Thanks for the "confirmation," I knew that it is a heavily discussed topic that many are secretly curious about.

K+ for posting it! Yes, K+ to other contributors also!

CH

[uniman]

Always heard IGT drew 10 cards with five in a stack waiting to be drawn and Bally's had a draw card behind every one of the five face up cards. That must be myth.
Thanks for the info, great stuff!
I often play at the Eureka in Mesquite NV. They have Jacks or Better 9/6 even in nickles. Can't find that in Vegas!

www.wizardofodds.com is a great place to get info on various game strategies.

[doublediamonddlx]

I was told at one time that regulations forced the manufactuers (IGT, WMS, etc...) that if there was a power surge mid draw and that computers had to keep the other 5 cards in memory so when the machines power came back that it would be able to finish the game with the other 5 cards (10 card draw). 

I was also told that the 5 draw cards "sit" behind the each set of hold cards and that the selection of the hold cards was very specific for the payout of the hand.
Maybe Stolistic can straighten us all out if he finds the right code....

KNAGL:

I think it would be interesting to see all the different odds based on verbage printed on table games....especailly for craps and blackjack.

[jay]

In terms of a game of poker, it would be of no concequence if the computer drew 10 cards from a 52 deck, presented 5 and then had the other 5 in reserve
Vs the drawing of just 5 cards followed by a continual shuffle of the remaing cards until they were needed. The outcome is still the same.

The draw cards however should not "sit" behind each of the drawn cards as then the sequence of what is dealt next is not sequential to the draw and would impact the outcome of the game depending what is held or not.

From a CMOS memory perspective (for game completion after a power outage) the order of 52 cards takes very little memory... at worst case 52bytes, which would easily be compressed to 13 bytes since 4 x 52 = 208 which is less than 1 byte of memory space, and with a fancy algorthym could be compressed even smaller. Even in 1982, 8K chips far exceeded the amount of memory needed to preserve this info.

[Op-Bell]

I'm very skeptical of what Vasquez told the American Casino Guide. I don't know exactly how everyone programs their shuffle and draw, but there are really only two ways to do it -

1. Have an array of 52 memory locations and actually shuffle the contents;
2. Just pick 5 random numbers 1..52, all different, and call those the card values.

If (1), then there's no need to "pick" cards out of the deck - you just take the first 5, whatever they may be, and then draw from the next 5, just like off the top of a physical deck. All you have to do is shuffle until the game starts, then stop shuffling until it's ended. The cards are "picked" for you just by virtue of being in the first ten locations. I don't believe the order of the card array gets disturbed after the game starts.

If (2), then there's no actual shuffle going on at all, is there? You're just running the RNG.

I've asked my programmer friends at IGT about the way they do it. What's stored in the NVR for game recall is just the cards currently displayed, not the discards or unpicked draws. The whole shuffled deck is in NVR too, plus the latest state of the RNG, so that it carries on where it left off after a power cycle. I can't think why the draw would be different after a power cycle, unless possibly the game always does at least one shuffle on power-up, so although you still draw from positions 6 thru 10 the cards have been changed.

The draw cards however should not "sit" behind each of the drawn cards as then the sequence of what is dealt next is not sequential to the draw and would impact the outcome of the game depending what is held or not.

In fact it makes no difference at all. If you need to draw 2 to make a hand, in either case those two cards have to be in two particular positions - either behind the two discards, or first and second in the draw. This used to be the subject of endless arguments on alt.gambling.poker, thanks to the misinformation in Crevelt's and Scoblete's books. You can absolutely prove the odds are the same by working through all the combinations, trying it both ways. I don't suggest anyone try it with a 52 card deck, but a 3-card poker game with 6 cards in the deck can be fully analyzed on paper in fifteen minutes. In any event, there are no "parallel" (behind the display) machines oday. There were once - I've got one in my hallway. It's the old TTL based Bally Computer Poker, which doesn't actually contain a computer and has no memory at all. It uses shuffle method 2, and at the start of the game it stores ten card numbers in two rows of 5 latch chips. When it draws, the card from the second row steps forward to the first row. But since the introduction of microprocessors and video, no other machine has ever done it that way.

[StatFreak]

I'm very skeptical of what Vasquez told the American Casino Guide. I don't know exactly how everyone programs their shuffle and draw, but there are really only two ways to do it -

1. Have an array of 52 memory locations and actually shuffle the contents;
2. Just pick 5 random numbers 1..52, all different, and call those the card values.

If (1), then there's no need to "pick" cards out of the deck - you just take the first 5, whatever they may be, and then draw from the next 5, just like off the top of a physical deck. All you have to do is shuffle until the game starts, then stop shuffling until it's ended. The cards are "picked" for you just by virtue of being in the first ten locations. I don't believe the order of the card array gets disturbed after the game starts.

