Poker For Roulette - side bet

Back up the duel a few paces. Re-read my statement.



In my post above, I noted that YOU said that Roulette is boring. I never stated in my post how I feel about it. You made an anology of my bet to the FireBet. That leads me to believe you find craps more exciting. I never stated how I feel about craps, just that I know there are people that find it more boring that Roulette.

Personally, yeah, I find craps (even without the FireBet), a LOT more exciting than Roulette. The last few times I've been in a casino, the only reason I ventured over to the Roulette tables were to take a glance, and possibly a photo, of the history display.


I believe that among the active members here (i.e. the people that would vote), most would say Craps is more exciting than Roulette. Therefore, there's no reason for the poll.

However, among the general gambling table-playing public, I have a feeling more people will say that Roulette is more exciting. Or at least easier to understand, and therefore, of the two, the game they play more often.

You were able to take a photo of a Roulette table inside a casino?

I once got a photo of a Ricochet table, not tryin to clutter your thread or anything, but I'm kind of proud of it ...







BUT

I got in big trouble for it, they took me into the security office and threatened to shove a swordfish up my butt.


How did you get away with it? Were those the ones on your PokerForRoulette web site ?

BTW, what is the patent number for PokerForRoulette, I didn't find it anywhere on the web site or the brochure.

Photos? I just ask for permission - and, in some cases, hand the pit boss a business card.

I turn the flash off, and take it just of the history display.


Patent? Still pending....

The Excalibur used to let people take photos inside of the gaming floor. I have no idea if it still does. I've also taken photos in the Monte Carlo without asking permission and had no "heat", but those photos weren't directly of the table layout.

Well, they didn't really shove a swordfish up my butt, they told me the flash I used messed up some of their equipment.

DJ, what's the pending patent NUMBER?


BTW, the Wizard has used the above photos in his description of the game invented by Pat Bowling, Ricochet

wizardofodds.com/ricochet If you read this, Mike, Pat said its OK with him for you to use 'em, and he thanks you.


I took a photo of the Sandia casino in New Mexico from the outside...



I asked them if I could take any photos inside. Then they told me that the entire casino and all the land (yes, including the mountain) can't be photographed because it's all sacred. Its a digital photo, though, so I figured the old ways don't apply.

Man you guys are missing a heck of a blizzard tonight... I can't open my front door!

Quote: discflicker

DJ, what's the pending patent NUMBER?

I'm pretty sure DJ filed a provisional, which you won't be able to review publicly.

I know that to be the case, Dave did cover it.

Maybe this is a questions for the DYI provisional patents thread, but if someone wants to make public a product based upon a provisionally patented method (a public claim of sort, for example, DJ's web site), if the public can't see the document to obtain detail so as to question or dispute possible infringement, doesn’t that make it a free-for-all zoo of claims and disputes for the lawyers to all make money upon trying to resolve ownership? Like, say someone already owns a similar wager and become aware of DJs web site, and feel it might be infringing... how can they know for sure, or are they just supposed to wait until DJ makes a bunch of business connections and plans and launches his product?


DJ wants to keep this thread on topic. I have some procedural questions on the implementation. I will try to keep these questions on topic by not mentioning anything about my solutions here. For example, I'm wondering how DJ envisions the jackpots to be initialized, for example, does the house instantly re-fill the (for example) $5,000 starting jackpot, or does that get filled first from non-jackpot bets? Another, if I may be so bold, is when a jackpot or partial jackpot is hit, are the winnings split up between the multiple winners, or does the house pay full amounts for all winners?

Questions like this will shed light on the impact of, say a short run of 7 identical numbers. If the provisional does cover these details, that’s great. In this case, does DJ envision that these various approaches to handing jackpots are somehow made available to end-users as a set of options, or will he decide upon only one way of implementing these 4 combinations? (I have a feeling he'll say that's up to the distributor he finds).

DJ, if you feel these questions are off topic, I will stay completely off this thread. If not, I have a bunch of questions like it, and a good reason to be asking about these details beyond any proposals.

No, I don't think these questions are off-topic. But thanks for understanding my desire to keep this thread primarily about my idea. For the record, I have a link to this thread on the History page of my Poker For Roulette website, and therefore what to keep it pretty much on-topic.


In a prior post about the patent number, I responded that it's pending. When I posted that, the term 'provision' simply didn't come to mind, but a provisional patent is what I have.

I have the same concern about the inability to search provisional patents, and the posibility of multiple people filing overlaping / infringing claims during the one year provisional period. I guess that's one of the down-sides of using the provisional patent system.


Most of your other questions are covered on my website, but I'll answer them here.

You're right that some of the questions will be answered that it's up to the distributor. That's not to say I haven't thought about the question. It's that I believe that a distributor will have more experience, and knowledge of the regulations, than I do. Some of the details will be actually be up to the casino. It is quite possible that this same bet, running in different casinos, will have different rules, or at least different pay tables.

But the minor details? Frankly, I'm perfectly happy letting the distributor deal with the minutiae.


I don't know how the jackpot re-seed money is generated in progressive games such as Carribean Stud or Let It Ride. Suffice to say that since only a portion of the coin-in goes to the progressive, whatever formula they use must be supported by the math of the game.

My math shows that a NON progressive paytable, with HIGHER jackpot prizes than the progressive's seed value, still yeilds a nice house edge. Therefore, I'd actually think that the full jackpot be awarded, and the casino re-seed it on their own.

For the record, I'm open minded to customizing the paytable, progressive meter increments, and re-seed value and method as requested / needed.


