Deconstructing the Atkins Diet Slot Machine
https://wizardofodds.com/games/slots/atkins-diet/
Does anyone know why the reels have a length of 32?
I ask because I'm building a simple 3 reel slot machine as part of my learning the C# programming language. I have a simple RNG and assigned values to reel positions, but I am working on the challenging part where I have to make the game somewhat challenging. Understandably, if you make each reel have more "rows" then you can increase the number of combinations, but why 32 in the case of the Atkins Diet slot game?
As an example of the extreme.... there's a slot machine out there called the Green Machine. You can probably YouTube some videos to see it in action.
But 3 "stops" show up on each real, and unlike most slot machines where you need to match 3 symbols on a payline, for this machine, whatever dollar amount shows up on the reel, is what you win.... and the reels are independent of each other. Top prize on each reel is $1000 on the $1 coin version of this machine.
So let's think about that for a minute.... If the casino wanted to offer 100% payback, and the $1000 symbol was the only symbol on the reel... the reel would have to be 3000 stops in length. Now consider that they want to make some money so let's say the payback is 90%. Then the reel has to be 3333 stops in length.
Now lets consider that there are other prizes and that the top prize probably doesn't return more than 5% of each dollar spent.... well then the reel length needs to be 3333/.05 = 66,660 stops long!!!!!!!
In some of the Wiz's analysis of other video slots, the reels within a game don't even need to be the same length.
Quote: rsactuaryfor a video slot machine, reel length can be whatever you want it to be.
As an example of the extreme.... there's a slot machine out there called the Green Machine. You can probably YouTube some videos to see it in action.
But 3 "stops" show up on each real, and unlike most slot machines where you need to match 3 symbols on a payline, for this machine, whatever dollar amount shows up on the reel, is what you win.... and the reels are independent of each other. Top prize on each reel is $1000 on the $1 coin version of this machine.
So let's think about that for a minute.... If the casino wanted to offer 100% payback, and the $1000 symbol was the only symbol on the reel... the reel would have to be 3000 stops in length. Now consider that they want to make some money so let's say the payback is 90%. Then the reel has to be 3333 stops in length.
Now lets consider that there are other prizes and that the top prize probably doesn't return more than 5% of each dollar spent.... well then the reel length needs to be 3333/.05 = 66,660 stops long!!!!!!!
OK wow that is a crazy example. I'll check the vids of it on YouTube.
Thanks for the quick answer!
Now this is pure conjecture, but for the Green Machine, it's entirely possible that the first reel is set up at 85%, the second reel at 87.5%, the third reel at 90%, fourth at 92.5% and fifth at 95% to mimic the better return for higher amount paid that slot machines tend to do.
No way to actually prove it without getting the PAR sheets, which typically aren't public.
Quote: DiscreteMaths2Check out this video on virtual reel mapping: https://www.youtube.com/watch?v=Ur53m9zQ5sY I believe in the Wizard's example he does 1 stop to 1 number and then the game relies on symbol distribution. With virtual reel mapping not only can you set weights for the symbols but you set weights for the stops, so your actual reel length is arbitrary.
this is true for mechanical reel slot machines. Not so for electronic wheels, which the Wiz assures us each symbol gets an equal weight.
Quote: rsactuarythis is true for mechanical reel slot machines. Not so for electronic wheels, which the Wiz assures us each symbol gets an equal weight.
Is that the law in Nevada ? For New Jersey I could find no such law you are just limited to "Slot machines shall not offer a play with odds greater than 100 million to 1" , " shall not permit a symbol above or below the top jackpot symbol to be mapped more than six times more frequently than any other symbol on a reel strip." and then just some more rules for bonus games, RTP, etc.
Quote: DiscreteMaths2Check out this video on virtual reel mapping: https://www.youtube.com/watch?v=Ur53m9zQ5sY I believe in the Wizard's example he does 1 stop to 1 number and then the game relies on symbol distribution. With virtual reel mapping not only can you set weights for the symbols but you set weights for the stops, so your actual reel length is arbitrary.
That is a very helpful video. I had a hunch it was built like that when I was looking at this research paper during my early stages of planning the program:
http://jgi.camh.net/doi/pdf/10.4309/jgi.2009.23.5
Page 93 shows how the RNG has a range of 1-72 while the reel itself is 1-22.
On my program, I initially designed the reels to have 6 rows, with symbols of A, B, C, 7, Cherry, and BAR. (note that I didn't think to include spaces but will now).
Winning combinations include matching 3 symbols or the single occurrence of a cherry.
