Biased Dice: The science begins

After years of scoffing at the biased dice conspiracy theorists and having everyone weigh in their opinions, with the Wizard's nod of approval, this thread is specifically to begin a scientific study into the fairness of actual casino dice in play in actual casinos.

I am agreeing to keep things factual, and I invite others to participate in this study. Profiling the outcome of dice in real casino conditions is actually quite simple. There are multiple ways to do this, and I encourage as many people as possible to participate in this scientific study into the fairness of dice. To kick things off let me give you the methods that I have for collecting samples of dice outcomes in real casino environments:

Method 1: count using chips on the rail.

In the first method, you maintain seven stacks of chips. The first six represent the counts of faces one through six. The seventh is a common count of the other six stacks. So when all stacks have at least 5 in the count, you reduce the first six stacks by five units and add five units to the seventh common stack. For the common stack, you can also use other things besides chips to represent hundreds (like a penny).

Using this method, the most chips you will generally need are 4x6 white chips 4x6 red chips and 3x6 green chips. Using pennies instead of black chips takes care of needing an excessive amount of buy-in dedicated to counting. So generally get $25 in white, $125 in red, and usually you won't need as many greens as the green levels stay more close to one another, so you can typically get by $200 or even less in green chips. If you are unable to afford a $300 buy-in, you can use coins, but buying in for $400 to $500 and playing with your chips arouses less concern among the casino than putting non-chip items in the rail (or at least excessive non-chip items). You can also just use memory or written notes for black chips and of course once you get to the point you need black chips you have been there all day long anyway.

Method 2: Bank Tally Counters



The second method is to buy a bank tally counter and just press a button for each face count. If you decide to help out often enough, this is a much less expensive and easier way to keep tallys as you don't have to take as much time doing it and you can do it more often without getting confused with moving chips around if you also want to play the game. I have only used the tally counter once. It was at Gold Coast, and the crew was very interested and it turned out to be a lot of fun. Before I used the tally counter the first time I thought the casino would have more of a problem with it, but they were very entertained by the whole idea. They also use the same tally counters (but just one not six) to count the drops for the fire bet over there.

If you want one of these Tally Counters, I recommend that you make a custom mount and it fits very nicely right into the rail. It only took me about ten minutes to make my mount out of aluminum from Home Depot, and I can make one for you and ship it to you for free if you want to use this method in exchange for your help in counting.

Why so much counting?

This is a labor intensive process. We could argue all day about all the things that could affect the outcomes to be skewed unrealistically, but what we are looking for is skewed face results that do not fit fair distributions. We want counts from as many different casinos and days of the week and other factors as possible. I'm sure there will be lots of questions and suggestions, but the details of which dice are selected and other factors about how long the dice have been in play and so on even if they were suspect, we don't care anything except actual face outcomes from the dice. And the more the better. This means that if you have a die-down, you count the face of the die that did resolve legally. You don't count dice that go in the bowl, in the bank, etc, that lead to a no roll, but if you see one of the two dice resolve legally, you count it.

What are the possible results?

I believe that everyone on this forum (including me in fact) believe that we will have evenly distributed faces when all the data is combined together for a long enough period of time. But there is a possibility that we will find that some faces come up more often than what could be explained by pure chance.

What else should we pay attention to?

Any detail of the dice that you are recording are important. With enough training you can recognize the exact color value of specific manufacturer's dice. There are subtle details to dice that give away who makes the dice. Also you can ask who makes the dice and places will tell you. Building a database of what casinos are using what dice is part of what we would like to know when profiling dice, but also if it is possible to know for sure who made the dice you are profiling, that is helpful. Also making a note of the serial number and the font used for the serial number and any other details that may seem unimportant can be helpful in associating any unusual outcomes with a particular manufacturer or other subtlety to a casino die that may have not been considered possible to lead to bias.

I will start a blog with face counts and hopefully we can get to enough face counts to be meaningful. If you are willing to play at the casino while doing your face counts, I recommend betting the don't pass with whatever amount of odds you want to gamble and no other bets to get your cost per roll at $0.02 per roll. Throw the dice as randomly and quickly as possible and encourage the crew to keep the dice moving. An empty table is your friend as you want as many outcomes in as short a period of time as possible paying $0.02 per sample to roll them if you have to or finding some other shooter willing to randomly shoot the dice for you and you can just watch and click. If you record outcomes, using random shooters is ideal. Even though we have no firm proof that DI is possible, there's no reason to get into a conversation about whether the skew was DI or biased dice, so just don't record if someone looks to have a hope of controlling the dice. You want to record outcomes of random shots that hit the back wall .. IE: the same throws the casino likes to see are the ones you want to record: pick'em up and shoot-em!

