Calculating Slots Return Blind?

Hey there, I was just wondering what the best way of calculating a slots return rate without any given information, eg reel mappings. Do most casinos allow notes to be taken while you play (I think I remember reading some do on the Wizard of Odds site)? Is note taking the only way to accomplish the task, or are there any casinos in which asking for the mappings would work? XD I've just been thinking about this recently since it seems like it'd be fun to do. Any info would be great! Thanks!

Without the PARS sheet for the specific machine you're playing, no, there's no real way to calculate the return. You can approximate it, though, by taking notes and working from there. There's an easy way, a hard way, and a not-quite-so-hard way.

The easy way is finding a bank of slots that are marked "X% RETURN". IANAL, but I'm pretty sure that legally, all slots in that bank have to be set to have that return.

The hard way is reconstructing the PARS sheet. Over time, you can reconstruct an approximation of the reel striping and the probability of each symbol (or blank) popping up. Using the paytable (which you're given), you can approximate the return.

If you don't feel like taking the time to reconstruct the PARS, the not-quite-so-hard way is to simply take the winnings from each spin in a list (0,0,0,3,0,6,0,0,0,15,etc.), add them up, and divide by the coins wagered (usually 2 or 3 per spin).

One downside of the hard ways is that to get any kind of accuracy, you need a lot of spins. Hundreds is OK, thousands would be better. The Wizard did almost 4000 to clock one machine. With a high variance machine (usually they have a large jackpot payout), you would need a lot more spins to minimize variance.

The other downside is that now you have a return that only reliably works on the machine you clocked. A given style of machine could have many different PARS sheets, each giving different returns, even with the same paytable.

As far as taking notes goes, I've done it before with video poker, and I've never had any problems with the casino staff. Nosy neighbors are another story.

I agree with the above post. For a video slot you pretty much have to take a tally of the frequency of each symbol on the reels, and then do the math. The bonus should also be straightforward, where you know exactly the rules and possible outcomes, like a pick 'em bonus that always has the same prize distribution. All things considered it will take about 8 hours, and the return will only be valid for that one machine.

Taking notes should not be a problem. The only time I had a problem was in taking notes at a Fortune Cookie machine at the Suncoast, when three guys in suits approached me and said I was not allowed to play and write at the same time, and if I did it again they would have security escort me to the door.

I am not a slots player, and I certainly don't know how the machines actually work, so maybe this is a good thread in which to ask for a little education.

I had read the linked article by the Wizard when it was published and didn't understand whether the reel position weighting he described was the way machines actually functioned or just a framework for interpreting results. I had this concept in my head that the older (antique?) machines may have had mechanical devices (perhaps with physical weights or different width slots for a pin to drop into or something) to stop the wheels, with the weights and widths giving the casino's preferred payout.

My naive impression had been that the modern machines were just electronic chips programmed with a payout pattern (frequency of each type of payout) that was constantly "spinning" equivalent to a roulette wheel, with the pressing of the button or pulling of the handle providing an interrupt like the ball dropping. Then, based on the payout indicated by the "ball position" the reels were just stopped electronically in some position that would provide that payout.

Even though I don't play the machines, I am curious as to how they actually function, both the old ones and the modern ones. Can someone educate me briefly or point me to the right spot? Thanks.

Quote: Doc

I am not a slots player, and I certainly don't know how the machines actually work, so maybe this is a good thread in which to ask for a little education.

I had read the linked article by the Wizard when it was published and didn't understand whether the reel position weighting he described was the way machines actually functioned or just a framework for interpreting results. I had this concept in my head that the older (antique?) machines may have had mechanical devices (perhaps with physical weights or different width slots for a pin to drop into or something) to stop the wheels, with the weights and widths giving the casino's preferred payout.

My naive impression had been that the modern machines were just electronic chips programmed with a payout pattern (frequency of each type of payout) that was constantly "spinning" equivalent to a roulette wheel, with the pressing of the button or pulling of the handle providing an interrupt like the ball dropping. Then, based on the payout indicated by the "ball position" the reels were just stopped electronically in some position that would provide that payout.

Even though I don't play the machines, I am curious as to how they actually function, both the old ones and the modern ones. Can someone educate me briefly or point me to the right spot? Thanks.

I echo your sentiments exactly. I didn't know you could tell something by how many times a symbol came up. I always compared it to a scratch off lottery ticket; You get the 2 gold bars way more often than normal distribution would suggest, but the 3 gold bars is the jackpot...