Let's just say that my initial testing makes you win way too frequently. Since each reel has 6 possible stops, and a cherry is on each reel, you can win quite easily.
I now need to expand the size of my virtual reels and make the cherry less frequently hit by the RNG.
Quote: desertairIn this article written by the Wizard, he explains the entire process for creating a slot machine and describes the probabilities and such. Very helpful.
https://wizardofodds.com/games/slots/atkins-diet/
Does anyone know why the reels have a length of 32?
The "obvious" answer is, a power of 2 is used as most RNGs generate a number in the range of a power of 2 (if there are 5 reels, a number from 0 to 230-1 is generated; this is five 6-bit (i.e. 0 to 31) numbers, one for each reel). 32 might be chosen because it is the smallest power of 2 greater than 20, which I am led to believe is the "traditional" number of symbols on a "old style" slot machine; it could also be because, as there are 5 reels, 32 symbols per reel require 30-bit numbers, but 64 symbols would require 35-bit ones, which are larger than a 32-bit integer.
(Speaking of which, I was a little surprised to see "modern" slots for sale at a pinball show I went to near Sacramento recently, but then again, I definitely remember seeing video slots in Reno in 1985, and California has a "if it's at least 25 years old, it's an antique and you can sell it" law.)
Quote: rsactuarythis is true for mechanical reel slot machines. Not so for electronic wheels, which the Wiz assures us each symbol gets an equal weight.
I thought it was that multi-line slots had reels where each symbol had an equal weight (otherwise different lines would have different EVs), whether electronic or mechanical. Then again, do any one-line-only electronic (i.e. video) slots even exist, other than computer/online emulators of existing mechanical machines?
Quote: DiscreteMaths2Is that the law in Nevada ? For New Jersey I could find no such law you are just limited to "Slot machines shall not offer a play with odds greater than 100 million to 1"
This is not the first time I have heard this "law" mentioned here. The closest I can find to this in either the Revised Statutes or the Gaming Regulations is, a slot machine can have a jackpot with odds of 100 million-1 or higher of hitting it, but if it does, then the odds of hitting it must appear on the machine. Pardon my Latin, but while there may be a de facto (i.e for all intents and purposes) limit of 100 million to 1, it is not de jure (i.e. a legal requirement).
Yes, every multi-line game I've ever done (or seen par sheets for) has equally-weighted stops for exactly that reason. But some early single line video slots had weighted reels. Silicon Gaming had several: Arabian Riches, Banana-Rama, Silver Belle Express, Buccaneer's Gold, etc.Quote: ThatDonGuyI thought it was that multi-line slots had reels where each symbol had an equal weight (otherwise different lines would have different EVs), whether electronic or mechanical. Then again, do any one-line-only electronic (i.e. video) slots even exist, other than computer/online emulators of existing mechanical machines?
Quote: ThatDonGuyThis is not the first time I have heard this "law" mentioned here. The closest I can find to this in either the Revised Statutes or the Gaming Regulations is, a slot machine can have a jackpot with odds of 100 million-1 or higher of hitting it, but if it does, then the odds of hitting it must appear on the machine. Pardon my Latin, but while there may be a de facto (i.e for all intents and purposes) limit of 100 million to 1, it is not de jure (i.e. a legal requirement).
NJ: N.J.A.C. 13:69E Gaming Equipment 13:69E-1.28A Standards for the approval of a slot machine game "(c)
Slot machines shall not offer a play with odds greater than 100 million to 1"
For Nevada I cannot find such a law. I would assume if it exists it would be in Regulation 14
Manufacturers, Distributors, Operators of Inter–Casino Linked Systems, Gaming Devices, New Games, Inter–Casino Linked Systems, On-Line Slot Metering Systems, Cashless Wagering Systems, Mobile Gaming Systems, Interactive Gaming Systems and Associated Equipment; Independent Testing Laboratories and I can't find it.
Quote: DiscreteMaths2Quote: ThatDonGuyThis is not the first time I have heard this "law" mentioned here. The closest I can find to this in either the Revised Statutes or the Gaming Regulations is, a slot machine can have a jackpot with odds of 100 million-1 or higher of hitting it, but if it does, then the odds of hitting it must appear on the machine. Pardon my Latin, but while there may be a de facto (i.e for all intents and purposes) limit of 100 million to 1, it is not de jure (i.e. a legal requirement).