I think we went through this before: counting even if you counted millions of rolls would not stand up in court. It would just be circumstantial evidence. To prove a bias you would somehow have to examine and prove either by weight or size or construction that the dice were not manufactured to have an even distribution. This could only be done with dice that have not yet been rolled or used in a game.

Any attempt to indicate that any sort of "count" showed any bias and the "defense" would say "dice is a random game and any number of faces could show at any time." Further the dice makers would blame the tables for impeding a smooth and even roll of the dice -- unlike dice on a spinning device.

In short, give it up already. OR find some way to dissect dice to determine how each section of the die weighs and what its dimensions are.

Quote: AlanMendelson

In short, give it up already

Geez, Alan, how about just blocking the thread?

block the thread and just let the nonsense run rampant? No.

Look if you want to do something scientific then do something scientific. tracking rolls of the dice wont cut it.

Quote: AlanMendelson

I think we went through this before: counting even if you counted millions of rolls would not stand up in court.

Yes. We went through this before.

Even if were so that an obvious bias (reflected by a large sample) "wouldn't stand up in court," it doesn't matter, because:
1. conclusions of courts often have no basis in scientific reason nor in reality;
2. no one is trying to convince a jury nor a judge -- this thread is merely an exploration of a hypothesis.

Quote: AlanMendelson

It would just be circumstantial evidence. To prove a bias you would somehow have to examine and prove either by weight or size or construction that the dice were not manufactured to have an even distribution.

Not that it matters to this thread, but circumstantial evidence has won many court cases.

Also, determining that a die is off-balanced or "uneven in construction" (shaved?) would also be circumstantial evidence (not direct evidence) as to such a condition causing a results bias. Combining such a physical irregularity with peculiar results (or vice versa) could corroborate such evidence.

Quote: AlanMendelson

Any attempt to indicate that any sort of "count" showed any bias and the "defense" would say "dice is a random game and any number of faces could show at any time."

Just because one of the hypothetical parties makes such a statement doesn't mean that it is so, nor does it mean that that party wins the court case.

Furthermore, statistical evidence from a large sample of rolls could be corroborated by expert testimony such as that from an actuary like Wizard who, incidentally, is the one who suggested this new thread to count a large sample of throws in judging bias.

Quote: AlanMendelson

Further the dice makers would blame the tables for impeding a smooth and even roll of the dice -- unlike dice on a spinning device.

Huh? Whatever.

Quote: AlanMendelson

In short, give it up already. OR find some way to dissect dice to determine how each section of the die weighs and what its dimensions are.

Dissecting the dice is not a very good idea, as has been discussed in another thread.

Quote: AlanMendelson

Look if you want to do something scientific then do something scientific. tracking rolls of the dice wont cut it.

You certainly know your science.

Hey guys, this might prove interesting and informative. Let the guy have his go at collecting data...though I think his question was already answered in New Jersey at the Golden Nugget. Some Pai Gow players answered it for him. Only they noticed an exploitable eight deck shoe...and they exploited it.

Anybody ever notice an exploitable die is gonna do the same. Them dice have short lives at the table though. So if anyone really wants to think that after the frequent examination by staff and occasional examination by players the dice are noticeably and significantly nicked then let the guy go and collect the data on it.

I ain't gonna stand there clicking clickers all shift long, but if he is willing to that is fine with me.

Its just that I don't think he is gong to be seeing anything but callouses on his fingertips. Any exploitable bias will be senses by the sharpies and bet on while he is clicking away.

Good luck.

For those of you who think tracking rolls will prove something answer this question:

How will you explain all of your statistics when you realize that you are all tracking DIFFERENT PAIRS of dice?

Do you start to understand that this is an exercise that is meaningless?

IF by some miracle you threw the SAME pair of dice a hundred thousand times, AND the dice were never worn or damaged, such a "test" might reveal something. But no... you are going to be recording rolls on different dice from different casinos, from different days, with different elements of wear... and you think that is supposed to present something meaningful?

Get real.

Now if you want a fair test do this:

Buy from each manufacturer six thousand dice that it makes for each casino. These would be new, virgin dice, never rolled.
Then, one by one roll each of the dice one time and record the roll.
Then after rolling each of the six-thousand virgin dice one time, see if you get a random distribution of the faces -- one-thousand for each face.

Let me know what you find.

You won't find a significant bias unless you reproduce the initial conditions at each throw - which is nearly impossible at a live table.

That is why scientists make any kind of statistical tests in a controlled environment (i.e. a laboratory), where they spent large amount of effort to stabilizing those properties which they think are important. Otherwise, even if there is a deterministic outcome of your dice throw, it will be masked by your varying initial conditions.

