Forums » Backgammon
Fixed dice
Replies: 96   Pages: 7   Most recent reply posted by: OPERATOR (Feb 23, 2012 8:03 AM)
Welcome, Guest
Login
Guest Settings
 Reply to this Topic  Post New Topic  Search Forum
Previous Topic
  Back to Topic List
Next Topic
Replies: 96   Pages: 7   [ Go: « 1 2 3 4 5 6 7 » ]
DEVELOPER

Posts: 513
Registered: Sep, 2004

Re: Posted: Feb 18, 2014 3:16 PM
Reply
"C.
in real game each games " Dice " would be justified by its own, like, if you got some 3-1 1-5 2-3 1-2 , then in the rest the probability of have a 1 or 3 is lower than others. it is called conditional possibility. Yet in the game the otherwise is seen !!!"

In order to clear up "conditional possibility":

"What is the probability of a 6 IF you have got a 6 previously"

It's 1/6

"What is the probability of a 6 IF you have got two 6's previously in a row"

It's 1/6

"What is the probability of a 6 IF you have got 10 6's previously in a row"

It's 1/6

"What is the probability of a 6 IF you have got 10000000000000 6's previously in a row"

It's 1/6

So, "in the rest the probability of have a 1 or 3 is " THE SAME.

You may want to roll a virtual dice here:

http://www.random.org/dice/

nefrin

Posts: 21
Registered: Sep, 2010

Re: "Fixed dice" Posted: Feb 18, 2014 7:35 PM
Reply
Dear Flyordie
Sorry for my bad English
And sorry not to be accurate :
A :
By X-6 I didn't meant straight 6 on each side, any special pair of number with a 6 like :
The probability of [ 1 & 6 or 6 & 1 ] simply it is 2 pair in 36 pair of numbers. That makes the possibility of 1/18.
.
but for any pair of [ any number & 6 or 6 & any number ]
it is 11/36. ( still near and below 1/3 ).
.
& i wanted to say it happens with more frequencies " IN RELATIVE TO GAME PLAY AND SITUATION"
B :
It is true , shot the number 6 you need one on these :
{ 2-2 , 3-3 , 5-1 or 1-5 , 4-2 or 2-4 + any 6 }
1/36 + 1/36 + 2/36 + 2/36 + 11/36 = 17/36 , it is near 1/2 and still much higher comparing to 5/36 for hitting " 9 ".
and i think it is still more repeated than it should be " ACCORDING TO GAME PLAY for hitting a sole stone. "
C :
SGF would not be useless :
2 times 3-5 means 1/18 * 1/18 = 1/324.
3 times pair of 1-1 is alike 1/36 * 1/36 * 1/36 = 1/46656.
if this set of numbers happens just once that would be fine but what you call it when you see these things frequently ????????
.
My numbers maybe have not been exact but the logic behind them is true.
How can i collect may game records ??
That is why i believe SGF button is essential.
So every one can see what things are really happening.

nefrin

Posts: 21
Registered: Sep, 2010

Re: "Fixed dice" Posted: Feb 18, 2014 8:32 PM
Reply
Dear DEVELOPER
Please, imagine we are playing dice, with just 1 dice.
We roll 2 times and we bet on a " double 6 ".
1. What is the possibility of having two 6 in the set ???
.
2. Then what is the possibility of having number 6 in 2nd roll
Just IF you had the number 6 in the 1st roll ???
.
3. Ok , then let me bet you wont get a 6 after 10 rolls which in all you have already had 6 ?
.
&. And now we rolls 1000 times. As it is wished by you.
If you got 500 roll of 6 , then in the rest the frequency of 6 must be much lower, otherwise we wont see an even distribution of numbers.
& by the bigger numbers we should see a better distribution.
.
It is simple and straight.


Battle Royale

Posts: 45
Registered: Jan, 2007

Re: "Fixed dice" Posted: Feb 19, 2014 2:41 AM
Reply
"1. Lots of times it begins with 3-1 2-4 6-1 3-5 2-4."

Feel free to record your games again. If any of your games will start with these five pairs in this order, you will get eternal VIP status.

Best comment yet

Genius!!

FlyOrDie

Posts: 457
Registered: Nov, 2001

Re: "Fixed dice" Posted: Feb 19, 2014 8:28 AM
Reply
"3 times pair of 1-1 is alike 1/36 * 1/36 * 1/36 = 1/46656."

that is true if you roll only 3 times.

During a backgammon match there are several rolls, and there are usually maybe 4-5-6 doubles per player per match. Let's suppose there are 10 all together.