Quote: wildqat

*snip*

If you don't feel like taking the time to reconstruct the PARS, the not-quite-so-hard way is to simply take the winnings from each spin in a list (0,0,0,3,0,6,0,0,0,15,etc.), add them up, and divide by the coins wagered (usually 2 or 3 per spin).

One downside of the hard ways is that to get any kind of accuracy, you need a lot of spins. Hundreds is OK, thousands would be better. The Wizard did almost 4000 to clock one machine. With a high variance machine (usually they have a large jackpot payout), you would need a lot more spins to minimize variance.

*snip*

Reconstructing the PARS would be one of the things that would make it fun! I don't mind having to play a lot, I tend to end up in casinos often as of late, so I thought doing this as an experiment would just keep me more entertained than merely pressing a button. :P

Also, one of my favorite types of slots is like All That Glitters (That was the only name of one I could remember). It seems like it'd be simple to figure out the probabilities for each symbol to drop. I think I'll start with that kind.

Thanks!

Quote: Doc

Even though I don't play the machines, I am curious as to how they actually function, both the old ones and the modern ones. Can someone educate me briefly or point me to the right spot? Thanks.

I think it is easiest to explain how video slots work, like my Atkins Diet slot. I try to explain how it is programmed here.

Quote: Wizard

I think it is easiest to explain how video slots work, like my Atkins Diet slot. I try to explain how it is programmed here.

Thanks! That is a good explanation.

My thinking was that if the payout were going to be determined by five random number generators and five strips, why not just use one random number generator and a function to give the HA and variance you think is appropriate? Then just display an array of graphics that gives that payout. Would that violate some regulation or some player expectation? With a video game, it seems that you could generate whatever array of icons you want after determining the necessary payout for each spin. Mechanical reels would just have to be stopped intentionally at the necessary positions to give the appropriate payout.

Guess maybe my thinking is incomplete because I don't play the machines. I know a lot of people who really enjoy them, but I have never drawn any entertainment from them, and entertainment is about the only reason I participate in gambling activities (not quite foolish enough to believe that I am going to win in the long term.)

Quote: Wizard

The only time I had a problem was in taking notes at a Fortune Cookie machine at the Suncoast, when three guys in suits approached me and said I was not allowed to play and write at the same time, and if I did it again they would have security escort me to the door.

Their paranoia was irrational, right? It's IMPOSSIBLE to cheat at slots by taking notes. Right? If casinos are ok with baccarat and roulette players taking notes, why not slot players?

Quote: Doc

Would that violate some regulation or some player expectation? With a video game, it seems that you could generate whatever array of icons you want after determining the necessary payout for each spin. Mechanical reels would just have to be stopped intentionally at the necessary positions to give the appropriate payout.

I think in Nevada the outcome of each reel has to be independent. In the late 90's I played a game in Atlantic City where the reels were not independent, that probably worked as you described, choosing an outcome first, and then finding a way to stop the reels to match it. That is also how class II slots work, choosing random bingo balls, scoring the patterns, and then finding a way to stop the reels to match the win from the bingo balls.

Quote: reno

Their paranoia was irrational, right? It's IMPOSSIBLE to cheat at slots by taking notes. Right? If casinos are ok with baccarat and roulette players taking notes, why not slot players?

Right and right. You'd have to ask them why they did it. The head engineer at IGT could not beat his own slot machines, except if a progressive machine was unusually high. They made a remark to me that I wasn't playing the machines in the way they were meant to be played.

Sometimes it seems that casino operators can be just as superstitious as players. They love and encourage people taking notes at the craps and baccarat table, because this simply means that the player is pursuing some kind of betting strategy which will keep them at the table longer. But they get upset at people taking notes at a slot machine, when presumably the result will be the same.

I suspect that the Wizard gets more grief because people heard about his survey, but they haven't taken the time to read his method of operation. They just don't want him taking notes.

I think that server based machines will allow the casinos to reveal the settings. The mystery setting will be replaced with progressive increases in the payback. For example you may sit down at a machine and it reads 83%. But after five minutes of play it is readjusted to 90%. Then it begins to slowly increment by a 1% every 15 minutes at first, and then every hour. The player is rewarded for his longevity. Each increase in the percentage will be accompanied by some change in the paytables (most often on the jackpot setting).

I think that this type of system will remove widespread fears that machines have been tightened up following the recession. It will also remove fears that the casinos will misuse the server settings to screw players.

