# ugly night a the tables

Discussion in 'General Craps Discussion' started by scott22, Feb 5, 2012.

1. guido1, Mar 8, 2012

### guido1 Member

Joined:
Mar 12, 2011
Messages:
88
1
Location:
Newport Beach, CA
falcon...bro'

we also have to stamp out the 1/66 or 1.51% house edge myth on the place 6 and place 8 bets.

Another total total myth!
(because everyone and us left out the variance)

we will return in another thread for that!

I hear you falcon brother!!!
No freakin' way everyone that makes place bets on the 6 and 8 will lose 1.51% of the total bets made.
Its a lie! An untruth!

an expected value without value!

It is a lie so the truth will not be told.
But falcon will tell the truth about place bet ev and variance.

Yahoo!!!

#101
2. falcon, Mar 8, 2012

### falcon Member

Joined:
Jan 17, 2012
Messages:
420
6

#102
3. falcon, Mar 8, 2012

### falcon Member

Joined:
Jan 17, 2012
Messages:
420
6
guido1: All of the above will be required knowledge for MS's virtual college of craps BS degrees and as a condition for graduation matriculating students must stand on top of an unused craps table in a crowded casino and shout out the very meaningful equations you have provided.

Unfortunately, the MS college will not field a competitive basketball team for March Madness this year.

falcon

#103
4. falcon, Mar 8, 2012

### falcon Member

Joined:
Jan 17, 2012
Messages:
420
6
Scream it from the mountain tops and prove it with real play.

falcon

#104
5. basicstrategy777, Mar 8, 2012

### basicstrategy777 Member

Joined:
Jul 4, 2010
Messages:
9,227
8,422
Gender:
Male
Location:
CT.
There is no "perfect".

The law of averages asserts that in many dice rolls, the fraction of 7's will be approximately .16667. Computer simulations verify's this, but the term 'approximately' begs to be quantified. The reason you use 'approximately' or 'close to' is because of chance variation, an inherent part of randomness. The law of averages is a limit theorem ; it states precisely what will happen in an infinate number of observations, something we will never see. When we talk about the long run in practical terms, we mean a large but finite number of observations, and so we must allow for chance variation. Chance variation is best inderstood in the context of how things vary arounfd the average value.

The law of averages says that in a large number of trials, the actual number of occurences will be close to the average number of occurences, with some variation due to chance.

The casinos stake their financial survival on this math and what it says.

It is unfortunate you cannot understand or accept what is said by many people smarter than you. That is why you are a fool.

777

#105
6. Sancho Panza, Mar 8, 2012

### Sancho Panza Member

Joined:
Jan 30, 2010
Messages:
430
18
Conveniently omitting the fact that "multiple consecutive" wins are twice as likely as the "multiple consecutive losses." Meaning that the "great difficulty for a meaningful recovery" is not only erroneously overstated, but also entirely the reverse.

#106
7. The Midnight Skulker, Mar 8, 2012

### The Midnight Skulker Member

Joined:
Jan 28, 2010
Messages:
4,035
4,230
Gender:
Male
Location:
Idaho, USA
Up front I will give you a tip of the hat for trying to get the quotes right. Only a slight tip, however, since your technique still needs some work.
11.1% is 1 time in 9. So you are saying that the PL bettor nets a win every 9 -- um -- what? Times the bet is resolved? Times the bet is resolved on the comeout roll? Percentages have to apply to some total number of things. To what does your 11.1% net advantage apply?
I readily concede that when you're on the wrong side of an Ice Age a meaningful recovery is difficult. However, that recovery can be accomplished with point cycle wins in addition to comeout wins. (BTW point conversions happen about 40.6% of the time.)
Actually Basic Sevens beat me to the punch on this one. He taught a college level class on craps, complete with student manual which he wrote. The class included an exercise which had the students roll dice 144 times IIRC and record and submit the resulting distribution. Again IIRC the purpose of the exercise was to show that while few if any of the 144-roll sets were perfectly distributed, the actual distributions tended to be clustered around that perfect distribution.