NJ: N.J.A.C. 13:69E Gaming Equipment 13:69E-1.28A Standards for the approval of a slot machine game "(c)
Slot machines shall not offer a play with odds greater than 100 million to 1"
For Nevada I cannot find such a law. I would assume if it exists it would be in Regulation 14
Manufacturers, Distributors, Operators of Inter–Casino Linked Systems, Gaming Devices, New Games, Inter–Casino Linked Systems, On-Line Slot Metering Systems, Cashless Wagering Systems, Mobile Gaming Systems, Interactive Gaming Systems and Associated Equipment; Independent Testing Laboratories and I can't find it.
I can't find it either now; maybe they ditched it. Nevada used to be a "de facto" 1 in 100 million. NJ and MS are 1 in 100M. Many Midwest riverboat states are 1 in 50M and GLI-11 recommends maxing out at 1 in 50M.
Quote: DiscreteMaths2Quote: ThatDonGuyThis is not the first time I have heard this "law" mentioned here. The closest I can find to this in either the Revised Statutes or the Gaming Regulations is, a slot machine can have a jackpot with odds of 100 million-1 or higher of hitting it, but if it does, then the odds of hitting it must appear on the machine. Pardon my Latin, but while there may be a de facto (i.e for all intents and purposes) limit of 100 million to 1, it is not de jure (i.e. a legal requirement).
For Nevada I cannot find such a law. I would assume if it exists it would be in Regulation 14
It's in the separate Technical Standards for Gaming Devices and Associated Equipment, Standard 2.070:
"If the odds of hitting any advertised jackpot that is offered by a gaming device exceeds 100 million to one, the odds of the advertised jackpot must be prominently displayed on the award glass or video display."
Odyssey multi game slot/VP machines were fairly innovative. IMO they were better than anything IGT had at the time. If I was going to purchase a slot machine I would probably go for one of them.Quote: MathExtremistYes, every multi-line game I've ever done (or seen par sheets for) has equally-weighted stops for exactly that reason. But some early single line video slots had weighted reels. Silicon Gaming had several: Arabian Riches, Banana-Rama, Silver Belle Express, Buccaneer's Gold, etc.
Did they get bought out by IGT?
Arizona Charlie's had a bank of .25 4play FPJW along with some drawings and other small promotions. That was a Fun play. Here's the machine in action including the infamous Vacation USA bonus slot. https://www.youtube.com/watch?v=ox8jk--dZaM
------------------------------------------------------
Does anyone remember Hot Reels multi-line slots you held symbols and drew. They were set over 100% and very fun as well. (they lacked a bonus round). Great for comps and mail, of course free play wasn't mainstream like it is nowadays. Usually you just got cash vouchers.
I'm not sure if Silicon made them or not. I can't find any information on them.
Quote: ThatDonGuyQuote: DiscreteMaths2Quote: ThatDonGuyThis is not the first time I have heard this "law" mentioned here. The closest I can find to this in either the Revised Statutes or the Gaming Regulations is, a slot machine can have a jackpot with odds of 100 million-1 or higher of hitting it, but if it does, then the odds of hitting it must appear on the machine. Pardon my Latin, but while there may be a de facto (i.e for all intents and purposes) limit of 100 million to 1, it is not de jure (i.e. a legal requirement).
For Nevada I cannot find such a law. I would assume if it exists it would be in Regulation 14
It's in the separate Technical Standards for Gaming Devices and Associated Equipment, Standard 2.070:
"If the odds of hitting any advertised jackpot that is offered by a gaming device exceeds 100 million to one, the odds of the advertised jackpot must be prominently displayed on the award glass or video display."
Good catch. Their organization is strange I didn't think to look in that standard.
Quote: DiscreteMaths2
Good catch. Their organization is strange I didn't think to look in that standard.
Neither did I, I knew I saw it somewhere before though. Thanks ThatDonGuy.
Quote: ThatDonGuyThe "obvious" answer is, a power of 2 is used as most RNGs generate a number in the range of a power of 2 (if there are 5 reels, a number from 0 to 230-1 is generated; this is five 6-bit (i.e. 0 to 31) numbers, one for each reel).
Correct. I picked 32 because it is a power of 2, which is also what random number generators are based on. Programming habit.
Quote: AxelWolfOdyssey multi game slot/VP machines were fairly innovative. IMO they were better than anything IGT had at the time. If I was going to purchase a slot machine I would probably go for one of them.
Did they get bought out by IGT?
I had a Silicon Gaming Odyssey machine but when I moved it was so heavy and awkward to move that I just gave it away. I kind of regret that now.
Yes, IGT did buy Silicon Gaming but they only did so to acquire some I.P. The company was already pretty much dead at that point.