On a side note, I do understand how Poker Bad Beats are calculated. But part of the reason is, the poker rooms are very open about it. Unlike the Caribbean Stud or Let Ir Ride, there are no intermediate or partial payments, it's easy to describe. The jackpot values displayed in the poker room is only 80% of the collection. That 80% gets divided when the bad beat is hit, and the remaining 20% become the new bad beat jackpot. Or only 16% with 4% (20% of 20%) being reserved for the NEXT jackpot. The casino had to seed it from their own pockets only once. I would not be surprised to learn that a small percentage is shaved off the top until the casino's initial seed money is returned. In fact, it' probably "initial seed plus interest". Whatever.


In my game, Progressive Jackpots are NOT shared. Since every player is betting on the same outcome, every player deserves the full jackpot payout when it hits. The distributor I met with about the Hit It Again idea said that may be a stumbling block. Not that he didn't believe the math, but that casinos have a way of not wanting to let go of money, even if the math supports it. Go figure!

He was saying that there may need to be an aggregate cap. As a result of THAT detail, I created two pay tables for Poker For Roulette with medium jackpots and two with small jackpots. Read more about it on my Paytable page.

When I think about the aggregate problem, it seems like a non-issue if, say, 5 players hit it at once. However, in Asia, there are installations of RapidRoulette (or it's competitors) that have over 200 player terminals running off a single wheel. If all those players were betting and the jackpot hit, it's east to see why the casino would prefer the aggregate cap. And THAT'S what I created the four paytables. The smallest jackpot seed, even if paid 200 times, is still low enough that the casino will feel it, but more or less brush it off.


A short run of 7 identical numbers, indicates 3 Five Of A Kind jackpot hits in a row.

What isn't obvious is that prior to the first, there is a Four Of A Kind hit. Any Four Of A Kind pays a percentage of the jackpot. It also lowers the jackpot meter by the amount paid to one person. When the next spin produces the Five Of A Kind, it is now paying the full reduced value. This means that for players that had bet on every spin, the Four Of A Kind was irrelevant. I will point this out to the distributor.

I have considered that the system can "lock-in" the jackpot value at the time the bet is placed. However, in your 7 identical numbers secnario, this would mean that the next number also generates a Five Of A Kind. Even greedy gamblers would admit that it should be paid at the re-seed value. That wouldn't happen if the jackpot was "locked in". In your scenario, there is yet another Five Of A Kind. Casinos would freak out if they had to pay the full, locked-in, jackpot value.

Note that, except for certain Wild Five Of A Kind combinations, Five Of A Kinds are followed by a Four Of A Kind. As stated, it lowers the jackpot meter. This will result in the jackpot being lower than the re-seed value. Whether the Jackpot then be "topped off" or not is another decision that I'll defer to the distributor or casino.


I hope this answers those questions. Feel free to ask some additional questions.

Well alrighty, now were talkin!


In order to make your wager into a complete package that a Casino might buy, I'm thinkin these choices should be already considered and everything already worked out. Again, I'm not critisizing you, but if you read your own descriptions, you'll see that everything isn't worked out from the standpoint of profit/loss over these various options.

Similar to the details of "payouts over the various ways jackpots are implemented", I think there are also some things you need to work out in the game procedures. I will be asking you questions about that as well. Some of these issues might effect payouts as well.

If everything WERE worked out, I'm thinkin that would be a great way to boost sales; your customers would NOT have to do all this work up front before actaully considering buying it. Do you agree that they would need this information before buying?


What I said about having other motives deals with my own implememntation, which promised I won't get into in this thread; Since you have boldly gone where not many game designers have, and that is crossing the digital divide on Roulette tables and introducing a computer interface via these dealer ABD procedures, the benifits are AMAZING once that divide is crossed. For example, if the Roulette game has established an electronic ABD state, they could then use a new invention (that have thought about patenting) that places a vail of light around the entire table, so if anyone crosses it with their hand, an alarm is sounded. I have thoght about a LOT of things like this, but they alll depend on having the electronic ABD signaling.

OK, I won't try to put you to sleep right here with a 90 page rant, so I will be cheking back with additioal questions and issues.


THANK YOU DJ for allowing me to discuss this here.

marty

Some of the points you bring up, require additional math analysis. That sort of stuff is beyond my abilities. And even if I could do it, it would still need to be verified by a recognized casino math expert. The Wiz is the first, and maybe only, person I'm calling to do that, once I'm ready to go.

When I said that some decisions are being left to the distributor, I also mean decisions about who will do the math analysis will be up to them. If they tell me that they will handle the math, fine. But they will probably tell me that the math is my responsibility. When they do, I'll bet that they'll also tell me what sort of reports that are required, both to get this approved by Gaming, as well as to show to the casinos.

I'm not trying to dodge your questions. It's just that I don't know how to do some of the analysis that you're talking about. I agree that it needs to be done, but until I get a distributor that wants to move forward with this, that's stuff that will wait.

You've mentioned that I'll need it before I can sell my idea. Mind you, I do not consider the distributor to be my customer. They will be my partner. The casino is OUR customer. Only once they think we've got all the bases covered, will we start knocking on casino doors.


I'm not sure what you mean by "ABD". Do me a favor, if you can explain it in a sentence or two, please do. Otherwise, please create a new thread for it.


For what it's worth, my math page shows the number of times each combination will occur amoung all 79 million combinations of 5 spins. In short, it's showing the statistical chance of occurance. Many of the additional questions need to have a probability chance determined. That's way over my head.