A similar situation exists with the Perfect 1980. The probability of 1980 actual rolls of the dice being distributed exactly as shown in the spreadsheet I posted previously is very small, but of the millions of possible such distributions it is the most likely. Furthermore, the probability of getting any distribution of 1980 actual rolls decreases as the difference between that distribution and the perfect one increases. IOW the Perfect 1980 gives the expected value, the average distribution if you will, of a large number of 1980-roll sets. In any event, to address your complaint that the Perfect 1980 is inapplicable to this analysis because the chances of it being obtained are remote, I have redone the spreadsheet using probabilities. To no one else's surprise, the results are the same: a straight Pass Line bettor taking no odds is expected to lose less than the player who places the point after it is established.
And so we see that your ignorance is intentional. Faced with incontrovertible evidence of your error you deride your opponent and dismiss the topic as being of no consequence. Nevertheless, I have enjoyed the repartee. Adieu.

#### Attached Files:

• ###### PLvP.jpg
File size:
82.6 KB
Views:
207
#107
8. guido1, Mar 8, 2012

### guido1 Member

Joined:
Mar 12, 2011
Messages:
88
1
Location:
Newport Beach, CA
The Law of Averages is NOT the same as the Law of Large Numbers.
From wikipedia:
"The law of averages is a lay term used to express a belief that outcomes of a random event will "even out" within a small sample.
As invoked in everyday life, the "law" usually reflects bad statistics or wishful thinking rather than any mathematical principle."

The Law of Large Numbers, the real deal and theorem, deals with ratios, percentages and averages. Like 1 in 6 or 1/6 or .16666

The more trials, the percentages get closer to the theoritical average or ratio NOT necessarily the absolute number of occurrences or outcomes.

In reality as the number of trials increase the absolute differences tend to get larger, not smaller as the percentages get closer to expectation.

From the below picture is a quick random walk data set.
The delta column is the difference between heads and tails.
It kept growing as the # of trials increased. Most times it does.
But the actual percentage of heads per trials went down each time, very close the the expected 50% value.

falcon and I were discussing this last night at the Craps table while we were both hitting on the hottest cocktail waitress ever!
She hit us both, but it was worth the price of admission.
Falcon, you awake yet!

#108
9. basicstrategy777, Mar 8, 2012

### basicstrategy777 Member

Joined:
Jul 4, 2010
Messages:
9,227
8,422
Gender:
Male
Location:
CT.
There are layman here.

I used a term most here would understand..... and qualified it to the extent where it was precisely correct and most mathemeticians would know what I was saying.

Did you have a problem understanding what I said ?

777

#109
10. guido1, Mar 8, 2012

### guido1 Member

Joined:
Mar 12, 2011
Messages:
88
1
Location:
Newport Beach, CA
You said:
"The law of averages says that in a large number of trials, the actual number of occurences will be close to the average number of occurences, with some variation due to chance."

We all need to talk more about the Law of Large Numbers and everyone will learn the real truth.

The law of averages has no mathematical validity.
Period.
None.
That is why your statement of the law in your above post needs to be pointed out as a false statement. It should not be used by anyone.

The Law of Large Numbers is the only correct law to use here. It has been proven mathematically by many.
Everyone needs to Stop using the Law of Averages as it is not a law but an incorrect belief.

You know the difference.
So one should use the correct term, not a useless law that contains errors.

There exists not one proof of the "law of averages" but many proofs that it is not even a law.

Just like most Craps players say a shooter "craps out" after rolling a 7 with a point established.
It is not correct but many still say it.
That does not make it acceptable because many will understand it.
It needs to be shown why it is not correct and stop the use of it.

I know Falcon and I just hate it when others push untrue statements.
The truth needs to be told 100% of the time.