If (2), then there's no actual shuffle going on at all, is there? You're just running the RNG.

I've asked my programmer friends at IGT about the way they do it. What's stored in the NVR for game recall is just the cards currently displayed, not the discards or unpicked draws. The whole shuffled deck is in NVR too, plus the latest state of the RNG, so that it carries on where it left off after a power cycle. I can't think why the draw would be different after a power cycle, unless possibly the game always does at least one shuffle on power-up, so although you still draw from positions 6 thru 10 the cards have been changed.
...

I used method 2 when I wrote my programs, but that was a long time ago. If I understand you correctly, the folks that you talked to at IGT used method 1. However, you said that unpicked draws were not stored, but then said that the entire shuffled deck was stored, which would seem to be a contradiction. Did I misunderstand something?


BTW, Norm Wattenberger, in designing his CVData blackjack analyzing software, actually went to the trouble of setting up a sub-system within the program to allow the user to build specific complex physical shuffles to emulate a particular casino's method while still allowing for a flexible amount of random variation in the number of cards grabbed and their interlacing during the shuffle, and so forth.

Activating that part of the program slows down the analysis significantly, as one would expect, since it is so CPU intensive, and even though he put a lot of work into designing this sub-program, he discourages using it in his manual, stating that using the sophisticated shuffle routine "rarely makes a significant difference" in the outcome vs. using a normal RNG shuffle, which runs many times faster.

Of course, a typical suggested run is 300 million rounds, and the program can run up to 300 billion rounds in one analysis. I typically set the program to run 2 billion rounds when analyzing new strategies or counter measures to rule changes.

[Op-Bell]

If I understand you correctly, the folks that you talked to at IGT used method 1. However, you said that unpicked draws were not stored, but then said that the entire shuffled deck was stored, which would seem to be a contradiction. Scratch Head Did I misunderstand something?

Yes, you misread what I said. Only five cards are stored as part of the saved game records, just the outcome, not the discards and undrawn cards. And the entire deck resides within NVR, but as it is constantly being shuffled, the only importance of this is that it lets the shuffle continue where it left off after a power hit instead of starting afresh from an orderly state. During the course of the game, other values are temporarily saved in NVR so the game can be resumed if interrupted. I don't know exactly what these are.

The shuffle algorithm I always use is one I first saw in a text blackjack game written in 1975 for the 8080, and I've never found a better one. Pseudo-code:

Code:

array deck(0..51)
for i=0 to 51
   j = int(rand*52) // value 0..51
   k = deck(i)
   deck(i) = deck(j)
   deck(j) = k
next i

[CommTech]

there are some machines around that have a much lower payout for the Royal.

They are rare, but they can be found.

The advantage to playing these machines is that because of the lower payout on the Royal, the machines will actually pay the smaller jackpots more often.

And the Royal itself will actually hit slightly more often.

Actually, this is a True statement. You guys are forgetting about the Wild Card Games.  Atlantic City Jokers Wild for example would only pay 500 Credits for a Royal Flush (Full 4000 for 5 of a kind), but you would also hit the lower amounts more often due to the wild card factor. You would obviously also hit Wild Royals more often, but with the same lower payout (Lower payouts on the Deuces Wild machines but also more chances to hit Wild Royals).
The article did not say your odds are better with these games, just that these games exist.

[StatFreak]

If I understand you correctly, the folks that you talked to at IGT used method 1. However, you said that unpicked draws were not stored, but then said that the entire shuffled deck was stored, which would seem to be a contradiction. Scratch Head Did I misunderstand something?

Yes, you misread what I said. Only five cards are stored as part of the saved game records, just the outcome, not the discards and undrawn cards. And the entire deck resides within NVR, but as it is constantly being shuffled, the only importance of this is that it lets the shuffle continue where it left off after a power hit instead of starting afresh from an orderly state. During the course of the game, other values are temporarily saved in NVR so the game can be resumed if interrupted. I don't know exactly what these are.

The shuffle algorithm I always use is one I first saw in a text blackjack game written in 1975 for the 8080, and I've never found a better one. Pseudo-code:

Code:

array deck(0..51)
for i=0 to 51
   j = int(rand*52) // value 0..51
   k = deck(i)
   deck(i) = deck(j)
   deck(j) = k
next i

Thanks Op-Bell. I appreciate your helping me understand how IGT implemented things.
I like your shuffle algorithm: simple and efficient. Of course, the tricky part of any gambling program is creating an RNG that passes muster.

[StatFreak]

there are some machines around that have a much lower payout for the Royal.

They are rare, but they can be found.

The advantage to playing these machines is that because of the lower payout on the Royal, the machines will actually pay the smaller jackpots more often.