As I mentioned previously, the math for figuring out the statistical analysis was over my head to the point where I simply wrote a program to cycle thru and analyze all 79 million combinations. That takes my computer just under an hour. I suppose I could modify my program to do random analysis and produce pseudo probabilities, but, I already know that the RNG in my computer isn't random enough for the task.

Hmmm.... While typing that sentence, I suddenly want to do it, even if the results are NOT truly random. I'm gonna do it. Expect the results in a couple days.

Quote: discflicker

In order to make your wager into a complete package that a Casino might buy, I'm thinkin these choices should be already considered and everything already worked out. Again, I'm not critisizing you, but if you read your own descriptions, you'll see that everything isn't worked out from the standpoint of profit/loss over these various options.

DJ isn't selling to casinos. He's selling to distributors who will bear the responsibility of polishing up the game, handling regulatory, and selling to casinos. That's virtually always the way it works. Very few game developers run the full gauntlet because most game developers don't want to become a licensed gaming vendor in multiple jurisdictions. That's why distributors exist. You can keep more of the pie if you do everything yourself, but the pie is almost invariably smaller.

DJ : You should look at the patterns of resolution not just each individual possibility of 5 numbers... because of the strings of bets there is a correlation between each result (Bet A+1 contains 4 of the numbers of Bet A). A Monte Carlo simulation using your pay tables would reveal if there are large gaps between pay offs, especially if those pay offs suddenly result in a large lay out by the table (the 4/5 of a kind issue you mentioned).

Simulating those in a script wouldn't be too hard (the hardest part is identifying a sequence as match to a pay out), and then you can run several million output sequences... would be good enough to see if there is a problem, and any relatively robust pseudo number generator will do for this sort of test.

MathExtremist -

Thanks for the confirmations.

Sometimes I *know* I'm right, but DiskFlicker makes me question myself.

On the other hand, I don't mind questioning myself. It makes me give everything a re-think. Including doing simulations with my computer. I almost can't wait to get home and start coding....

Quote: thecesspit

A Monte Carlo simulation....

Do you mean implementing it at a local church Monte Carlo night? Simulate it live, with me working this side bet, manually doing all the stuff the electronics would do, while a dealer handles the rest of the game?

I don't think that would show me anything.

First of all, how many spins occur in a typical setting? 200 on a good night? Oh, sure, there will be plenty of pairs and flushes, maybe one or two trips, but I doubt I'd see many of the other combinations. Those photos of history displays I've been taking are interesting, but are rare.

On a side note, my wife wants me to play Roulette in a casino for an hour or two to get a feel for how well this bet will work. I highly doubt that doing so will teach me anything that I can't learn by looking at the history displays, with most of them being boring, showing nothing more than a pair or two....

Quote: DJTeddyBear

Do you mean implementing it at a local church Monte Carlo night? Simulate it live, with me working this side bet, manually doing all the stuff the electronics would do, while a dealer handles the rest of the game?

Heh, no. Monte Carlo simulation is the technical name for running a simulation using a random number generator for millions of trials. Contrast with an analytical solution, where you directly calculate the probabilities for each possible outcome or series of outcomes. Thecesspit's suggestion is a good one - you should produce a histogram of 2, 3, 4, 5-result aggregates so the casino has an idea of how the bet will behave over time.

Quote: MathExtremist

Heh, no. Monte Carlo simulation is the technical name for running a simulation using a random number generator for millions of trials. Contrast with an analytical solution, where you directly calculate the probabilities for each possible outcome or series of outcomes. Thecesspit's suggestion is a good one - you should produce a histogram of 2, 3, 4, 5-result aggregates so the casino has an idea of how the bet will behave over time.

What he said, basically.

I think your wifes suggestion is a good one... feeling how the game is played and how the excitement (or lack of it) of the bet feels would help you with the mechanics of a Roulette table. You might learn something new about Roulette the game feels and improve your bet. It might be a waste of 2 hours and $100 bucks.... but just the feel of 5 spins and how long that takes for the bet to resolve, tracking the hands while playing, etc...

Quote: MathExtremist

Heh, no. Monte Carlo simulation is the technical name for running a simulation using a random number generator for millions of trials. Contrast with an analytical solution, where you directly calculate the probabilities for each possible outcome or series of outcomes. Thecesspit's suggestion is a good one - you should produce a histogram of 2, 3, 4, 5-result aggregates so the casino has an idea of how the bet will behave over time.

OK, so, and once again, I don't want to clutter up the thread, if DJ goes homes and whips up a routine like I described, he could plug it into a host testing system to perform empirical (Monte Carlo) testing upon it.

THEN AND ONLY THEN, DJ can play around and tweak the payout tables and jackpot use parameters used in his callable routine, and for each combination:

1) Figure out exactly what the payouts really are, what the worst case scenarios could be (see note below on that one)
2) Figure out the optimal way to use the jackpot, as discussed previously

I said this before but MathEx twisted it around into something I claimed only SpikerSystems can do.

Note below: This bet has a danger of hitting the house really hard in obscure runs of identical spins. I KNOW that anyone can take a theoretical approach to determining this stuff, but without the skill of a Wizard, I wouldn't try it without having the empirical tests to back me up, especially because of the impact of the other factors discussed. By using a system to test such an algorithm, DJ will be using the Monte Carlo... it cannot lie.