The part in The Law of Averages is not correct by those that say "the actual number of occurences will be close to the average number of occurences, with some variation due to chance."
Simulations easily prove that statement to be incorrect. The actual number of occurrences can be close or far away. That is what a random walk is all about.
The Arcsine Law is a proof of what can happen in a random walk, NOT the incorrect Law of Averages.

I say down Down with the Law of Averages.
Up with the Law of Large Numbers.

Another freakin' shooter just "crapped out" and now we have a new shooter.

#110
11. 7Craps, Mar 8, 2012

### 7Craps Member

Joined:
Feb 1, 2012
Messages:
64
1
Even after most are told the difference between a "7 out" and a "Crap out" about 99.99999999999999999% of those that have been corrected just say Crap out instead of 7out.

Oh well.
That shows that most Craps players do not care about being correct.
Even tho' it is not correct in usage, most do not care.
Like bananas from Africa. Who cares??
Oil from where!
Sell me 2,000 barrels for \$115 each!

my 2 cents.
The Law of Averages is no law, it is a fallacy based from the misunderstanding of the Law of Large Numbers.

It came from those that did not pay attention in math class, yes, the ones that just did not care.

If you are not a George, you are a flea and can now go home.

#111
12. basicstrategy777, Mar 8, 2012

### basicstrategy777 Member

Joined:
Jul 4, 2010
Messages:
9,227
8,422
Gender:
Male
Location:
CT.
Guido.......The question becomes, is 'law of averages' commonly used as I have used it and accepted by learned and un-learned people when discussing gambling.

I have looked around the internet and books in my library and many people use the term "law of averages' interchangably with "law of large Numbers' . Everybody understands what it means to covey.

So Guido, do I believe this guy...who has considerably more credentials than you in the area of mathematics and gambling....or do I side with you.

Mike Orkin ....CSU Hayward

Statistics Department

--------------------------------------------------------------------------------
Michael Orkin, Ph.D., U.C. Berkeley, 1971. Currently Associate Dean of the School of Science Professor of Statistics
--------------------------------------------------------------------------------
Areas of Interest: Probability; Game Theory; Gambling; Statistics and the Law; Multi-Media Educational Technologies.
Published Articles: Can You Win?, W.H. Freeman and Co., New York, 1991.

and other articles on probability and game theory .

Professional Activities: Appears regularly on television and radio shows to discuss gambling odds; Hosts weekly TV interview show on CSUH cable TV.

in his book "What are the Odds" he specifically uses the term "law of averages" on many occasions( not law of large numbers).......also, when he spoke before this audience, he used the term "law of averages at least 3 times. see here:

Many dictionaries label Bernoulli's theorum as " the law of averages"

http://dictionary.reference.com/browse/bernoulli's+theorem

It appears strongly, the term "law of averages" is accepted in the mainstream lexicon as well as by PhD's.

777

#112
13. falcon, Mar 9, 2012

### falcon Member

Joined:
Jan 17, 2012
Messages:
420
6
The law of averages says that in a large number of trials, the actual number of occurences will be close to the average number of From the below picture is a quick random walk data set.
The delta column is the difference between heads and tails.
It kept growing as the # of trials increased. Most times it does.
But the actual percentage of heads per trials went down each time, very close the the expected 50% value.

falcon and I were discussing this last night at the Craps table while we were both hitting on the hottest cocktail waitress ever!
She hit us both, but it was worth the price of admission.
Falcon, you awake yet![/quote]

You know, being an old short fat dumpy dude (that is me), I thought you had scored. Your personality and sense of humor really captivated her. She kept coming back for more intellect and wit. She just could not stay away.

I need to ask you on the coin flipping cart. Did the coins hit the floor or were they caught? There is a difference, and results can be skewed.

falcon

#113
14. slorch, Mar 9, 2012

### slorch Member

Joined:
Feb 16, 2012
Messages:
21
0
Location:
SE Texas
Looks like that trial was using a loaded coin...

In no series of flips did Tails outnumber Heads?