If you have a double already, chances are 1/36 that you will get the same in the next roll, and 1/36 * 1/36 (1/1296) that you will get it in the next two rolls, making it three in a row.

So we have 1/1296 chance for every double that happens ~10 times in a match, which gives that there is a ~1/130 chance that you will see the same double three times in a row during a match, as opposed to 1/46656, which is not an insignificant difference.

In peak hours, there are ~200 ongoing matches simultaneously on our server, so probably one or two of them has the "3 same doubles in a row" event, which you classified as "cannot be random in no way".

FlyOrDie

Posts: 457
Registered: Nov, 2001

Re: Posted: Feb 19, 2014 9:03 AM
Reply
"3. Ok , then let me bet you wont get a 6 after 10 rolls which in all you have already had 6 ?"

"If you got 500 roll of 6 , then in the rest the frequency of 6 must be much lower, otherwise we wont see an even distribution of numbers."

random does not mean "even distribution of numbers" in the short run.

You have the same 1/6 probability of rolling a 6, even after a straight 500 roll of 6's.
The probability of the result of a dice roll is independent of any previous roll.

We "expect" the outcome of consecutive dice rolls to alternate between the six numbers much more than they really do. We easily mistake long
series of consecutive results, or clusters, as patterns when they are in fact truly random. To the untrained eye, this makes random results appear to be 'fake', and fake results seem more 'real'.

It is a common activity among math teachers when teaching students probability theory to ask them to flip coins 100 times and record the results. They can usually easily tell who 'faked' the results, and whose results are 'real'. The real results will usually contain long series of consecutive results, 5, 6 maybe even 7 consecutive heads or tails. 'Fake' results will contain only series of two or three consecutive heads or tails.

5-6-7 consecutive heads might seem unrealistic for the untrained eye.

For a dice roll, the numbers are 'a bit' different, as there are 6 possible outcome (or 36 for double dice), but getting the same roll 3 times in a row is definitely not impossible, and will happen 'frequently' if you play many games regularly.

DEVELOPER

Posts: 513
Registered: Sep, 2004

Re: "Fixed dice" Posted: Feb 19, 2014 9:38 AM
Reply
> Dear DEVELOPER
> Please, imagine we are playing dice, with just 1 dice.
> We roll 2 times and we bet on a " double 6 ".
> 1. What is the possibility of having two 6 in the set ???
> .


1/36

> 2. Then what is the possibility of having number 6 in 2nd roll
> Just IF you had the number 6 in the 1st roll ???
> .


1/6


> 3. Ok , then let me bet you wont get a 6 after 10 rolls which in all you have already had 6 ?
> .
> &. And now we rolls 1000 times. As it is wished by you.
> If you got 500 roll of 6 , then "in the rest the frequency of 6 must be much lower, otherwise we wont see an even distribution of numbers.
> & by the bigger numbers we should see a better distribution.
> .
> It is simple and straight.


NO. "in the rest the frequency of 6 " will be NOT lower UNLESS it's a fake random series.
Sorry, but you don't understand conditional probability and random event independence, if you truly believe the opposite.

The probability of the event "500 rolls of 6 in a row" is 1/6^500 (~8.4e-390)
The probability of the event "501 rolls of 6 in a row" is 1/6^501 (~1.4e-390)
However, the probability of a 6 AFTER "500 rolls of 6 in a row" is 1/6.
You may classify the probability of 1.4e-390 as impossible. However, if you have already rolled 500 6's in a row, then you are just experiencing a very rare event with a probability of 8.4e-390. So, if you roll another 6 after 500 6's that's not a miracle. It will happen one time out of 6 on average: if you roll 500 6's in a row many-many times then you will experience that you will roll another 6 in 1/6 part of the cases.




nefrin

Posts: 21
Registered: Sep, 2010

Re: Posted: Feb 19, 2014 4:58 PM
Reply
Dear DEVELOPER
Ok , Lets Just forget the conditional probability and so on and
Please think of Distributions.
1_ if there are 100 rolls, how many 6 do you expect to have in all ???
and now,
2_ think you already have had 16 roll of 6 in the first 50,
What the possibility to have another 16 in other 50 would be ???
.
3_Think of natural distribution , what would be the possibility of something happens with the 2x of times in a set of dice ?
.
The thing i am talking about is that Maybe " According to Special Situation's " in the game the frequencies are not distributed evenly while in the whole it is adjusted to be normal.
.
To Find out the truth there is just one way : to have game records.