One more thing, and DJ granted me 2 sentences to explain it,

ABD stands for All Bets Down, as described in my 1990 systems overview manual, my point being that there really is a lot to think about in order to cross the "digital divide" on a live Roulette table, and the system that host tests your payout routine needs to already have that stuff all worked out in order to accomplish the Monte Carlo, in fact, its part of the considerations.

I think I can plug a routine structured as I described it into my payout routines and be able to demonstrate how it really works in about a couple of weeks of coding, and immediately, we can all play against each other over the internet, and then if DJ wants to run Monte Carlo simulations, he can easily do 30 concurrent tests, each using various strategies all at about 1 roll per second.

Quote: discflicker

I got in big trouble for it, they took me into the security office and threatened to shove a swordfish up my butt.

You're not supposed to take a picture of a casino employee actually vacuuming up people's money!!!

The ricochet table uses stacks of color-coded same-denomination chips just like a roulette table, so she couldn't cash 'em in without someone being curious.

Hey DJ just a thought, NOTHING TO DO WITH MY STUFF, if you haven't already specified this in your provisional, your concept can easily apply to an electronically assisted game of craps!

(1) (2) (3) (4) (5)

These positions can hold a great variety of comparable entities. In assembly operation terms we call this "lift and carry" from station to station.

If the point sum rolls of a craps game are used, the payout would be strange because of the varying odds of rolling various point numbers, (there are ways around that, see note below), but here is the really cool part...

These bets would all lose if a 7 out occurs first !!!

So, as the bet advances from station (1) to (2), it might not SURVIVE all the way to (5), but if it does, perhaps the addition of partial wildcards wouldn't be needed because anything that does SURVIVE to (5) already should payout big by itself.

note below: one way is to have the player choose a single point number to fill, for example, the 6 point, and only the 6 point are considered among the 5 stations for 3 of a kinds, etc. To make a poker straight, though is tough, and so, the other way is to use color distinguishable dice, say blue and white, ... these would NOT have ANY impact upon the craps game, but (for example) the white die could be used to provide the number, and perhaps the relative number (the blue die is lower than he white) can be used to make up "red and black colors" for flushes (in Poker-For-Roulette, "flushes consist of only the two red and black color combinations as well).

DJ !!!

Think about including verbiage in your utility to specifically allow for the possibility of such applications to your bet.

And if you think that really makes good sense, maybe you should consider changing the name of the wager from Poker-For-Roulette to something like Poker-Survivor-Side-Bet because YOUR BET CAN BE USED IN KENO, BLACKJACK, Etc, Etc Etc. !!!!

Don’t forget THE FUTURE OF GAMING REVOLVES AROUND ELECTRONIC IMPLENTATIONS SUCH AS THESE.

You are crossing the digital divide... the line drawn by the ABD signal... MAKE THE MOST OF IT!!!

marty

Quote:

if DJ wants to run Monte Carlo simulations, he can easily do 30 concurrent tests, each using various strategies all at about 1 roll per second.

That's much too slow. I have an RNG sim of one of my roulette games and it churns through 100M roulette spins in 12.7 seconds. That's on my laptop which only has a 1.3Ghz dual-core Pentium, and I'm just using the free MS Visual C# 2010 Express.

If your code can only generate thirty spins per second, it's not going to be useful in Monte Carlo simulations. Getting to 1M spins will take you over 9 hours, and for DJ's bet a sim should be far more than 1M spins due to the correlation effects noted above. The chances of 5 of a kind are only 1 in 2.08M, etc.

I'd advise you to use a different technology stack if you're going to do random simulations. It doesn't look like your demo harness will cut it. My advice: stick to minimal UI while your sim is running. I spit out a status line every 10M spins or so, just to see how things are going (and to prove the code is still running). But a big heavy GUI like you have with all those buttons and textboxes will hurt your runtime.

Quote: MathExtremist

I have an RNG sim of one of my roulette games and it churns through 100M roulette spins in 12.7 seconds. That's on my laptop which only has a 1.3Ghz dual-core Pentium, and I'm just using the free MS Visual C# 2010 Express.

Wow! That's dang fast. For the record, I'm estimating that 100m random spins will take me a little over an hour when I'm done coding. Mind you I'm using a freeware version of Basic (NOT Virtual Basic). Once I am satisfied with short term results (say 5m spins), I'll set it up to run 18 hours...

Quote: MathExtremist

... one of my roulette games ...

How many roulette games do you have? What do you do with them? What do you do in the gaming industry? FYI: This might be fodder for a new thread, or private message....

Quote: MathExtremist

If your code can only generate thirty spins per second, it's not going to be useful in Monte Carlo simulations. Getting to 1M spins will take you over 9 hours, and for DJ's bet a sim should be far more than 1M spins due to the correlation effects noted above. The chances of 5 of a kind are only 1 in 2.08M, etc.

He said ONE roll per second, not thirty.

Also note that there are 30 combinations of 0 and 00 which pay the full jackpot, in addition to the 38 natural five of a kinds. That means, the full jackpot has a chance of 68 in 79M or 1 in 1,165,223.

Quote: discflicker

ABD stands for All Bets Down, as described in my 1990 systems overview manual.

Oh. Silly me. I thought you were refering to an industry standard abbreviation that I hadn't heard before.

Quote: discflicker

I think I can plug a routine structured as I described it into my payout routines and be able to demonstrate how it really works in about a couple of weeks of coding, and immediately, we can all play against each other over the internet, and then if DJ wants to run Monte Carlo simulations, he can easily do 30 concurrent tests, each using various strategies all at about 1 roll per second.