#114
15. falcon, Mar 9, 2012

### falcon Member

Joined:
Jan 17, 2012
Messages:
420
6
[q=MS]: Actually Basic Sevens beat me to the punch on this one. He taught a college level class on craps, complete with student manual which he wrote. The class included an exercise which had the students roll dice 144 times IIRC and record and submit the resulting distribution. Again IIRC the purpose of the exercise was to show that while few if any of the 144-roll sets were perfectly distributed, the actual distributions tended to be clustered around that perfect distribution.

A similar situation exists with the Perfect 1980. The probability of 1980 actual rolls of the dice being distributed exactly as shown in the spreadsheet I posted previously is very small, but of the millions of possible such distributions it is the most likely. Furthermore, the probability of getting any distribution of 1980 actual rolls decreases as the difference between that distribution and the perfect one increases. IOW the Perfect 1980 gives the expected value, the average distribution if you will, of a large number of 1980-roll sets. In any event, to address your complaint that the Perfect 1980 is inapplicable to this analysis because the chances of it being obtained are remote, I have redone the spreadsheet using probabilities. To no one else’s surprise, the results are the same: a straight Pass Line bettor taking no odds is expected to lose less than the player who places the point after it is established.[/quote]

Since we have Basic 777's here, could you answer some questions related to your classes and the tossing experimentations?

When your students ran the tossing experiment, did they use two different colored dice?

How many students were in a pod or group doing the tossing and recording the data? Did each do tossing and data recordings?

Do to the possible length of time involved, was fatigue ever a factor so as to skew data collection and were there more than one individuals involved in independently recording the same data so as to validate the tosses with result checking of the two or more data sheets done by a third party?

If the dice were the same color, how could there possibly be any validity to the data when point differentiation can not be actual and pure?

Because of the "millions of possible such distributions," why should anyone rely on computer generated "reliabilities," "expected outcomes," and virtual dice tosses using the law of large numbers when the reality of play provides none of that?

And finally, what are the actual odds of producing the "perfect 1980"?

I also have no argument with the factual assertion of straight PL bettors taking no odds losing LESS than the player who places the point after it is established. That fact was never in my assertions; anyones simple math would show that to be true.

falcon

#115
16. basicstrategy777, Mar 9, 2012

### basicstrategy777 Member

Joined:
Jul 4, 2010
Messages:
9,227
8,422
Gender:
Male
Location:
CT.
One piece of homework was for each student to roll regulation casino dice 144 times ( 36 x 4 ) and record each roll.

The idea was to show how disparate each roll could be RE: the numbers showing.....they would clump ...they would hide...

At the end of 144 rolls they would add up how many times each number showed and write it next to the 2,3,4,5,6,7,8,9,10,11,12.

The chart in the textbook would show how many times each number should have shown per predicted normal distribution...

2-4 times...3-8 times...4-12 times...5-16 times...6-20 times...7-24 times....8-20 times...6-16 times...10-12 times...11-8 times...12-4 times

This exercise showed the students something very valuable........i.e. that any number could show up at any time and that over time each number would show close ( sometimes exactly) to the number of times probability 'predicted' they would show.

777

#116
17. kaysirtap, Mar 9, 2012

### kaysirtap Member

Joined:
Nov 2, 2011
Messages:
483
22
Falcon, just cut to the chase already. What are you claiming this time?

#117
18. guido1, Mar 9, 2012

### guido1 Member

Joined:
Mar 12, 2011
Messages:
88
1
Location:
Newport Beach, CA
Sure it is used everywhere, just like a shooter "craps out" is used by almost all craps players when he rolls a 7 while a point is established.

It matters to me. 100% truth that is.