So make some clever device which will allow you to throw the dice in a very reliable fashion. For a dice bias test I would probably use some kind of upside-down centrifuge where the dice is set with the same orientation to the wall at a defined friction. Spin up the centrifuge (the dice will stick to the wall) and let it freely slow down. Once the decreasing frictional force applied by the centrifuge drops below the gravitational force, the dice drops out by itself. This will always be with precise the same rotational frequency, and hence velocity, regardless how you spin up the centrifuge. Then you have a dice with precise orientation and velocity. Let it fall on your desired surface and only then you should collect your results.

Quote: MangoJ

Scientists make any kind of statistical tests in a controlled environment (i.e. a laboratory), where they spent large amount of effort to stabilizing those properties which they think are important. Otherwise, even if there is a deterministic outcome of your dice throw, it will be masked by your varying initial conditions.

To be called science or scientific, there are two additional minimal requirements before a hypothesis can be found valid.
1) The tests have to be double blind.
2) The results have to be reproducible.

Quote: tupp

1. conclusions of courts often have no basis in scientific reason nor in reality.

Maybe you have never sat on a jury that weighed forensic evidence. Some of us have. And where it is possible to produce physical evidence that is a major step in forming conclusions on the way to justice and, presumably, the truth.

I promised to participate in this thread if I saw some evidence collected under casino conditions. So far, I haven't seen any. For example, here is the kind of table I'd like to see. This was based on random numbers drawn in Excel.

Face Up	Number
1	158
2	172
3	161
4	183
5	174
6	152
Total	1000

Ahigh, I know you were collecting data when I saw you at the M, so may I see your data thus far?

What baffles the snarf out of me is that people that beleive the game is totally a random game,
people that throw the dice and they land all over the table, each throw completely different than
the next, actually spend hours ranting about biased dice.

Alan has it correct, when every table has a different set of dice atleast each day, sometimes
changing more than every 24 hours, you can't possibly provide information that would be of use to
anyone .

You may as well beleive in global warming.


Dicesitter

Quote: dicesitter

You may as well beleive in global warming.
Dicesitter

At least global warming provides for the collection of a statistically relevant set of data involving one earth and one atmosphere. Claiming measurable dice bias is like claiming global warming exists, but doing it with tiny data sets across multiple earths.

I am very skeptical that anything will be found, but to the naysayers:

Imagine if a bunch of recreational craps players were to combine their data, run tens/hundreds of thousands of rolls through a chi-squared test, and find that there is an unexplainable deviation from the norm for some face? What would we say then?

If some casino dice are biased, and others are not, the sum of their face distributions is also biased. If they are biased in random ways, we probably won't notice. If they are biased in the same way, we will. I think anyone who wants to participate should. I think the Wizard is taking a fine stand here. Put out the math in nice tables, and trained people will analyze it.

I'd bet money we won't find anything, but why do people care what other people do in their spare time? They're not juggling kittens.

Bingo. If a sub set of the dice are biased in the same way, the total data will show bias.

By the way, again I point out there is a source for thousands of dice rolled using commercial boardgame dice.

Quote: Wizard

I promised to participate in this thread if I saw some evidence collected under casino conditions.

I've expressed elsewhere that, in my view, bias dice--if it even exists--cannot be exploited. (I also recently asked Ahigh or others, in that thread, to respond to several questions about his views on bias dice, but so far no response.)

I also believe this whole proposed study of looking at actual rolls from different casinos will not conclusively tell us anything about bias dice including whether bias dice exists or how the dice may be skewed or, most importantly, the degree of exploitation that can be utilized by either the casino or the player.

Perhaps over many, many hundreds of thousands of rolls under exacting controls involving a mechanical throwing device as opposed to humans and involving a table that does not change over the course of the test--using fresh dice every 8 hours from the same "batch" of one particular manufacturer (since comparing dice from different batches or from different manufacturers would not be supportive of anything)--then maybe we might discover that dice are not perfect. Big "maybe." I have no doubt whatsoever that such an exhaustive test will at least prove one thing: that the dice are, indeed, perfect enough.

Nevertheless, for the sake of Ahigh's desire (and perhaps others as well) to pursue this topic, I offer the following:

There is a book entitled Craps -- A Smart Shooter's Guide written by Thomas Midgley and published by GBC Press (Cardoza Publishing) in 2011.

Midgley recorded more than 7,500 actual rolls of the dice at five Las Vegas casinos (apparently over 40 hours). He used this data throughout his book in support of his ideas on how to approach the game (as opposed to using mathematical formulas). The book states that Midgley concludes that "dice operate in cycles and that is why systems players lose." The book's subtitle is: "The Book That Mathematically Shows You How To Profit From The Cycles And Patterns Of Dice!"

I do not present or cite this book as authoritative. I have a very dim view of most books on craps--especially if the book espouses any sort of "betting system."