Seems like it shouldn't take that long to code, but whatever.

And 30 concurrent tests? What the heck for? "Strategies" normally suggests player strategies. I would think that I only need one test, then feed the result into Excel to tweak or confirm my payouts.

Quote: discflicker

Hey DJ just a thought, NOTHING TO DO WITH MY STUFF, if you haven't already specified this in your provisional, your concept can easily apply to an electronically assisted game of craps!...

It's got nothing to do with my stuff either.

I'm not interested in revolutionizing the world.

I'd just like to make a few bucks off a concept that people have told me is pretty cool.

Quote: DJTeddyBear

He said ONE roll per second, not thirty.

He also said something about 30 tests in parallel, so I was assuming he meant all 30 tests running at 1 roll/second each. If it's 1/30th the speed, that's even worse - it'll take over 11 days to generate 1M trials, and nearly 3 months to generate 10M.

Point is, both he and you should be using different software technologies to do anything that involves randomly generating millions of trials. I tend to use C# more than Java these days, but I admit that I haven't used Eclipse in a few years and it's probably better than it was. But either of them should be orders of magnitude faster than a BASIC interpreter. Also, BASIC or VB might be good for easily slapping together GUIs (actually, so is VC# - it's drag-and-drop) but you should never use a constantly-refreshing GUI while doing RNG trials. Your processor will spend much more time updating the screen than actually generating data.

I'd be curious to see your algorithm for detecting each of the Poker Hands... I'm slowly working through a pattern matcher myself, and not really getting anywhere fast (slow, I can do slow, but it seems... inelegant).

I do this in Groovy (which is an extension of Java), but that's because I'm a scripter at heart and total efficency isn't my main motivation. I'm using IntelliJ as the IDE, by the way, and very happy with it in Version 10.

Quote: thecesspit

I'd be curious to see your algorithm for detecting each of the Poker Hands... I'm slowly working through a pattern matcher myself, and not really getting anywhere fast (slow, I can do slow, but it seems... inelegant).

I do this in Groovy (which is an extension of Java), but that's because I'm a scripter at heart and total efficency isn't my main motivation. I'm using IntelliJ as the IDE, by the way, and very happy with it in Version 10.

Check the open-source poker hand evaluator for hints on using lookup tables and bit masks to make this really fast. Data representation is everything.

I won't give you the algorithm, because that's proprietary. I'll make it public after I get the utility patent finished - but only if my attorney says it's OK to do so.

But I'll give you a hint:

The routine acts like novice poker player. It sorts the cards.

I.E.: If the results are 16, 5, 32, 16, 21, it's difficult to see the pair. But sorted, 5, 16, 16, 21, 32 makes that pair far easier to see.

Add wild cards and it gets trickier, but still do-able.

Math, Math, Math... Do you think I'm a total IDIOT ??? You're gonna sit there and tell me how to code a simulation? Youir gonna tell me what language to use?
PLEASE! First you tried to convince everyone that these studies weren't needed. Now you're gonna tell me how they SHOULD be conducted? PLEASE!!

DJ call me if you need help, I have this whole thing figured
out.

Please follow a simple flow-chart for determning the payouts:

Break it down into 2 searches, one with and one without the wilds.
Break the wilds search down into 2 searches, one for European, one for American Roulette.
For each of these 3 sub-searches:
Eliminate the winners by looking for the biggest payout bets first, followed by the next, etc.
You DO NOT have to test for all of the combinations, only all the winning combinations.
When you find ANY winner, you are done with that sub-search. If anything matched, payout the higest payout of the 3 sub-searches.


If you want, I can start it myself or get you started because obviously we're all interested.

Whoever said 2 week is too much for these code changes to my system has no FREEKING clue what he's talking about, if you dont know what your talking about, "silence is better".

Here is a detailed overview for creating the module to fit into my testing system.

If you dont want to code the module per my specifications, please have a look at this paper anyway because it lists several PARAMETERS that need to be considered, for example, if the jackpot is being split up or not, jackpot limits, etc..



marty

No worries... was looking for a short cut, but the tip for the open source poker hand analyzer will also help.

Cheers, good luck with the tests...

On your payout and math pages, I'm confused why a Wild five of a kind (which I assume is 5 green numbers in a row) is less common than A natural 5 of a kind (1,1,1,1,1 - for example).

Quote: thecesspit

On your payout and math pages, I'm confused why a Wild five of a kind (which I assume is 5 green numbers in a row) is less common than A natural 5 of a kind (1,1,1,1,1 - for example).

There are 36 numbered naturals, plus the 2 that are either all 0 or all 00. That's 38 naturals.

If there are four 0 and one 00, there are 5 ways they can appear. Ditto for four 00 and one 0.

If there are three 0 and two 00, there are 10 ways they can appear. Ditto for three 00 and two 0.

That's 30 ways of having 5 greens that aren't either of the two naturals.

---

2 ^ 5 = 32. Subtract the two naturals, and you have 30 combinations of five spins with 0 and 00 results.

Quote: discflicker

DJ call me if you need help, I have this whole thing figured out.

I also have it figured out.

As I've said, repeatedly, I used a program of my own design to cycle thru the 79 million combinations to come up with the numbers on my math page. Therefore, I certainly DO know how to take 5 random spin results, and check for winning combinations.


I may not know how to implement it into your system, but, since I have no desire to do so, I don't need to know how.

Good luck pulling together the simulation. Call me if you need any help. I'm glad I was able to spark your interest. Thanks!


I cant wait to play it!!