It will also reward longevity in play, with more players trying to crank their machine up to a high percentage. I hope that it doesn't involve the increased used of adult diapers (but anything is possible).

It may have to involve the use of loyalty cards and an upper time limit of 4 hours. Otherwise there would be slot teams that would try and keep a machine isolated and running for weeks on end to try and keep the house average as low as possible. The players would think in terms of "sessions" and the reward would be very high returns for the last hour.

Quote: pacomartin

It may have to involve the use of loyalty cards and an upper time limit of 8 hours. Otherwise there would be slot teams that would try and keep a machine isolated and running for weeks on end to try and keep the house average as low as possible.

As long as it never went above, say 98%, they'd encourage people to play in teams, and continue to lose money on a continuous basis. Bac is less than 2% house edge, and I doubt they mind people playing that for hours on end.

Quote: pacomartin

Sometimes it seems that casino operators can be just as superstitious as players. They love and encourage people taking notes at the craps and baccarat table, because this simply means that the player is pursuing some kind of betting strategy which will keep them at the table longer. But they get upset at people taking notes at a slot machine, when presumably the result will be the same.

I'd be happy if they let you take notes at the blackjack table.

Quote: pacomartin

I think that server based machines will allow the casinos to reveal the settings. The mystery setting will be replaced with progressive increases in the payback. For example you may sit down at a machine and it reads 83%. But after five minutes of play it is readjusted to 90%. Then it begins to slowly increment by a 1% every 15 minutes at first, and then every hour. The player is rewarded for his longevity. Each increase in the percentage will be accompanied by some change in the paytables (most often on the jackpot setting).

As the law stands, this would most likely be illegal, even if it's just twiddling with the jackpot. Server based machines have to be idle for at least four minutes before the payout is adjusted. Even if the pay table isn't changed, and the jackpot goes up to twiddle the return, some enterprising lawyer will figure out a way to successfully argue that they're a type of single-machine progressive, and there are other laws in place that prohibit progressives from just disappearing, as they would likely do once the machine goes idle.

Quote: wildqat

The other downside is that now you have a return that only reliably works on the machine you clocked. A given style of machine could have many different PARS sheets, each giving different returns, even with the same paytable.

Another problem is that the machine you clocked on your last visit could have been changed before your next. How would you know unless you happened to see the machine open? While this requires paperwork, it is not uncommon for casino management to make adjustments from time to time. Also, how often have you arrived at your favorite casino, leave alone the ones that you visit less frequently, to find that machines have been rearranged or removed. If this change affects the machines like the one that you've clocked, how would you know which is which?

Quote: wildqat

As far as taking notes goes, I've done it before with video poker, and I've never had any problems with the casino staff. Nosy neighbors are another story.

For the most part, I haven't had a problem with taking notes or using guides. Usually, I am not even approached. When I have been approached, it has always been for writing, not reading, at a machine. Usually, when I tell the management or security personnel who do inquire that I am writing down paytables for my own personal use, they are satisfied and leave me to it.

Sometimes, the circumstances do not make a difference. If a casino management wishes to bring heat, they can find just about any reason.

At Casino Fandango in Carson City, I was asked by security not to write down the paytables of Multi-Strike Deluxe, as it violated copyright laws and exposed the casino to liability.

More recently at Baldini's in Sparks, I was asked why I was only writing down progressives (I was, in fact, focusing on paytables for VP progressives). I replied that I prefer to find the best game to play in order to "lose less" so that I may "play longer." Then, when I was asked if I work for [another local casino], I realized that they were concerned that a competitor was scouting their goods. I thought it was silly that they thought someone would be wearing a T-shirt from the casino while scouting, especially one that advertised a beer event that they put on, but far be it from me to question the logic of casino management! I assured the suit in the group (I had been approached by a suit and three or four gaurds) that I am doing this for my own use and was asked not to continue, to which I complied. Later, I was told by the same suit that it was OK to proceed, and once again that it was not OK. After leaving my paper & pen in the car and returning, I was eventually asked (politely) by a different security guard to "call it a night," as "writing down progressive paytables is [prohibited by] Gaming." Throughout this process the staff was polite and, as far as I could tell, I was cooperative, even to the point of letting them take a photocopy of my DL, as I was informed that I may return. This was about six months ago, and I have not been back, only because this casino is not close to others that I frequent.