However, Midgley does include the results of the recorded rolls, and it is for this purpose only that I refer to the book.

The five casinos from which Midgley collected his data were Frontier, Hilton, Caesars, Sands, and Nugget. During the time the data was recorded, the dice were thrown 7,654 times, making 15,308 rolls of an individual die. Midgley includes a table showing the number of times each die number appeared at each casino. From this table, the totals for all five casinos for the 40 hours are as follows:

Die showing the number 1: 2,498

Die showing the number 2: 2,647

Die showing the number 3: 2,606

Die showing the number 4: 2,590

Die showing the number 5: 2,462

Die showing the number 6: 2,505

Based on the 15,308 rolls of the individual die, the mathematical probability for each number appearing is 2,551.

Midgley then proceeds to present the recorded results of the dice combinations of the throws--broken down by casino--so that you can see how many 2s were thrown, how many 3s, how many 4s, etc. And in each case, he presents the information precisely--so that, for example, with the number 7 for the two dice, he tells you (again for each casino) how many were 1-6 and how many were 2-5 and how many were 3-4. As well as for the other dice combinations.

As for the sum of both dice for all casinos, the times each number combination appeared:

Dice totaling 2: 210 (Probability is 213)

Dice totaling 3: 433 (Probability is 425)

Dice totaling 4: 656 (Probability is 638)

Dice totaling 5: 858 (Probability is 850)

Dice totaling 6: 1065 (Probability is 1063)

Dice totaling 7: 1265 (Probability is 1276)

Dice totaling 8: 1067 (Probability is 1063)

Dice totaling 9: 867 (Probability is 850)

Dice totaling 10: 652 (Probability is 638)

Dice totaling 11: 399 (Probability is 425)

Dice totaling 12: 182 (Probability is 213)

Frankly, I don't think this is supportive of anything--other than based on a "limited" run of 7,654 rolls of the dice, there will be slight variations from true mathematical probability. But we all already knew that.

But if you want to see "some evidence collected under casino conditions," then here it is.

Here is my analysis of the Midgley results.

First, one die at a time.

Die	Observations	Expected	Chi-squared
1	2498	2551.33	1.114885
2	2647	2551.33	3.587188
3	2606	2551.33	1.171327
4	2590	2551.33	0.586012
5	2462	2551.33	3.127951
6	2505	2551.33	0.841434
Total	15308	15308.00	10.428795

The p value for a chi-squared statistic of 10.428795 with 5 degrees of freedom is 0.063958.

Next, the sum of the two dice.

Total	Observations	Expected	Chi-squared
2	210	212.61	0.032067
3	433	425.22	0.142264
4	656	637.83	0.517420
5	858	850.44	0.067125
6	1065	1063.06	0.003557
7	1265	1275.67	0.089191
8	1067	1063.06	0.014636
9	867	850.44	0.322286
10	652	637.83	0.314650
11	399	425.22	1.617048
12	182	212.61	4.407296
Total	7654	7654.00	7.527541

The p value for a chi-squared statistic of 7.527541 with 10 degrees of freedom is 0.674878.

So, these throws look random to me.

Exactly.

What bothers me is that there are claims (Harley) that say that you can tell that dice are biased in a few tens of rolls. Now, if you have data that says that a die manufacturers produce faulty dice with the same dice across thousands of dice in the same set of casinos, then there could be data that supports or refutes that theory.

Chi-Squared and Binomial distributions are the *only* proven standards out there. Deviations from the random curve can be measured with a probability value. The more data that you collect, the more likely that the probability value will fall to 0 if the dice are indeed random.

Anything else is human observational bias -- our mind seeking patterns in life in order to make sense of the world.

The other way to determine that dice are biased is through physics. You can model the exact characteristics of the dice and put it into a modeller and simulate millions of throws to determine bias. But even still you have theory because you are inputting the model based on characteristics you believe to be true.

Going back to the global warming argument, the fact that man-made carbon dioxide is increasing in the environment does not mean that the earth will get warmer. There are many many other factors besides carbon dioxide at play in the global model. Therefore, you can only prove global warming is real through observation. Similarly, even though a caliper might show different measurements on the sides of the dice and the weighting of the dice may be different doesn't mean that it will result in biased results.

Lately, Ahigh has been attempting to show something of a string of consecutive winning sessions to prove something. I don't know what that is - that he's a great gambler, that he is influencing the dice, both... who knows... it's a mystery.

But if we're talking bias, the composition of a dice doesn't prove bias. You can theorize based on the composition of the dice that there *should* be a bias but without the experimental data (ie, real throws) to back it up, it's just theory.

No more comments on global warming in this thread. If you want to discuss that please make a separate thread for it, or add to an existing one.