And the Royal itself will actually hit slightly more often.

Actually, this is a True statement. You guys are forgetting about the Wild Card Games.  Atlantic City Jokers Wild for example would only pay 500 Credits for a Royal Flush (Full 4000 for 5 of a kind), but you would also hit the lower amounts more often due to the wild card factor. You would obviously also hit Wild Royals more often, but with the same lower payout (Lower payouts on the Deuces Wild machines but also more chances to hit Wild Royals).
The article did not say your odds are better with these games, just that these games exist.

Well, he doesn't refer to wild cards at all in his statements, and I wouldn't call wild card games rare. He is certainly implying that one can hit a regular Royal more often -- again, he says nothing about wild cards, and the wild card games that I've seen have two separate payouts for a wild Royal and a natural Royal, so any responsible statement would need to point out the difference, IMO.

Can you post the paytable for the Atlantic City Jokers Wild?

In the Joker's Wild poker that I have in "Frugal VP" (a VP analysis and training program similar to Bob Dancer's), a wild Royal pays 500, five-of-a-kind pays 1000, and a natural Royal still pays 4000 coins. In addition, the odds of getting the Royal with best play are in the 39000-40000 to 1 range, similar to jacks or better.

There is a Five Jokers variant listed in the program that pays 4000 coins for five Jokers and for a natural Royal, but then the any five-of-a-kind pay drops to 75 and the wild Royal drops to 100. The odds of getting a natural Royal with best play on this machine shoots to over 77000 to 1, and the odds of getting the Five Jokers is listed as 171,355.6 to 1.

PS: The odds of getting a natural Royal with best play on games with wild cards are generally worse than on games without because the player holds the wild cards more often and does not hold cards to the natural Royal as often as they would on non-wild games. For standard Deuces Wild, for example, the odds of getting a natural Royal are in the 45000-46000 to 1 range, compared to the 39000-40000 to 1 range for Jacks or Better.

So, if the statements in the article were referring to wild card games, then the statement that the Royal hits slightly more often is false, and if the statement referred to hitting wild Royals, then it is also false, because those hit MUCH more often, not slightly more often.

[Op-Bell]

I agree with Stat, the article does imply that the chances of smaller payouts are increased because the Royal pays less. I think the writer probably heard this is true for regular reel jackpot games and doesn't understand that pokers are different. Having said that, just because something looks like a poker machine doesn't mean that it plays like one. I've seen machines overseas - not in the US - that played like slot machines, with small but very frequent pays for 4 of a kind. They wouldn't be allowed in Nevada, which is the area being written about. As far as I know, all the other gaming jurisdictions in the US follow the same rules, though if there's any doubt, just pull up the Gaming Compact or Lottery Regulations and look up the technical standards.

Talking about five-wild hands, I hit the biggest jackpot in my life on a deuce-joker machine in Biloxi once. It dealt me three deuces and a joker, and I drew the fourth deuce for a progressive of something like 12,750.

[StatFreak]

...
Talking about five-wild hands, I hit the biggest jackpot in my life on a deuce-joker machine in Biloxi once. It dealt me three deuces and a joker, and I drew the fourth deuce for a progressive of something like 12,750.

It must have been a great feeling to see that last deuce fall into place! I hope that you did some serious celebrating that evening.

[Op-Bell]

I hope that you did some serious celebrating that evening.

The following evening, actually. I was with a party and we all had comp'd rooms at Treasure Bay. I'd been losing all weekend. I couldn't sleep, so I went down to the boat alone at about 4.30am with the last of my cash. My luck didn't change and I was playing single credits to eke it out, but when I got down to 20 left in the world I thought fcuk it and hit the max bet button. That was the hand that hit. It didn't register at first, because I was hoping for a court card to make a wild royal.

[StatFreak]

I hope that you did some serious celebrating that evening.

The following evening, actually. I was with a party and we all had comp'd rooms at Treasure Bay. I'd been losing all weekend. I couldn't sleep, so I went down to the boat alone at about 4.30am with the last of my cash. My luck didn't change and I was playing single credits to eke it out, but when I got down to 20 left in the world I thought fcuk it and hit the max bet button. That was the hand that hit. It didn't register at first, because I was hoping for a court card to make a wild royal.

I'd have been thinking the same way, hoping for a wild royal at 60 credits and not daring to think that I'd get the fourth deuce for 12,750.
Was it dollars or quarters?

[Op-Bell]

Quarters. I was so broke by that time, if there was a penny machine I would have been playing that instead. My sister in law said I was a cheap-ass wuss (she never plays less than dollar machines) and I hadn't won 3250 bucks but actually lost 9500.     I don't know what she would have said if I hit it with a single credit played - well, I wouldn't have mentioned it, probably.