In both cases, justified or not, I complied, as arguing with them in their house is probably not going to be viewed favorably. Besides, I was outnumbered! lol

I cannot say what kind of heat you might find by jotting down every outcome of a series of plays. On one hand, you're playing, so you're a "paying customer." On the other hand, writing for a prolonged period could bring unwanted attention - I've been "ratted out" by a waitress after her third unsuccessful visit while I was trying to clock multiple meters on a VP progressive. Perhaps ordering something and tipping would have satisfied her, but it would have also broken my concentration!

Quote: Doc

My naive impression had been that the modern machines were just electronic chips programmed with a payout pattern (frequency of each type of payout) that was constantly "spinning" equivalent to a roulette wheel, with the pressing of the button or pulling of the handle providing an interrupt like the ball dropping.

Continuous 'spinning' is where I have little trouble. The old 3-reel machines had programs on e-proms, and CPU just kept running it. Due to memory, and other limitations that seemed plausible (and all the how the slot machines work articles point to that). Now the new CPU's are very powerful, no memory limitations. So does the old theory of continuous spinning still hold true? I question that because many new machines are muti-denomination as well as multi title machines. It would be much easier to have a built in RNG or the spinner for each title then to fetch the numbers from a single RNG algorithm.

Quote: Hyperlisk

Hey there, I was just wondering what the best way of calculating a slots return rate without any given information, eg reel mappings. Do most casinos allow notes to be taken while you play (I think I remember reading some do on the Wizard of Odds site)? Is note taking the only way to accomplish the task, or are there any casinos in which asking for the mappings would work? XD I've just been thinking about this recently since it seems like it'd be fun to do. Any info would be great! Thanks!

The wizard calculates the odds of a slot machine blind with only 4000 plays. He deduces that it pays back 95.53% with 1 coin and 95.66% with 2 coins. He explains in detail the calculation. I don't know if you understand virtual reels or not, but it would be a fairly simple calculation without virtual reels. But there are no machines without virtual reels anymore.

In his survey of the casinos of Vegas done 12 years ago, he had the benefit of being given PARS sheets. These sheets show you the possible return settings, and he was able to write a computer program which would tell him what the setting the casino had chosen for that machine.

A particular brand of machine has typically 8 possible settings for returns. The casino selects their setting by inserting the appropriate computer chip. Usually the settings are something like 85%, 87.5%, 90%, 92.5%, 94%, 95%, 96%, 97.5% or something similar. Out of 71 casinos, and dozens of machines in each casino, he only ever found a single machine set for the loosest possible setting. He suspects that it was in error.

Casinos that don't expect repeat customers like the Airport, or big flashy strip casinos tend to use the tightest possible setting. Locals casinos with repeat customers tend to use a much looser setting (but with only 1 exception, never the loosest possible).

Quote: pacomartin

A particular brand of machine has typically 8 possible settings for returns. The casino selects their setting by inserting the appropriate computer chip. Usually the settings are something like 85%, 87.5%, 90%, 92.5%, 94%, 95%, 96%, 97.5% or something similar. Out of 71 casinos, and dozens of machines in each casino, he only ever found a single machine set for the loosest possible setting. He suspects that it was in error.

Casinos that don't expect repeat customers like the Airport, or big flashy strip casinos tend to use the tightest possible setting. Locals casinos with repeat customers tend to use a much looser setting (but with only 1 exception, never the loosest possible).

You must be referring to IGT 3 reel games with the approximate payback percentages....Would the strip properties have them set at the lowest rates (i.e. highest hold)? I would hope that the 25 cent ones are set at the 90% or better amounts with the amount of people that visit the strip each year....

Quote: Wizard

Taking notes should not be a problem. The only time I had a problem was in taking notes at a Fortune Cookie machine at the Suncoast, when three guys in suits approached me and said I was not allowed to play and write at the same time, and if I did it again they would have security escort me to the door.

I've also been tossed for note-taking, but it was not atg the Suncoast or anywhere in Vegas. They sent me away and told me not to come back for 24 hours, and if I did it again, then I'd be tossed for a year.

Quote: Wizard

Right and right. You'd have to ask them why they did it. The head engineer at IGT could not beat his own slot machines, except if a progressive machine was unusually high. They made a remark to me that I wasn't playing the machines in the way they were meant to be played.

The comment to me was that I was stealing proprietary information, or something along those lines, I don't recall the exact verbiage.