Quote: boymimbo

Lately, Ahigh has been attempting to show something of a string of consecutive winning sessions to prove something. I don't know what that is - that he's a great gambler, that he is influencing the dice, both... who knows... it's a mystery.

I think Ahigh reports both wins and losses, as others on the forum also do. However, nobody ever focuses on the motivations of why the other members post their wins/losses.

Quote: boymimbo

But if we're talking bias, the composition of a dice doesn't prove bias. You can theorize based on the composition of the dice that there *should* be a bias but without the experimental data (ie, real throws) to back it up, it's just theory.

Well, the theory is that a shaved die will have a greater propensity to land on the shaved side and it's opposite face and a loaded die will have a greater propensity to land with the heaviest face down and the heaviest face up.

Let's do try to stay on topic here also. What I am specifically interested in is someone else willing to do some clicker work. If I do all the work myself, this isn't going to be fun nor am I going to be able to get enough samples to matter. So let me start off with this offer! I will pay $100 cash to any participant who buys there own clicker and collects more than 6000 real samples (from 3000 throws) with a sufficient number of relative details to convince me that they did the work (I can refuse payment for any reason, but I will be fair .. just not a sucker!). I will pay another $100 for each additional 20,000 samples (I may agree to a pro-rated payment as well on a case-by-case-basis). In order to qualify for payment, I expect to have photos of your chips on the rail along with something in writing proving that the samples were collected. In effect, I am splitting the gambling cost of $0.01 per sample if you were to be betting the line in order to collect the samples on an empty table with a sufficiently random roll.

This is offer is for Las Vegas only!!

I will also make custom mounts for anyone buy the 6-gang clickers. This offer may expire at any time.

This in effect will cut the house edge on a line bet with full odds in half. Those of you wishing to get a better deal who are math guys should jump on this opportunity!

You will need to get an "okay" from me to start doing the work before you begin so that I don't over-commit on a budget to pay for samples.

If you are interested, PM me and tell me what you plan on setting out to do so that I can set aside cash to pay you for samples.

Here's a link to buy bank tally counters. The cost for your tally counter should be below $50. The 6-bank counters are sold out, but you can probably call and find out if they will be available soon or buy two 3-bank counters.

http://www.tallycounterstore.com/Bank_Tally_Counter_p/014.htm

Quote: dwheatley

I am very skeptical that anything will be found, but to the naysayers:

Imagine if a bunch of recreational craps players were to combine their data, run tens/hundreds of thousands of rolls through a chi-squared test, and find that there is an unexplainable deviation from the norm for some face? What would we say then?

If some casino dice are biased, and others are not, the sum of their face distributions is also biased. If they are biased in random ways, we probably won't notice. If they are biased in the same way, we will. I think anyone who wants to participate should. I think the Wizard is taking a fine stand here. Put out the math in nice tables, and trained people will analyze it.

I'd bet money we won't find anything, but why do people care what other people do in their spare time? They're not juggling kittens.

I hate to be blunt but I would assess the data this way: garbage in, garbage out. I am surprised that the Wizard would accept any data except scientific data gathered under scientific conditions.

If you guys came to me with this kind of data and told me you had "news" about biased dice (if that's what it showed) I would tell you I couldn't accept it because you have no controls, you have mixed data, mixed conditions (different dice) and no verification of the data since it comes from different people without verification.

Sorry. Ahigh would be doing better science if he just took one new die from each casino and on his table at home rolled it a thousand times and videotaped the results. Now THAT would at least give us some controls to work with: one die from one casino, one table, a thousand rolls, videotaped. But if that showed a biased die, then what? Yeah, Ahigh found ONE biased die... and dice are thrown in pairs so what about the other? And what about the 10 or 120 dice that a casino uses in the course of a day?

I know there are some who want to keep this "investigation" alive but come up with something meaningful and not this... forgive me for being blunt... dribble.

Quote: AlanMendelson

If you guys came to me with this kind of data and told me you had "news" about biased dice (if that's what it showed) I would tell you I couldn't accept it because you have no controls, you have mixed data, mixed conditions (different dice) and no verification of the data since it comes from different people without verification.

Even then, what's the best case scenario? Say you do accept the data. We've now "proven" that a casino used biased dice in the past. No conclusion can be drawn about future uses of biased dice, and no governing body can intervene to address the "problem."

Quote: rdw4potus

Even then, what's the best case scenario? Say you do accept the data. We've now "proven" that a casino used biased dice in the past. No conclusion can be drawn about future uses of biased dice, and no governing body can intervene to address the "problem."

Agree.

the only way to prove a "case" is to get a whistleblower from the dice manufacturers or from the casinos to come forward and say there is a plan to introduce biased dice and to show a box of new dice to present in court for court supervised testing to prove their bias.