I did some contract work for Silicon Gaming in the 1990's.
Robert Winter has some Odyssey info on his site:Quote: AxelWolfDoes anyone remember Hot Reels multi-line slots you held symbols and drew. They were set over 100% and very fun as well. (they lacked a bonus round).
I'm not sure if Silicon made them or not. I can't find any information on them.
http://www.robertwinter.com/slot/odyssey/
And yes, Silicon did make Hot Reels. Specifically, I did -- that was one of the games I designed. I had to integrate Excel with an external DLL too, that was a pain.
I don't recall doing a model over 100%, though it was a long time ago so I could be mistaken. We weren't trying to make a VP-like game, the goal was to have a typical slot hold. Many of the strategy plays were obvious, but not all of them, so I think I figured about 3-4% suboptimal play from the typical player. The true (optimal) hold was about 4% and therefore the expected hold was about 7-8%. Of course, it wasn't dealt out of a known deck like VP so you'd never be able to discern the optimal play without knowing the reels, and that requires either looking at the par sheet or reverse engineering. I recall it performed about where we wanted it to at least from a percentage standpoint. By late 1999 we were running into hardware issues with the CRTs so we had less win/day than we should have.
I did the initial models for 3 Reel Hold-Up too, and that one survived and still lives online at various IGT Interactive casinos.
Specifically, Silicon had just inked the deal for The Price Is Right, beating out IGT in the process. As I recall, IGT's offer to acquire SGIC came within 3 days. And you're right, the company was on its last legs. We had defaulted on our debt and the stock had cratered.Quote: DRichYes, IGT did buy Silicon Gaming but they only did so to acquire some I.P. The company was already pretty much dead at that point.
I did some contract work for Silicon Gaming in the 1990's.
Most of the company was spun off into WagerWorks and moved to San Francisco, then five years later WagerWorks was also acquired by IGT, turned into their online RMG group, and then rebranded as IGT Interactive.
Which games did you work on?
Quote: MathExtremist
Which games did you work on?
My work was on the back end helping to design the wide area progressive.
its been somany years now but I'm 99.9% sure on top of the games they advertised over 100% return. I think they said 101%/102%.Quote: MathExtremistRobert Winter has some Odyssey info on his site:
http://www.robertwinter.com/slot/odyssey/
And yes, Silicon did make Hot Reels. Specifically, I did -- that was one of the games I designed. I had to integrate Excel with an external DLL too, that was a pain.
I don't recall doing a model over 100%, though it was a long time ago so I could be mistaken. We weren't trying to make a VP-like game, the goal was to have a typical slot hold. Many of the strategy plays were obvious, but not all of them, so I think I figured about 3-4% suboptimal play from the typical player. The true (optimal) hold was about 4% and therefore the expected hold was about 7-8%. Of course, it wasn't dealt out of a known deck like VP so you'd never be able to discern the optimal play without knowing the reels, and that requires either looking at the par sheet or reverse engineering. I recall it performed about where we wanted it to at least from a percentage standpoint. By late 1999 we were running into hardware issues with the CRTs so we had less win/day than we should have.
I did the initial models for 3 Reel Hold-Up too, and that one survived and still lives online at various IGT Interactive casinos.
I know I had a strategy card for 2 different versions. Yellow and pink Version?
I remember comparing 2 strategies made by 2 different people and they were the same.
Played them at many places espesilly heavy at the Rio including points comps and mail. Circus Circus IIRC they had at least 1% cash plus comps I can't remember if they did much mail. The Hard Rock was paying double TOP LINE Slots only jackpots. I can't even remember what the cycle was or what it paied.
Even without all the extra stuff the over all results were good.
Eventully a few other AP's were hitting them hard as well.
It would make sense that they would set them that high concidering the average person would have a hard time playing them without tons of mistakes since it's not obvious like video poker. How the heck would someone know if they should hold a wild and a red seven or a flag. A dealt wild and mixed sparkplugs or just the wild?
Anyway, it seems to me that it wouldn't be much of a stretch to utilize a similar technique to write software that could very rapidly map out any video slot machine. Use either a smart phone camera or a hidden camera to snap photos of completed spins. You would basically just push the take photo button after each spin. The photo would be sent to an app on a phone, that would do the analysis and not only map the reels, but be able to tell you precisely when you had all symbols mapped so that you don't need to do even 1 extra spin.
