Lisbon Portugal Member #167185 June 29, 2015 22 Posts Offline

Posted: July 4, 2015, 6:06 am - IP Logged

Hi

Once again: How can players use this knowledge in order to beat the odds in POWERBALL?

This kind of work represents a great job and effort coming from you but this was already done thousands of times by mathematicians and enthusiasts...

And really, considering prediction as a sort of war against reality I do not see how can you defeat and open breaches in reality`s walls with these kind of weapons. Reality is made of fractal mathematics (and this is already an approach) and most of the people are trying to fight it with Euclidian mathematics and bell shape curves, arithmetic, geometric, harmonic averages, etc.

These are too much weak weapons. Like I said once is like fighting a machine gun with a plastic revolver or a cancer with tea...

Reality does not calculates the average. Reality is simply... real...

So, predictive algorithms must be more realistic...

Once again: How can players use this knowledge in order to beat the odds in POWERBALL?

This kind of work represents a great job and effort coming from you but this was already done thousands of times by mathematicians and enthusiasts...

And really, considering prediction as a sort of war against reality I do not see how can you defeat and open breaches in reality`s walls with these kind of weapons. Reality is made of fractal mathematics (and this is already an approach) and most of the people are trying to fight it with Euclidian mathematics and bell shape curves, arithmetic, geometric, harmonic averages, etc.

These are too much weak weapons. Like I said once is like fighting a machine gun with a plastic revolver or a cancer with tea...

Reality does not calculates the average. Reality is simply... real...

So, predictive algorithms must be more realistic...

RENTAP

We're reposting your reply from the topic How to simulate a distribution and use the lottery's own randomness against itself since this is the relevant topic.

We will get back to this a little later.

We are testing and finalizing a =DisimulateNexus(Numbers, Distribution, N_{1}, N_{2}) function for the other topic.

You have made a great statistical work. But the question is: what is this for? Can you build a predictive algorithm from this knowledge?

Here's an example: Take the Megaball, once again.

If the common player bet 1 number he will get 0.0666 hits per draw (1/15 = 0.0666).This is the REFERENCE or the so called Theoretical Average (TA). My question is: using this system (3 sums) how many hits can you achieve per draw?

If after 100 samples, for example, you get 0.0666 hits a draw, then there is no advantage to follow it;

If you get a higher value means you can beat the TA so, in this case you are predicting;

finally,If you get a lower value you are also predicting because, in this case, you may bet against the algorithm.

However, I fear it is not possible to predict using this knowledge because there is a synchronizing problem, I mean, although similarities in simulated and real patterns, you need to synchronize your model with the reality of the game. Several years ago when I started to develop predictive algorithms I lost a few hours or even days (I can`t remember exactly) with part of this same exercise you did and I quickly came to dead ends. However, you went a little bit further introducing simulated information.

But, I insist: is that enough to beat the odds, i.e. to get more than 0,0666 matches per drawing? If so, the system works.

But let`s see some obvious problems present in the 3 sums technique.

If the first 2 numbers of the sum are 1 + 1 + ... = 24 and knowing that the ideal sum of Megaball is 24, it means that the third number would be 22. Now, the ball 22 does not exist!!!

In the case of the first 2 numbers are 15+13+…= 24 the third number would be -4.However, Megaball does not contain negative numbers.Therefore, a brief review finds that there are dozens of black cases where the system does not work, because there are no enough numbers.

But the real problem is to construct predictive algorithms based on Euclidian maths when the reality is fractal. We can not fight fractal maths with Euclidian ou Gaussian maths. That is the reason why is so difficult to predict...

That is like fighting a war tank with a plastic revolver.

Regards

RENTAP

Presented 'AS IS' and for Entertainment Purposes Only. Any gain or loss is your responsibility. Use at your own risk.

Order is a Subset of Chaos Knowledge is Beyond Belief Wisdom is Not Censored Douglas Paul Smallish Jehocifer

West Concord, MN United States Member #21 December 7, 2001 4087 Posts Online

Posted: July 6, 2015, 7:54 pm - IP Logged

Quote: Originally posted by JADELottery on July 4, 2015

We're reposting your reply from the topic How to simulate a distribution and use the lottery's own randomness against itself since this is the relevant topic.

We will get back to this a little later.

We are testing and finalizing a =DisimulateNexus(Numbers, Distribution, N_{1}, N_{2}) function for the other topic.

You have made a great statistical work. But the question is: what is this for? Can you build a predictive algorithm from this knowledge?

Here's an example: Take the Megaball, once again.