The arguments that have been going on here resemble the types of arguments that go on over beers at a bar. And I dont drink.

Quote: Wizard

Here is my analysis of the Midgley results.

So, these throws look random to me.

Wizard, this is wonderful work, but maybe as a baseline study only. I do not have Midgley's book in front of me now, but if I recall correctly this roll data is from 1978ish ? No wonder 2 of the casinos are now obsolete. Caesar's is NOW using some of the worst reported biased dice, maybe they had to change their operations formula from using fair dice to semi-survive.

That is a mystery to me also. All of us have won several times in a row, and all of us have lost several times in a row.

NOw i am a dicesitter, but winning several times in a row does mean that all of a sudden the casino has no chance to
win. and losing several times in a row does not mean that all of a sudden i forgot how to roll dice.

In terms of showing you have come control over the dice... it is simple, if your throw is random, if your dice land
at different places on the table, if they come to rest all over the table, never showing any control at all, you
are a random thrower...end of story...

that does not mean you cant win several times in a row, it means it is the luck that comes with anything.

Luck is a nice friend to have, he will visit now and then, but never move in with you.


dicesitter.

Quote: dicesitter

NOw i am a dicesitter

What is a "dicesitter?"

Is it the latest description of "dice setter?"

A new spin on "dice influencer" or "dice controller," perhaps?

I can certainly appreciate how sitting on the dice before rolling them could help, as I've pulled my fair share of good rolls from out of my ass.

I imagined it's because they can(*) 'sit' the dice snugly on the table from a throw, with a better than random chance of hitting his target numbers.

(claim they can, suggest they can, think they can... whatever).

Quote: rew4potus

Even then, what's the best case scenario?

I also continue to scratch my head in trying to understand the implication or inference to be drawn from "proof" that a casino has used biased dice.

Bias or skewed in what way? In favor of the casino or in favor of the player? How?

If biased dice are skewed in favor of the casino, then does that mean that the number 7 is more likely to appear? How do you then account for Don't players?

Is there any evidence that the casinos are complicit in using biased dice?

I also do not believe the gathering of this data as proposed will be "proof" of anything--beyond the fact that the throws are random with results approaching mathematical expectation as more information is gathered.

But as was stated:

Quote: dwheatley

but why do people care what other people do in their spare time? They're not juggling kittens.

To each his own.

Yeah but we won't know until we try.

I would think that 12 would be most likely to come up on house-biased dice. But I don't know.

Also, in the other-information category, both of my cats very much enjoyed being gently thrown.

I don't expect anything except for a scientific analysis to occur. I doubt that there will be proof of anything either way with the results, but as the Wizard points out, samples from real casino environments collected in a scientifically astute manner is a missing ingredient.

Using this counter during play alters the social play dynamic dramatically. In many ways it is actually a lot of fun. But this is not the type of thing that you would want to do with a controlled shot. I think this is a better setup for someone who wants the best possible deal possible (IE: a hard-core don't player who plays the don't because the edge is better). That's why I am offering to cut the edge for this type of player in HALF if they can provide data with reassurances that the data was gathered with with proper diligence.

I would prefer not to gather any of this data myself just to remove myself from the possible scenario that I fudged the numbers for some reason. Anyone else who has a strong opinion about biased dice should also ideally not be involved in gather the data. That is why I would prefer to pay for some don't betters to gather the data and to pay them for it.

The value of the payment is tiny compared to the total effort to collect it. It only makes sense for the guy who is really counting every penny on every roll.

So far nobody has expressed interest in gathering data, and frankly I don't expect that to change.

I don't want to gather data either as I prefer to have a different experience at the table and also to attempt to influence the dice rather than record the bias the dice might have from an intentionally random throw.

But if we could find a few don't betters who like to play the game a lot, there is a chance for a nice symbiotic relationship if there are any statistically minded don't betters willing to help.

I am willing to front the cost of the counter for the first person to sign up for this task (and reduce the initial $100 payment on the first batch of samples by $50) and pledge to record 3,000 samples minimum withing a period of time of no more than one month.

Quote: Ahigh

So far nobody has expressed interest in gathering data, and frankly I don't expect that to change.

Are you saying you only want face counts? No “keep track of which dice” or “note when they’re changed” or anything else? Just face counts?

OK, a scientific study could be fun.

How many dice manufacturers are there? Is there any way to tell which one supplies which casino(s)? I see the data collection methods described, and the mathematical data analysis is relatively straightforward. But what about the sample groupings to be analyzed? If Wynn and Caesars get dice from different places, we shouldn't aggregate their dice for analysis.

Quote: Bohemian

Caesar's is NOW using some of the worst reported biased dice

Reported where? "NOW using"--meaning that it is ongoing? And this is common knowledge? "Worst" biased dice in what respect?