Quote: discflicker

Math, Math, Math... Do you think I'm a total IDIOT ??? You're gonna sit there and tell me how to code a simulation? Youir gonna tell me what language to use?
PLEASE! First you tried to convince everyone that these studies weren't needed. Now you're gonna tell me how they SHOULD be conducted? PLEASE!!

I haven't told you anything other than one trial per second is too slow for a Monte Carlo simulation. I'm not trying to convince "everyone" of anything; I'm just trying to give DJ some feedback on how to properly simulate his game. Choice of language for simulations is indeed important, and a better platform and/or tighter code will help DJ generate results faster. If your code takes three months to generate 100M trials, it is practically useless for Monte Carlo simulation.

Quote: thecesspit

No worries... was looking for a short cut, but the tip for the open source poker hand analyzer will also help.

Cheers, good luck with the tests...

The shortcuts are really dependent on exactly what kinds of hands you're trying to detect, and what you're drawing from. Bit strings and masks work well in poker because you only have one of everything. Roulette uses repeated independent trials which actually makes it harder. Even then, my first attempt would be based on frequency analysis - sort the non-wilds by frequency, then apply the wilds to that. If the most frequent occurs 4x, you have quads; if it's 3x, check for a pair: full house if yes, trips if no. A frequency of 2/1/1 plus one wild => trips. You can only have a straight if the most frequent occurs 1x and max-min = 4, and so on. You can shuffle these tests around to increase efficiency.

By the way, DJ, your website is triggering a malware error with my free copy of avast! antivirus (Win 7, Chrome). You might want to check that out. (Edit: see vscan.urlvoid.com and put your url)

Also, your own hand evaluator might be proprietary, but I wouldn't put that specifically into your claim language. If you do, anyone who uses a different hand evaluation routine can design around the claim.

Yea, DJ, I had some MAJOR problems with your other web site, so maybe you need to check for coodys.

MathExtreamist is recommending the same method I suggested, which is a process of elimitation. If you try the blaster approach, which is to loop through all combos, it will work, but will take the entire 13 seconds or whatever, which is OK but not for simulation testing where it will get called repeatedly.

There is also an intermediate solution, which may be considered since the cost of hardware is so cheap, and that s to "pre-blast" every combination into a memory-resident data array and keep it around as a static lookup table for each call. You could even save it to disk file and re-load it if you need to.

MathExtremist -

The way you describe it is pretty much how I check for winners.

Thanks, for the heads-up about the virus. I uploaded my side and now I get clean reports.

Quote: DJTeddyBear

There are 36 numbered naturals, plus the 2 that are either all 0 or all 00. That's 38 naturals.

If there are four 0 and one 00, there are 5 ways they can appear. Ditto for four 00 and one 0.

If there are three 0 and two 00, there are 10 ways they can appear. Ditto for three 00 and two 0.

That's 30 ways of having 5 greens that aren't either of the two naturals.

---

2 ^ 5 = 32. Subtract the two naturals, and you have 30 combinations of five spins with 0 and 00 results.

Okay, thinking about it some more, I was guesstimating that any one natural would be rarer than a 5-wild. But actually, there's 38 natural 5 of a Kinds, (38*1/38^5) which is bigger than the set of 5 wilds (1/19^5) (ignoring that 5 0's and 5 00's... ).

Anyways, as for the scanning the hands, the tips given are useful. It was -sort of- how I was going about it, but with less refinement.

I had an interesting weekend.

In just a couple hours, I modified my original program that cycles thru the 79 million combinations, to run thru 500 million random spins, analyze every spin along with the four preceding spins, and record the output.

I ran three separate trials of 500 million spins. That was actually the easy part. It took a little more work to get analysis in Excel.

I'll probably publish my data and Excel document in a couple days, but here's a preview of my results.


In addition to tracking the poker hand results, I tracked the spin result frequency. No number hit more or fewer than +/- 2.5 Standard Deviations from the average.

As expected, numbers ran hot and cold. I tracked the five longest durations between hits for each number. The average of the longest duration numbers did not hit was 631 spins. The longest of any of them was 763!


The Poker analysis:

500 Million - Data Set 1	Pay Table 1	Pay Table 2	Pay Table 3	Pay Table 4
Pairs Thru Quads	$ 371,194,221	$ 387,883,063	$ 390,588,993	$ 397,580,449
Standard Five Of A Kind	$ 31,598,100	$ 15,799,050	$ 14,001,800	$ 7,000,900
Standard Profit	$ 97,207,679	$ 96,317,887	$ 95,409,207	$ 95,418,651
Standard House Edge	19.44 %	19.26 %	19.08 %	19.08 %
Progressive Five Of A Kind	$ 34,080,000	$ 17,040,000	$ 13,816,000	$ 6,908,000
Progressive Profit	$ 94,725,779	$ 95,076,937	$ 95,595,007	$ 95,511,551
Progressive House Edge	18.95 %	19.02 %	19.12 %	19.10 %
Largest Progressive Pool	$ 18,065	$ 9,032	$ 7,828	$ 3,914
Largest Progressive Payment	$ 5,153	$ 2,577	$ 1,863	$ 932

EDIT - Days later...
I discovered that I used an old pay table for this. The actual house edge is about 16-17%. Correct data for multiple 500 spin simulations, will be on the website Simulations page.

One of the reasons I ran these tests was to see "clumping". I.E. Multiple consecutive hits of Five of a Kind. Yes, thanks to the wilds, there's plenty of that, but that's offset by occasional large gaps between hits. On this dataset, the there was at least one period of 140,536 spins with no Five of a Kind, wild or not.