Quote: IAchance5

You must be referring to IGT 3 reel games with the approximate payback percentages....Would the strip properties have them set at the lowest rates (i.e. highest hold)? I would hope that the 25 cent ones are set at the 90% or better amounts with the amount of people that visit the strip each year....

The payout percentages are not approximate, they are based on simple mathematics detailed in the PARS. Now they must be played for long enough so that the law of large numbers takes over, and the actual return approaches their theoretical payout.

The Strip averages 91.37% payback on their 25 cent machines over the last 12 months. I doubt very much that every single machine is set to return over 90%, although nothing can be determined absolutely from a mean. I suspect that almost every machine made by the big manufacturers is a minimum of 85% payback. Machines that pay lower than that are annoying to players, and are of limited value to slot owners.

From the Wizard's survey, we can guess that every machine he studied was set to the lowest possible payback at the airport. His results imply that the Venetian also used lowest settings (although he has never actually stated that observation.

Hi!

Maybe I do not understand that, the point of doing the calculation return simply to learn how to set it up a casino?

- In Europe, if a player will ask the casino manager, will answer what percentage of the automaton.

Or at different times, the gaming machine may have different percentages of return?

Quote: Dobrij

Maybe I do not understand ..........
Or at different times, the gaming machine may have different percentages of return?

Hi, I know I don't really understand either.
Most machines in Vegas are at a certain fixed but undisclosed payout. Sure signage may give some information particularly as to minimum payout but players want to know actual payout of the machine. This is expressed as a percentage of coin-in but does not mean that actual coin-in affects the results. It simply is an indication of what will be likely to happen over a long-term trial.

Only a relatively few machines, known as Server machines, may be changed from time to time by the casino without physical changing of a computer chip. These machines, by law, must be idle for "x" minutes prior to any such change being implemented via computer.

The one really puzzling thing to me about this is that it appears to be similar to asking for a woman's "vital statistics" before going out on a date with her when what you are really interested in is what will be happening at the end of the date. Having precise measurements of her dimensions is only thought to be a guide to her likely decisions for that night. However, its well known that if she discloses her intentions prematurely, its likely she won't get taken out for the dinner and drinks at all.

If payout percentages were available for a very short term time period... either the players would walk right on by or there would be a race between the players and the casino manager who wanted to pull the plug out of the socket.

If the payout is truly determined by random and independet reel positions, doesn't that mean that return percentage cannot be independent from lines played?

Example: The highest pay symbol appears once per reel. If you hit five in a row on line one, that makes it impossible for all of the other lines to hit the same 5-of-a-kind.

Can somebody shed some light on this? Does that mean for max return I should play just one line?

As far as anyone relying a machine needing to be idle for 4 minutes, I believe it can be forced idle unwillingly by locking the player out. I've seen machines "disabled by host" on numerous occasions and think this would qualify as being idle.

Quote: FleaStiff

If payout percentages were available for a very short term time period... either the players would walk right on by or there would be a race between the players and the casino manager who wanted to pull the plug out of the socket.

There is every reason to believe that on some machines it is. A few years ago I thought it would be nonsense.

Quote: Canyonero

Can somebody shed some light on this? Does that mean for max return I should play just one line?

This is also a very interesting question ! By the way the game with a fixed number of lines (multi-game machines) do have a higher rate of return

Quote: Canyonero

Can somebody shed some light on this? Does that mean for max return I should play just one line?

No, the return is the same regardless of the number of lines you play. I'm struggling for a good way to cleanly explain this though.

Quote: tringlomane

No, the return is the same regardless of the number of lines you play. I'm struggling for a good way to cleanly explain this though.

Still struggling? In the meantime I have convinced myself that you are right, but have failed to come up with a mathematical proof. The convincing I did by brute force though, and I am fully aware that this proves nothing. Anyway behold my E-Z-Slotz (tm), patent pending: (sorry, the formatting got totally messed up, I hope it is still readable)

E-Z-Slotz™
Reels: Reel 1 Reel 2 (identical) (this is the full reel layout of my two reel machine, with 4 symbols each)
x x
o o
o o
o o

Paytable:

x x 20

o o 1

possible combinations (of the whole playing field): (there are 16 possible combinations that can appear on screen after a round of play, doubles included, so each one below is equally likely)


Payouts: Line1 Line2 Line3 Line4
x x 20 1 0 0
o o

x o 0 1 0 1
o o

x o 0 1 0 1
o o

x o 0 0 20 1
o x


o x 0 1 1 0
o o

o o 1 1 1 1
o o

o o 1 1 1 1
o o

o o 1 0 0 1
o x


o x 0 1 1 0
o o

o o 1 1 1 1
o o

o o 1 1 1 1
o o

o o 1 0 0 1
o x


o x 0 0 1 20
x o

o o 1 0 1 0
x o

o o 1 0 1 0
x o

o o 1 0 20 0
x x

The (statistical, theoretical) sum of the payouts after 16 rounds of play for one line played and four lines played:

total payouts 4 lines: 116
total payouts line 1: 29

116/29=4 q.e.d. - the return is indeed the same for one and four lines.