The only news of which I am aware involving Caesar's is that Casino Horseshoe Cleveland was fined $180,000 for a variety of violations of internal procedures and state law--one of which involved dice.

For six months after the casino's opening in May 2012, the casino used dice in a game called Fortune Pai Gow Poker without submitting a sample or photograph to the Ohio Casino Control Commission and getting the commission's approval. There was never any indication that the dice were rigged. The dice had simply not been approved.

Is this the situation to which you refer that Caesar's is now using some of the worst reported biased dice?

Speculation is the Bohemian is Harley. And no offense to Harley or his followers, but let's keep this whole process on track to be scientific, alright?

Quote: rdw4potus

OK, a scientific study could be fun.

How many dice manufacturers are there? Is there any way to tell which one supplies which casino(s)? I see the data collection methods described, and the mathematical data analysis is relatively straightforward. But what about the sample groupings to be analyzed? If Wynn and Caesars get dice from different places, we shouldn't aggregate their dice for analysis.

Bud Jones
TK
Paulson
Midwest Gaming
Bee

These are the most commonly known dice used in Vegas.

I learned that Wynn uses Midwest Gaming and Paulson dice exclusively by asking a member of the staff.
Boyd properties are known to use Midwest Gaming dice very frequently.
Midwest Gaming dice appear to have smaller pips as a ratio of the size of the cube and are often easily spotted with practice.
Midwest Gaming dice are the only dice to be made in the USA.
The M uses Bud Jones dice. The M is the only place I have gotten a response from that uses Bud Jones branded dice.
Bee dice are very common. To get a list of casinos that use Bee dice, you can look at the rejected dice at the Gambler's Supply store that come from the factory.

Bud Jones, TK, and Paulson dice are all brands owned by Gaming Partners International.

http://www.gpigaming.com/casino-precision-dice

Bee dice casinos include:
* Silverton
* The M
* Red Rock
* Palazzo

This is just to get started, but yes, a trained person can spot a brand of dice by the DETAIL of the color hue of the dice, the font of the serial number and other details. Each manufacturer has multiple fonts available for serial numbers so it's not absolutely trivial to spot the manufacturer by eye, but I have done this at many casinos with verification from the staff what kind of dice they are using. Most recently, I told them at Lucky Club that they were using Bee branded dice because I recognized the exact hue and color of the dice as a Bee die.

The floor manager said, "I don't know what brand they are." She went back to check the foil and verified, "yep, they're Bee."

Well sitting the dice does not mean a thing if you cant throw them the same way.

I surely would not
want any part of your body getting sore from to much use, so relying on luck will prevent that.


dicesetter

Ahigh, what exactly is the motivation for doing this?

Was there a personal experience in which you somehow inescapably concluded that biased dice were used? Or was this something reported to you? When and where and under what circumstances and in what respect was the dice biased?

Considering today's regulatory control of casinos with provisions pertaining to dice specifications (granted not all states are the same and some states may have less and some more by way of specifications of dice), do you believe that the regulatory commissions have failed in their duty to protect the integrity of dice games by allowing biased dice to be used?

Surely you are not wanting to pursue this endeavor on a whim?

Quote: Ahigh

let's keep this whole process on track to be scientific

And you think the data that you hope to gather in the manner that you have suggested is pursuant to scientific method?

Quote: Jimbo

If biased dice are skewed in favor of the casino, then does that mean that the number 7 is more likely to appear? How do you then account for Don't players?

For one, I'm rarin' to go. Is that Caesar's LV only or all CZR? Any limits imposed? Last time that this was suggested, all such questions were painfully ignored. Please, Bohemian, give us more guidance. Which color is the most cockeyed dice? Can we parlay the 12?

There is a feedback loop where a casino looks at the hold. This feedback loop will naturally prefer dice that go against the largest bets on the table (the free bets).

That's enough evidence for me to be suspicious.

If the dice come out fair, that is a FANTASTIC RESULT!

I want unbiased participants collecting and documenting their gathering of data.

I want this issue to be closed off an attention paid to more important details.

This issue will be closed off by dice being manufactured more fairly if there is a problem.

If there's not a problem, it's a simple matter of gathering enough data to show that there is not a problem.

The best time to gather this data is the busiest times when there are the highest percentage of free bets on the table. That is the time when in theory the casino would benefit the most from dice biased against the free bets.

I am hopefully pretty unbiased about the reality of this happening. But having 100% unbiased people collecting the data I think is a critical component to moving forward.

I applaud the Wizard to getting involved, and hopefully he also appreciates my financial incentives to buy some data for this purpose using scientifically nit-picky people who bet the dont' very frequently already. There has to be some don't players looking to cut their edge who are out there in Vegas!!!