nefrin

Posts: 21
Registered: Sep, 2010

Re: Posted: Feb 19, 2014 5:08 PM
Reply
Dear Battle Royal
There are 36 pairs of numbers in total { 6 * 6 } , which i suppose the game starts more frequently with set of numbers : { 1-6, 6-1, 3-1, 1-3, 2-4, 4-2, 3-5, 5-3 }
And by that i mean it is more than 10/36.
.
Im sure in my past games iv seen that effect , but to find out that SGF of the games is needed. i offer the site puts that option in the games.

nefrin

Posts: 21
Registered: Sep, 2010

Re: Posted: Feb 19, 2014 10:13 PM
Reply
@ Flyordie :
of course by coins it happens to see 5 times tail in a row , it is because there are only two options :
1 __ Head or tail, and each has 1/2 of probability and in a row you may have 1/2^5 = 1/32 , which in 100 place ( to have a such a set in 100 throw makes it highly possible.
BUT even 1/130 is not likely to happen many times, which i feel i does on this site.
( Thinking of 60 rolls of dice, 6 set of pairs [1-1, 2-2,...], and 1/6 their probability ---> 1/36*1/36*1/36*60*6 = 1/130 ).
.
2__ There also other pairs happen frequently / pair of pairs :
( 60 rolls, 1/18*1/18 for each pair, 15 special pairs )
1 time of a special pair : 1/18 * 1/18 * 60 = 60 / 324.
2 times happens : [1/18^2]*30
N times : [1/18^2] * 60 * N
N times for any pair : [1/18^2] * 60 * N * 15.
& for triple of the pairs : [ 1/18 ^3] * 60 * N * 15.
Of course it seems legitimate to have such pairs in a game ,
But please follow to num 3 :
.
3__ Supposedly , the site generates Lots of " Random " numbers based on a " Randomness Extractor " , using some entropy or something, giving a " Uniform distribution (discrete) " ...
and Bla, Bla, Bla :
http://en.wikipedia.or g/wiki/Randomness_extractor
http://en.wi kipedia.org/wiki/Uniform_distribution_(discrete)
But still they are things remaining :
They way you distribute dice got some problems, they are patchy or granulated.
I mean :
there is normal to witness fluctuation in random numbers, but the way this happens is different, the number and fluctuations are inevitable but the greater they get the bigger sample you need to have an even " uniform distribution " ,
< alike of comparing a harsh noise to fair noise, with harsher noise you have greater wave legnts> ,
.
In the Game with real dice It is much finer and more even that of this site , So numbers are distributed better, and gameplay is much more enjoy-full.
Moreover IN SMALL SAMPLE NUMBERS it is a wast difference BUT STILL IN LARG SCALE SAMPLES you loose to see the difference.
So, not being enough to examine large scale samples of the site, one also should examine the fluctuations.
< like comparing sand to dust from a mile a way versus microscope !!!>
I think this why lots of times Dice sets get packed :
Fo example please compare these 2 sets :
[ 2-2, 2-1, 3-5, 6-6, 1-1 ] & [ 2-1, 2-3, 6-1, 1-5, 2-6 ].
as you see taking any pair as a special one [ 1-6 & not 6-1] each set has the same probability in 5 roll, but the first one is kind of odd in real game play.
ALL i mean is total randomness in the whole site is not enough :
A. Hence any game should be fitted and adjusted even.
B. The numbers doesn't get clumped.s
C. & that they are ways to hide inside these fluctuations, like hitting a sole stone if needed.
But if there was a way to get the frequencies of hit a sole stone "Accordingly to a Situation " then we could see what really is going on.
.
.
With best Regards.



FlyOrDie

Posts: 457
Registered: Nov, 2001

Re: "Fixed dice" Posted: Feb 20, 2014 9:49 AM
Reply
There is not much point in continuing this discussion if you cannot accept basic probability theory facts.

I can only repeat myself: if you see too many "four consecutive 1-1 or 4-4 doubles", please get a screen capture utility, record your games, and show us. It will not take too much time if your conceptions are true.

Please post your further comments in the other thread opened by you.

Thank you.

DEVELOPER

Posts: 513
Registered: Sep, 2004

Re: "Fixed dice" Posted: Feb 20, 2014 1:20 PM
Reply
> Dear DEVELOPER
> Ok , Lets Just forget the conditional probability and so on and


You can't forget conditional probability. It's all about conditional probability.
Rolls of a dice are perfectly independent from each other. This means that the probability of an outcome is the same as the conditional probability of that outcome. Actually, that's the definition of event independence. The dice doesn't have a "memory". It does not know what values it rolled previously.

> Please think of Distributions.