If the common player bet 1 number he will get 0.0666 hits per draw (1/15 = 0.0666).This is the REFERENCE or the so called Theoretical Average (TA). My question is: using this system (3 sums) how many hits can you achieve per draw?

If after 100 samples, for example, you get 0.0666 hits a draw, then there is no advantage to follow it;

If you get a higher value means you can beat the TA so, in this case you are predicting;

finally,If you get a lower value you are also predicting because, in this case, you may bet against the algorithm.

However, I fear it is not possible to predict using this knowledge because there is a synchronizing problem, I mean, although similarities in simulated and real patterns, you need to synchronize your model with the reality of the game. Several years ago when I started to develop predictive algorithms I lost a few hours or even days (I can`t remember exactly) with part of this same exercise you did and I quickly came to dead ends. However, you went a little bit further introducing simulated information.

But, I insist: is that enough to beat the odds, i.e. to get more than 0,0666 matches per drawing? If so, the system works.

But let`s see some obvious problems present in the 3 sums technique.

If the first 2 numbers of the sum are 1 + 1 + ... = 24 and knowing that the ideal sum of Megaball is 24, it means that the third number would be 22. Now, the ball 22 does not exist!!!

In the case of the first 2 numbers are 15+13+…= 24 the third number would be -4.However, Megaball does not contain negative numbers.Therefore, a brief review finds that there are dozens of black cases where the system does not work, because there are no enough numbers.

But the real problem is to construct predictive algorithms based on Euclidian maths when the reality is fractal. We can not fight fractal maths with Euclidian ou Gaussian maths. That is the reason why is so difficult to predict...

That is like fighting a war tank with a plastic revolver.

If the common player bet 1 number he will get 0.0666 hits per draw (1/15 = 0.0666). This is the REFERENCE or the so called Theoretical Average (TA). My question is: using this system (3 sums) how many hits can you achieve per draw?

If after 100 samples, for example, you get 0.0666 hits a draw, then there is no advantage to follow it;

If you get a higher value means you can beat the TA so, in this case you are predicting;

finally, If you get a lower value you are also predicting because, in this case, you may bet against the algorithm.

However, I fear it is not possible to predict using this knowledge because there is a synchronizing problem, I mean, although similarities in simulated and real patterns, you need to synchronize your model with the reality of the game. Several years ago when I started to develop predictive algorithms I lost a few hours or even days (I can`t remember exactly) with part of this same exercise you did and I quickly came to dead ends. However, you went a little bit further introducing simulated information.

But, I insist: is that enough to beat the odds, i.e. to get more than 0,0666 matches per drawing? If so, the system works.

Doing better than (1 / 15, 0.0666... or 6.67%) for the Megaball is an indication of a system or algorithm that works.

Let's look at the Theoretical Average (TA) using Excel's =Randbetween(Lo, Hi) as a prediction example, first.

The Randbetween is just a plain old random selector and is homogeneous throughout the numbers being selected from Lo to Hi.

This means the numbers tend to be selected without bias and over long sets of selections are nearly equal in frequency.

If we use Randbetween as our prediction, the Lo is 1 and the Hi is 15.

We ran a 1,500 sample set for each drawing on the Megaball using Randbetween(1, 15) and derived the hit percentage for each 1,500 sample sets.

Then, we repeated the same 1,500 sample set 3 more times for a total of 4 different 1,500 sample sets per draw.

Below is a chart with the frequency of the hit percentages.

The percentages are grouped into 1/3 % or 0.333...% increments.

The highest frequency hit percentage is where we would expect it to be, right at the 6.67% Theoretical Average (TA); we colored the 6.67% a darker shade.

Also, notice the hit percentages tend to cluster near the 6.67% TA with an average of 6.64%, a minimum of 4.80% and maximum of 9.27%.

Exactly 50% of the percentage hits are average and below; likewise, 50% are above average.

Now, let's look at a Three Sum Quantum Selector, the CombNexus(N, R, C, Z_{1}, Z_{2}).

A Three Sum Quantum Selector works on the principle of connecting the three aspects of Past, Present and Future in a unified sum and finding the Future by passing through a reverse function of a simple integrated probability.

This selector is biased on a function of an integrated sum and is not homogeneous throughout the numbers being selected.

We ran the same 1,500 sample sets per drawing 4 times.

The objective is to have a greater than 6.67% hit percentage.

Below we can see the hit percentages are spread out more, however, the peak hit percentage is now at 7.67%, a full 1% above the TA.

The average hit percentage is 6.87%

Also, the maximum extends out to near 18% at 17.73% and nearly double the TA maximum of 9.27%.

The chart is colored with Red below average, Green average and Blue above average.

58.1% of the total percentage hits are above average, Blue.