I like facts that are 100% accurate.
Not 92.67834% accurate like Enron accounting or US Fed Government Labor statistics.
or 18.435% accurate fact statements by Falcon.(margin of error: +/- 5.555%)

http://en.wikipedia.org/wiki/D'Alembert_system
"For example, if a fair coin is tossed repeatedly and tails comes up a larger number of times than is expected, a gambler may incorrectly believe that this means that heads is more likely in future tosses. Such an expectation could be mistakenly referred to as being due, and it probably arises from everyday experiences with nonrandom events (such as when a scheduled train is late, where it can be expected that it has a greater chance of arriving the later it gets). This is an informal fallacy. It is also known colloquially as the law of averages."
So, The Law of Averages IS an informal fallacy.
"What is true instead are the law of large numbers – in the long term, averages of independent trials will tend to approach the expected value, even though individual trials are independent – and regression toward the mean, namely that following a rare extreme event (say, a run of 10 heads), the next event is likely to be less extreme (the next run of heads is likely to be less than 10), simply because extreme events are rare."

Stewart N. Ethier: The Doctrine of Chances. Probabilistic Aspects of Gambling. 2010
page 315

"The d’Alembert system is named for Jean le Rond d’Alembert [1717–1783],
a French mathematician.
In 1744 d’Alembert published TraitÂ´e de l’Â´equilibre et
du mouvement des fluides, containing his “Theory of Equilibrium.”
It seems likely that he was therefore credited with this betting system, which relies
on the “Theory of Equilibrium” in probability, the belief (erroneous unless
p = 1/2) that wins and losses must eventually equalize."

"The term “conservation of fairness” comes from Dubins and Savage (1976,
p. 5), but the concept is much older. Theorem 8.2.3 is from Billingsley (1995,
Theorem 7.1) but dates back to Doob (1936), and it debunks the Monte
Carlo fallacy (so-named by Maxim 1904, pp. 36, 46), which, as noted in
Section 1.7, is the belief that a series of failures makes a success more likely;
the law of averages is often improperly cited as justification
."

page17
Understanding Probability
Chance Rules in Everyday Life
Second Edition
HENK TIJMS
Vrije University
"For example, suppose a fair coin is tossed 100 times, resulting in heads
60 times. In the next 100 tosses, the absolute difference between the numbers
of heads and tails can increase, whereas the relative difference declines.
This
would be the case, for example, if the next 100 tosses were to result in heads
51 times. In the long run, “local clusters” of heads or tails are absorbed by the
average.
It is certain that the relative frequencies of heads and tails will be the
same over the long run.
There is simply no law of averages for the absolute difference
between the numbers of heads and tails.
Indeed, the absolute difference
between the numbers of heads and tails tends to become larger as the number of
tosses increases. This surprising fact can be convincingly demonstrated using
computer simulation."

From your many posts, you believe that events are due after they have not shown for a certain length of time.
"Actual outcomes will be close to actual averages" is the motif of many of your posts.

This is a false belief. That is what The Law of Averages is. False.
Period.

Homework time

added: any "expert" that uses Law of Averages interchangeably with Law of Large Numbers as the same law is not any kind of expert at all.
One is a Law the other is a Fallacy.

#118
19. guido1, Mar 9, 2012

### guido1 Member

Joined:
Mar 12, 2011
Messages:
88
1
Location:
Newport Beach, CA
Not! In the dream, she liked you more. I could see your right hand was on the chips in your rack but I could not see your left. I think see liked where it was. You "old short fat dumpy dudes" have all the luck!
Neither.
The point of the results was as the number of trials increased, the relative frequency of both H and T converged towards 50% even at the same time the absolute difference between the 2 was increasing.
This is the Law of Large Numbers in action.

#119
20. guido1, Mar 9, 2012

### guido1 Member

Joined:
Mar 12, 2011
Messages:
88
1
Location:
Newport Beach, CA
Your observation or intuition has let you down.
This was an exact 50/50 experiment.

The data points were selected for table display only.
In the actual experiment, Tails was ahead for many of the flips. Especially during the 410 and 738th flip.
It just so happened that it does not show that in the table.
The graph of the walk shows it. But that is too much work to change the form and upload the photo.

I am now on spring break for the rest of the year!

homework time is to read about the Arcsine Law and Random Walks!

Falcon and I should discuss that later. Maybe in May 2012 when I will be in Tunica.

#120