> 1_ if there are 100 rolls, how many 6 do you expect to have in all ???
> and now,


Talking about distributions doesn't make any sense over 100 rolls. It makes sense over millions and billions of rolls. And even then, laws of probability tells you only the average of outcomes will CONVERGE to the probability. It doesn't say it will be EXACTLY the probability.

> 2_ think you already have had 16 roll of 6 in the first 50,
> What the possibility to have another 16 in other 50 would be ???


The first 50 rolls does not have ANY impact on the second set of rolls.
You don't have a deck of cards, so when you deal 4 aces in the first set you cannot deal another ace in the second set. Counting may work in blackjack, but not in dice games or roulette. Casino goers could tell a lot about this misconception.

> 3_Think of natural distribution , what would be the possibility of something happens with the 2x of times in a set of dice ?

I don't understand your question. What happens "with the 2x of times in a set of dice"? If you are asking if the exact same event happens once again, it's probability can be calculated. But a low probability won't guarantee it won't happen at all.

> The thing i am talking about is that Maybe " According to Special Situation's " in the game the frequencies are not distributed evenly while in the whole it is adjusted to be normal.

The rolls don't have to be evenly distributed over 20-30 rolls. Real dice rolls won't be either.
This is what you are saying: You examine 10-15 rolls and then make a statement: based on the previous rolls this or that next outcome is impossible. What you are doing is predicting the outcome of the next roll (ruling out outcomes is a prediction too, actually, everything is prediction that beats the probability of an outcome). If you are able to predict dice rolls then you are psychic and you can collect the $1million reward http://www.skepdic.com/randi.html

Let's simplify this to a single dice: you see the rolls 1,2,3,4,5 then you say the next roll must be a 6 because it has to be evenly distributed. Well, 1/6 of time it will be 6.

> To Find out the truth there is just one way : to have game records.

You are free to record your games.

Here on flyordie, there are 500k backgammon rounds in a month. In a single round there are 2x20-30+ rolls, say 50 on average. That means 25 million dice rolls in a month. Based on probabilities:
- that means approx. 694000 double 6's in a month
- that means approx. 19290 2-long double-6 series in a month
- that means approx. 536 3-long double-6 series in a month
- that means approx. 15 4-long double-6 series in a month
- that means approx. 0.4 5-long double-6 series in a month (which scales up to 5 over a year)
- that means approx. 2.5 5-long same-double series in a month
- that means approx. 3215 5-long any-double series in a month
- that means approx. 0.4 3-long any-double followed by a 3-long any-other-double series in a month (which scales up to 5 over a year)

So you record a game where you see a "5-long double-6" series or a "3-long any-double followed by a 3-long any-other-double series" and say that's NO WAY random because it has a probability of 1/(36^5) = 1/60.466.176 ???
Any of these should happen 5 times a year on average. Which means there can be a year when it won't happen at all, or it can happen 10+ times.

nefrin

Posts: 21
Registered: Sep, 2010

Re: "Fixed dice" Posted: Feb 23, 2014 12:12 PM
Reply
My argument is continued here :
http://forum. flyordie.com/thread.jsp?forum=48&thread=53707

BlackSmoke

Posts: 3
Registered: Sep, 2012

Re: Posted: Feb 23, 2014 4:35 PM
Reply
Can't FlyorDie see the discontent players regarding the fixed dice issue?

There is no doubt that there is a need to review the source code to the software. Had couple of players complained this issue would have been a personal issue but this has become a wide spread issue. And all participants are being made aware.

Why is the reluctance in Flyordie rectifying the software? Are you hiding something? Is there a purpose of stretching this issue? Fly ore die recently changed the allowed playing times an the way doubling the points. All is possible if the aim is to provide a honest game play.

I have recently visited other backgammon web sites an can tell you that this issue is not even visible in the game play. Why not get things right here?

nefrin

Posts: 21
Registered: Sep, 2010

Re: "Fixed dice" Posted: Feb 25, 2014 6:23 PM
Reply
Other Topics with the same subject :
http://forum.flyordie.com/thread.jsp?forum=48&thread=58985&tstart=30&tran ge=15
.
http://forum.flyordie.com/thread.jsp?forum=48&thread=37031&tstart=15&tran ge=15
.
http://forum.flyordie.com/thread.jsp?forum=48&thread=58489&tstart=45&tran ge=15
.
http://forum.flyordie.com/thread.jsp?forum=48&thread=58166&tstart=60&tran ge=15
.
http://forum.flyordie.com/thread.jsp?forum=48&thread=35720&tstart=75&tran ge=15

Replies: 96   Pages: 7   [ Go: « 1 2 3 4 5 6 7 » ]
Previous Topic
  Back to Topic List      Top of the page
Next Topic