Quote: Ahigh

I want unbiased participants collecting and documenting their gathering of data.

...it's a simple matter of gathering enough data to show that there is not a problem.

The best time to gather this data is the busiest times when there are the highest percentage of free bets on the table. That is the time when in theory the casino would benefit the most from dice biased against the free bets.

I am hopefully pretty unbiased about the reality of this happening. But having 100% unbiased people collecting the data I think is a critical component to moving forward.

Check back a page. I posed a question that needs answering and have another one lined up based off your answer...

Quote: Face

Are you saying you only want face counts? No “keep track of which dice” or “note when they’re changed” or anything else? Just face counts?

Yes just face counts. I don't care about the order or the sequence. Just the occurrences of face counts. You're basically just profiling the face outcomes as two samples per roll, or one sample per roll if one die goes down. That's it.

I do want data of what casino, what brand dice they are, what time of day and day of week it is if that data can also be recorded.

Quote: Ahigh

There has to be some don't players looking to cut their edge who are out there in Vegas!!!

Though it is clear from my previous remarks that I believe this endeavor will accomplish nothing, I am curious by your comment. I am a Don't player who is regularly in Vegas. I don't see what you are proposing in the data gathering process that will "cut my edge." (And did you really mean to say that as a Don't player, I have an edge that I am willing to cut?)

Quote: Jimbo

Though it is clear from my previous remarks that I believe this endeavor will accomplish nothing, I am curious by your comment. I am a Don't player who is regularly in Vegas. I don't see what you are proposing in the data gathering process that will "cut my edge." (And did you really mean to say that as a Don't player, I have an edge that I am willing to cut?)

If you bet the don't pass with maximum odds on a $5 table, you're edge is 0.40% or $0.02 per roll. I am offering to pay $0.01 for every two dice resolutions. That means you can play the don't pass with maximum odds for a cost of $0.01 per roll if you click these things and write down numbers for me.

Let me know if you have more questions, but you have to only be playing a single don't pass with max odds for this to work out as cutting your cost to play the game from $0.02 per roll to $0.01 per roll.

Quote: Ahigh

Yes just face counts. I don't care about the order or the sequence. Just the occurrences of face counts. You're basically just profiling the face outcomes as two samples per roll, or one sample per roll if one die goes down. That's it.

I do want data of what casino, what brand dice they are, what time of day and day of week it is if that data can also be recorded.

... and you answered my second question in your reply. Very good.

I would like to point out, also, that if you take suggested advice on when and where to play this way and you do find biased dice and bet the don't with max odds, you will win. This is all theory, but max odds with a single don't pass bet on the line is a bet that will benefit in the long run from biased dice. You will win from having more than 1/36 rolls being aces. And you will win from your free lay odds bets in the long run. Again, all just theory of the heavy six and heavy one faces coming up.

Understand that I am not claiming biased dice exist. Just saying that if the theory turns out to be true, you can profit as the result of finding these biased dice as you bet the don't with max odds. This gives someone a further incentive to try to sniff out these theoretical cubes.

You might also benefit from looking for any tables that only pay double on the 12 in the field. This culls out a loophole of theoretically biased dice (heavy six face and heavy one face) that they can be exploited easily by a small bet in the field over and over.

Quote: Ahigh

If you bet the don't pass with maximum odds on a $5 table, you're edge is 0.40% or $0.02 per roll. I am offering to pay $0.01 for every two dice resolutions.

Oh, I see what you are saying--you are assuming that Don't players play for only $5. I don't bet $5. My average bet is considerably more than that--which accounts for my confusion in what you were proposing.

Secondly, I believe it is the house that has the edge, not the player. And the house edge on the Don't Pass per bet made is 1.36% without odds and the house edge (per bet made) for 3x-4x-5x is, I think 0.27% (which is what max odds would be where I typically play in Vegas). I don't know why you suggest the Don't player has an edge of 0.40%.

Quote: Jimbo

Oh, I see what you are saying--you are assuming that Don't players play for only $5. I don't bet $5. My average bet is considerably more than that--which accounts for my confusion in what you were proposing.

Secondly, I believe it is the house that has the edge, not the player. And the house edge on the Don't Pass per bet made is 1.36% without odds and the house edge (per bet made) for 3x-4x-5x is, I think 0.27% (which is what max odds would be where I typically play in Vegas). I don't know why you suggest the Don't player has an edge of 0.40%.

Per roll. 0.40% per roll. And of course the house has the edge, not the player, talking about the math alone. I generally ONLY think about edges per roll as I don't even KNOW how many discussions turn to arguments comparing apples to oranges when it comes to edges in craps.

Per roll only with me. That's just my way. The minimum cost to play craps $5 and up is two cents per roll in general. A fact that I believe any experienced craps player should learn if they don't already know.