Again, this was just done to illustrate the concept for myself, I am aware that this doesn't prove anything conclusively for multi line slots in general.


[Just for fun: You have to bet 2 units per line to play, so the return of this machine is 29 / 2*16 = 90.6%]

Coming back to the original question:

Is the number of symbols on any (video!) reel generally known? Are we talking 20ish or 200ish? Having a high number of symbols would make it virtually impossible to calculate the return through observation. (Not to mention the bonus rounds, which will destroy this effort for most recent slots.)

Quote: Canyonero

Coming back to the original question:

Is the number of symbols on any (video!) reel generally known? Are we talking 20ish or 200ish? Having a high number of symbols would make it virtually impossible to calculate the return through observation. (Not to mention the bonus rounds, which will destroy this effort for most recent slots.)

Nice work on proving by example!

As for this though, not really, it depends on the machines. Older games tend to have a bit less I think than newer games. Some of the ones I tested from my IGT PC games. Payouts ridiculously high since it's a PC game, not a machine in the casino.

Texas Tea: 53/53/53/53/53 - Payback ~ 97.36% Hit Frequency ~ 9.83%
South Pacific: 46/46/40/41/38 - Payback ~ 90.09%
Twin Win: 76/70/55/82/68 - Payback ~ 96.58% Hit Frequency ~ 4.04%
Festival Fantastico: 55/55/55/55/58 - Payback ~ 96.12% Hit Frequency ~ 8.12%


Texas Tina: 35/35/35/35/35 - Payback ~ 95.67% Hit Frequency ~ 16.9% (Similar math model to "Money Storm" discussed in Harrigan/Dixon; if math is exactly the same, this is likely the 96% version, and I misestimated a bonus round and/or screwed up the math)
Lobstermania: 46/47/47/51/50 - Payback 91.77% Hit Frequency ~ 2.18% (Also in Harrigan/Dixon, which means I may have goofed this up as well since they don't quite match. Actually they are probably different, this is the 25-line version of the game.).
http://www.nh.gov/gsc/calendar/documents/20091117_harrigan_dixon.pdf

I do remember some games many years ago that had a higher payback percentage by playing less lines. Playing one line was the best if I remember correctly.

I believe it was just a design flaw because scatter pays were paying multiplied by the line bet instead of the total bet.

Quote: tringlomane

Nice work on proving by example!

As for this though, not really, it depends on the machines. Older games tend to have a bit less I think than newer games. Some of the ones I tested from my IGT PC games. Payouts ridiculously high since it's a PC game, not a machine in the casino.

Texas Tea: 53/53/53/53/53 - Payback ~ 97.36% Hit Frequency ~ 9.83%
South Pacific: 46/46/40/41/38 - Payback ~ 90.09%
Twin Win: 76/70/55/82/68 - Payback ~ 96.58% Hit Frequency ~ 4.04%
Festival Fantastico: 55/55/55/55/58 - Payback ~ 96.12% Hit Frequency ~ 8.12%


Texas Tina: 35/35/35/35/35 - Payback ~ 95.67% Hit Frequency ~ 16.9% (Similar math model to "Money Storm" discussed in Harrigan/Dixon; if math is exactly the same, this is likely the 96% version, and I misestimated a bonus round and/or screwed up the math)
Lobstermania: 46/47/47/51/50 - Payback 91.77% Hit Frequency ~ 2.18% (Also in Harrigan/Dixon, which means I may have goofed this up as well since they don't quite match. Actually they are probably different, this is the 25-line version of the game.).
http://www.nh.gov/gsc/calendar/documents/20091117_harrigan_dixon.pdf

Thx a lot for the data and the link, some interesting stuff inside!

I want to clarify, calculate the percentage of return, perhaps on:

- From any manufacturer?

- Which requires technical characteristics of the game that it was possible?