Even so, as shown in this table, the progressive pool doesn't get outrageously large. And since the partial payouts reduce the prize pool, the 100% payouts were at points considerably lower than that.



But the biggest result I got was that my guesses at the progressive percentages, seed value and meter increments, produced results where the house edge was almost identical to the non-progressive edge.

I'm a happy camper about that!

Dave,
Wow, looks good.
What might be interesting is to see the algebraic spreadsheet versus simulation runs out of the 501,942 total possible combinations of 5 'hits' against 38 possible numbers, plus the repeating numbers and wild combos. A VERIFIED breakout that has all combos, and adds to 100%. something like:
HAND HIT: occurences P(hand)
5 of a kind full wild
5 of a kind full natural
5 of a kind "mixed"
full house wild
full house natural
flush wild
flush natural
straight wild
straight natural
...
trips (wild)
trips (natural)
two pairs wild
two pairs natural
one natural pair
nothing
TOTAL 501,497+ 100%

And see how much the simulations deviate from the algebraic theoretical norm (at 5K spins, 500K spins, 5M spins 50M spins, 500M spins)
You could tinker with the HE on the progressive via payouts (5 of a kind full wild, 5-OAK full natural, 5-OAK mixed, etc.), and get a higher resting jackpot level, etc.
Edit: You would need the verified game numbers in a report form from Mike, Stacy, Eliot, Charles Mousseau, etc., for a potential field trial, and for a higher priority with a distributor, to sign you up. A "Gaming ready" math report from a pro, not a DYI, as fine as it is, just says "I'm ready." A valid patent + a bona-fide math report = ready status. Granted, the electronics and programming are daunting to a distributor, but the math report says, "that too is doable."

Thanks.

That kind of stuff is in the Excel worksheets I used to generate the results posted above.

I'm gonna publish the data and Excel document in the next day or two.

The delay is because I want to create the Excel document that does a very good job of summarizing everything, while making it easy for the average Joe to use, as well as to insert one of several data sets that I've saved.

Dave - That would be great, and of great interest.
Some questions:
1. Will you have the theoretical "flat algebraic" spreadsheet done, to compare it to various simulations results, showing SD, comparison analysis, etc.
2. Would you consider ME to do the gaming report, or input to the base report format? He is so well-acquainted with this project, and is a bona-fide, it's an idea.

P.S.

I know I'll need a report from one of the people you named.

It's just that with the conversation I had with MathExtremist and DiskFlickr recently, I started to wonder about the numbers myself - particularly the numbers for the progressive payouts, and resulting house edge. I don't think I'm mistaked in my belief that the progressive edge can only be estimated after running a simulation.

A few months ago, my brother asked me what happens if it turns out my math is wrong. My answer was, how wrong could it be? Gut intuition suggests that the numbers I have on my math page for the odds of each hand seems to be in the ball park. If they're wrong, then I'll need to tweak the payouts a little to compensate. No big deal, really.

On the Progressive secion of my Payout Page, I state:

Quote:

When the Five of a Kind standard payments above are totaled, they are 6.36%, 3.18%, 2.81% are 1.40% of the total collections, respectively. It is therefore recommended that the Progressive Jackpot meter increase 6¢, 3¢, 2.6¢ and 1.3¢ respectively, whenever at least one person places the wager on the next five spins.

Because of the potential for the progressive jackpot to grow much higher than the set payout in the standard paytable above, the seed value is recommended to be one half of the Jackpot payout for the respective standard payout pay tables.

While having the increment be slightly lower than the collections shown seemed logical, having the seed be half was just a W.A.G. - Wild Ass Guess.

I am somewhat shocked and inspired that the progressive house edge ended up being so close to the non-progressive house edge.

Quote: DJTeddyBear

On the Progressive secion of my Payout Page, I state:While having the increment be slightly lower than the collections shown seemed logical, having the seed be half was just a W.A.G. - Wild Ass Guess.

I am somewhat shocked and inspired that the progressive house edge ended up being so close to the non-progressive house edge.

The HE on progressives are up to 50% (!) Shuffle's Pai Gow progressive on their Fortune game is that steep; I guess that 20% is doable, but a gaming math pro's input on that is needed. Steep progressives are begining to be viewed as bad values by players, especially if the bet just duplicates an existing bonus bet with a lower player return. But your progressive is a new thing on a totally new game, really.

What was the mean time to hit each of the various rewards, would be interesting... as would some sort of "streakiness" factor (I am sure there is a better statistical term for such a thing, but it eludes me).

Quote: thecesspit

What was the mean time to hit each of the various rewards, would be interesting... as would some sort of "streakiness" factor (I am sure there is a better statistical term for such a thing, but it eludes me).

Mean time, etc are part of my Excel document.

"Streakiness", or 'clumping' as I call it, is harder for me to figure out how to summarize. But it can be seen simply enough when looking at the worksheet detail. The only problem is that particular worksheet is long - it's time consuming to examine visually...

DJ

Its good to see you trying to work everything out before trying to sell it.

I would also suggest that in order to lessen the impact of those rare occurances of 7 7s in a row, that you simply SHARE all of the jackpots among all the players who win.

If for no other reason than to lessen the predictable variance in payouts for the house (in order to estimate profits), it makes sense. But you can also look at it as a way of ELIMINATING a HUGE unknown in your calculations... how many people bet on a jackpot at the same time, and won it. This becomes a non-issue if you share, and a huge issue if you don't



BTW, have you considered the impact this particular bet has upon the strategy used in the game of Roulette?