Presented 'AS IS' and for Entertainment Purposes Only. Any gain or loss is your responsibility. Use at your own risk.

Order is a Subset of Chaos Knowledge is Beyond Belief Wisdom is Not Censored Douglas Paul Smallish Jehocifer

These are good results, however I have 1 question:

You refer you used 1500 samples on your test. But since October 2013 until nowadays

we couldn`t have 1500 samples in Megaball. Does it mean that the samples are simulated

(random number generator) and not real?

In that case, do you think it is possible that this good performance may be replicated when

you use REAL RESULTS of Megaball and other REAL GAMES?

Is it possible for you to select a past period of 60 (or more) REAL drawings, for example,

and make a retrospective work and collect the real performance?

If so, you may say: look, in these past 60 drawings I was supposed to get 4 hits (60/15=4)

betting 1 Megaball, but, in fact, I got 5 hits, which means a hit performance of 0,083 hits per draw,

25% above TA (0,0667); or, betting a set of 5 Megaballs I was supposed to get

0,333 (5/15=0.333) hits per draw or 20 hits (in absolute counting)

but I got 23 hits which means I got 0,383 hits per draw, or +15% above the TA.

I put this question because I am afraid there is a gap between RNG drawings and physical drawings,

I mean, I think they don`t follow the same rules and principles. So, for a matter of safety and precision

I think, whenever possible, we must support the test and studies in real results or, at least, results provided

by the same “media” of the game we are studying.

RENTAP

It would be impossible to get a real 1500 drawings sample of either of the multi-states games since neither have ever had 1500 drawings without a matrix change. MegaMillions has only had 178 drawings since its last matrix change so it's either 178 or make up your own.

* you don't need to buy more tickets, just buy winning ones *

Lisbon Portugal Member #167185 June 29, 2015 22 Posts Offline

Posted: July 7, 2015, 6:12 pm - IP Logged

Hi I would never trust in a forecasting model (if that is what you intend to develop) regarding a physical game based on simulated computer results. It is not reliable. It is a lottery itself.

Nevertheless, 178 samples is already a significant amount of samples. In fact, with 35, 40 samples is already possible to detect a trend, that is, whether or not the model works.

Most of the people think randomness is similar either in physical games (natural randomness) as in number generators (artificial randomness) but my experience tells me it is not. First of all, because one is natural and the other is "fabricated" by a software. RENTAP

New Mexico United States Member #86099 January 29, 2010 12784 Posts Offline

Posted: August 2, 2015, 3:29 pm - IP Logged

Quote: Originally posted by JADELottery on July 4, 2015

We're reposting your reply from the topic How to simulate a distribution and use the lottery's own randomness against itself since this is the relevant topic.

We will get back to this a little later.

We are testing and finalizing a =DisimulateNexus(Numbers, Distribution, N_{1}, N_{2}) function for the other topic.

You have made a great statistical work. But the question is: what is this for? Can you build a predictive algorithm from this knowledge?

Here's an example: Take the Megaball, once again.

If the common player bet 1 number he will get 0.0666 hits per draw (1/15 = 0.0666).This is the REFERENCE or the so called Theoretical Average (TA). My question is: using this system (3 sums) how many hits can you achieve per draw?

If after 100 samples, for example, you get 0.0666 hits a draw, then there is no advantage to follow it;

If you get a higher value means you can beat the TA so, in this case you are predicting;

finally,If you get a lower value you are also predicting because, in this case, you may bet against the algorithm.

However, I fear it is not possible to predict using this knowledge because there is a synchronizing problem, I mean, although similarities in simulated and real patterns, you need to synchronize your model with the reality of the game. Several years ago when I started to develop predictive algorithms I lost a few hours or even days (I can`t remember exactly) with part of this same exercise you did and I quickly came to dead ends. However, you went a little bit further introducing simulated information.

But, I insist: is that enough to beat the odds, i.e. to get more than 0,0666 matches per drawing? If so, the system works.

But let`s see some obvious problems present in the 3 sums technique.

If the first 2 numbers of the sum are 1 + 1 + ... = 24 and knowing that the ideal sum of Megaball is 24, it means that the third number would be 22. Now, the ball 22 does not exist!!!

In the case of the first 2 numbers are 15+13+…= 24 the third number would be -4.However, Megaball does not contain negative numbers.Therefore, a brief review finds that there are dozens of black cases where the system does not work, because there are no enough numbers.

But the real problem is to construct predictive algorithms based on Euclidian maths when the reality is fractal. We can not fight fractal maths with Euclidian ou Gaussian maths. That is the reason why is so difficult to predict...

That is like fighting a war tank with a plastic revolver.