Does something like this already exist? I might try this, but I don't want to reinvent the wheel.
sounds interesting and I would like to have it. Im not sure if getting caught with it would cause problems. I don't care if it's legal or not it's not worth getting 86ed over.Quote: cwazyA long time ago I wrote a bot for playing Bejewled for money at Worldwinner.com.. It was a relatively simple matter to have it classify the various symbols that appeared on the screen on the fly. It would essentially just take a screenshot, and analyze each symbol to see if it had already seen other symbols that were similar within a specified tolerance on the screen. If the symbol was outside of the similarity tolerance for any symbol already in the classification library, it was considered a new symbol. I got this down to 100% accuracy. Once all of the symbols on the screen were classified, it was a simple matter to figure out the best possible move, and then automatically move the mouse to make the move - all within milliseconds. The only issue with this was that Worldwinner quickly stops matching you with other players the more games you win, and eventually your tournaments simply never close and you stop making money.
Anyway, it seems to me that it wouldn't be much of a stretch to utilize a similar technique to write software that could very rapidly map out any video slot machine. Use either a smart phone camera or a hidden camera to snap photos of completed spins. You would basically just push the take photo button after each spin. The photo would be sent to an app on a phone, that would do the analysis and not only map the reels, but be able to tell you precisely when you had all symbols mapped so that you don't need to do even 1 extra spin.
Does something like this already exist? I might try this, but I don't want to reinvent the wheel.
Also....
The problem is...most slot machines deal with bonus rounds, they have all kinds of animations, many levels, screens, pick this pick that.. How is it going to calculate that? It may help cut down the work involved and the person can keep track of it separately.
If this is something that is mainly used on 3 reels regular slots like Blazing sevens well there's enough information regarding them slots. One just need to track the one they are playing to get a good idea and a hidden video of the game taken home and calculated will do just fine.
Very few people would be interested in this even if it worked on all slots. So there wouldn't be much profit. If you cage to much AP'so won't pay.
Quote: cwazyA long time ago I wrote a bot for playing Bejewled for money at Worldwinner.com.. It was a relatively simple matter to have it classify the various symbols that appeared on the screen on the fly. It would essentially just take a screenshot, and analyze each symbol to see if it had already seen other symbols that were similar within a specified tolerance on the screen. If the symbol was outside of the similarity tolerance for any symbol already in the classification library, it was considered a new symbol. I got this down to 100% accuracy. Once all of the symbols on the screen were classified, it was a simple matter to figure out the best possible move, and then automatically move the mouse to make the move - all within milliseconds. The only issue with this was that Worldwinner quickly stops matching you with other players the more games you win, and eventually your tournaments simply never close and you stop making money.
Anyway, it seems to me that it wouldn't be much of a stretch to utilize a similar technique to write software that could very rapidly map out any video slot machine. Use either a smart phone camera or a hidden camera to snap photos of completed spins. You would basically just push the take photo button after each spin. The photo would be sent to an app on a phone, that would do the analysis and not only map the reels, but be able to tell you precisely when you had all symbols mapped so that you don't need to do even 1 extra spin.
Does something like this already exist? I might try this, but I don't want to reinvent the wheel.
Would you need to take photos? If you had a camera that captured video, then just capture the wheels as they spin.
Of course, such a strategy has limited use with a single-line machine, as you can't determine how many virtual stops each symbol has.
Quote: ThatDonGuyWould you need to take photos? If you had a camera that captured video, then just capture the wheels as they spin.
Of course, such a strategy has limited use with a single-line machine, as you can't determine how many virtual stops each symbol has.
Even with video machines, the animations don't always match the actual reels. The ones that do use the actual reels will start spinning with the previous reel stops and then switch over to the reels leading up to the next reel stops. This sometimes leads to graphical glitches such as displaying two identical stacked symbols which don't ever appear stacked in that game.
Still, I have been able to determine the exact reels for video machines using high speed video.
There are probably hundreds of hours of video footage of slot gameplay already on YouTube. You don't need to play at all in order to map reels if someone else has already posted sufficient footage.Quote: cwazyDoes something like this already exist? I might try this, but I don't want to reinvent the wheel.
Quote: MathExtremistThere are probably hundreds of hours of video footage of slot gameplay already on YouTube. You don't need to play at all in order to map reels if someone else has already posted sufficient footage.
Except each game can have several different percentage returns.. which means different reels. So you may be trying to construct reels from two different games, essentially.
If you're trying to get an advantage on slot you should probably only play when you're almost certain you have an advantage. If I'm dealing with something that might be 3% or 4 % I'm probably not going to bother for long. I might play for a short while, but I'm not going to spend hours playing it.