For example, we all were in on the conversation about making a lay bet on the upcomming number when a Fire bet in Craps is on the 5th or 6th number.

Have you considered situations where a player who made a Poker-For-Roulette wager can use the fact that the next 0 or 00 will pay, and he can use that fact in hedging strategies at that time?

I think its another reason to use computer simulations to show your prospective customer that you have already considered the impact of hedging in your analysis.


Good luck!

A shared payout is unnecessary.

The math supports paying winners what they are due. If the casino is concerned about a large payout, there are alternate paytables which provide a much smaller jackpot.

Plus, the casino always has the option of a maximum aggregate payout, which would be unrelated to the paytable or the rules of the game.

FYI: This has already been considered and included on my website.

---

Unless I missed the point of that Fire Bet Hedge thread, the hedge turns a potential large payment into a guaranteed small payment.

While I can see the casino favoring players who do hedge, I can't believe a casino would be stupid enough to count on such a strategy when evaluating the potential of any bet.

Therefore, to put any data about hedging into my analysis would be a weasel maneuver that I won't entertain.

Quote: DJTeddyBear

A shared payout is unnecessary.

The math supports paying winners what they are due. If the casino is concerned about a large payout, there are alternate paytables which provide a much smaller jackpot.

Plus, the casino always has the option of a maximum aggregate payout, which would be unrelated to the paytable or the rules of the game.

FYI: This has already been considered and included on my website.

Sorry I missed that on your web site, but if "A shared payout is unnecessary", why do you consider it on your web site and make it available as an option?

Quote: DJTeddyBear

Unless I missed the point of that Fire Bet Hedge thread, the hedge turns a potential large payment into a guaranteed small payment.

While I can see the casino favoring players who do hedge, I can't believe a casino would be stupid enough to count on such a strategy when evaluating the potential of any bet.

Therefore, to put any data about hedging into my analysis would be a weasel maneuver that I won't entertain.

The reason the Big Red payout is always the worst bet in the house at a whopping 16.67 house edge is to assure nobody can use it in hedging strategies. The house IS aware of the potential for griders to sit on such a strategy and essentially play all day without much risk.

That particular use of the FireBet as a hedge is not what I'm trying to point out here, DJ, its the fact that your Poker-For-Roulette wager (and the FireBet) are not the same as making a side bet in Carribian Stud because these bets introduce external factors into SUBSEQUENT play. As such, as these bets progress into their later phases, the expected payouts for certain combinations CAN BE USED TO THE ADVANTAGE OF THE PLAYER in hedging strategies, as I showed in the above example.

I can sit here and further demonstrate how your Poker-For-Roulette wager DECREASES THE HOUSE EDGE in several situations. If I were running the Casino, I would want to know EXACTLY how much.

Why you would choose to ignore this FACT, and only hope that your perspective customers don't figure it out puzzles me... I mean, why would you not at least give it a bit of thought, just in case they ask?


Does anyone else have an opinion on this? Can anyone else convince DJ to at least consider it?

The Poker-For-Roulette wager has a house edge as do the general roulette bets available. By combining the 2 you do not decrease the house edge of the regular bets. Players are in a position to 'hedge' only if their Poker-For-Roulette wager is showing +'ve ev - all they are doing is using some of that ev to provide a payback of some sorts regardless of how the bet ends.

In fact you could argue that the Poker-For-Roulette bet increases overall general roulette betting due to players wishing to hedge in winning situations.

Quote: discflicker

Sorry I missed that on your web site, but if "A shared payout is unnecessary", why do you consider it on your web site and make it available as an option?

Perhaps i worded it badly in my post above. I meant to say that I considered it, and put on the website why the jackpot should not be shared. It's in the Exposure section of my Casino Advice page:

Quote: Poker For Roulette - Casino Advice - Exposure

Because all participating players bet on the same outcome, and win at the same time, the jackpot payout is PER PLAYER, and not split among the players. Therefore, the progressive jackpot total should increment a set amount, regardless of the number of players participating, for every spin where at least one player is participating. As a consequence, there is a possibility of having more players in the game than average when the progressive jackpot hits. This can result in a payout that is greater than the amount collected. However, this is a relatively small risk, and is complemented by the opposite possibility - that of having fewer players participating than average when it hits. Still, because of this, it is recommended that the casino budget of a portion of the coin-in as a reserve towards this exposure.

Additionally, if aggregate maximum payouts are used, it is recommended that a separate aggregate maximum payout be established for this side bet, which is higher than the aggregate maximum for the other bets.

Quote: discflicker

I can sit here and further demonstrate how your Poker-For-Roulette wager DECREASES THE HOUSE EDGE in several situations. If I were running the Casino, I would want to know EXACTLY how much.

You might decrease the variance, or increase the expected value, but you have no effect on the house edge.

Or are you one of those people that thinks a free odds bet decreases the edge on the pass line bet?

Quote: discflicker

Why you would choose to ignore this FACT, and only hope that your perspective customers don't figure it out puzzles me... I mean, why would you not at least give it a bit of thought, just in case they ask?

I won't leave it to the casino to figure out. I expect it to be one of the math reports I'll get. But I also feel confident that it's going to confirm my position and not your claim.

Quote: discflicker

Does anyone else have an opinion on this? Can anyone else convince DJ to at least consider it?

Switch -
Thanks for backing me up on this.