sums de patternsPrev Topic Next Topic

New Topic New Poll

dr san

bgonÃ§alves
Brasil
Member #92,558
June 9, 2010
4,155 Posts
Offline

Nov 25, 2025, 4:20 am

Hello, you can also create this filter: for example, from 49/6 you can create 4 groups like this: pp = even/even = 02, 06, 24, etc. where the first digit is even and the second digit is also even, ii = odd/odd = 13, 19, 35... where the first digit is odd and the second is also odd, pi = even/odd = 23, 27, 49, where the first digit is even and the second digit is odd, and ip = odd/even 16, 18, 34... the first digit is odd and the second is even =

pp
ii
ip
pi How to use: Use the filter by calculating the minimum x maximum of each group.
dr san

bgonÃ§alves
Brasil
Member #92,558
June 9, 2010
4,155 Posts
Offline

Nov 25, 2025, 4:38 am

hello= powerball

02 04 06 08 20 22 24 26 28 40 42 44 46 48 60 62 64 66 68 =par/par
11 13 15 17 19 31 33 35 37 39 51 53 55 57 59 =ii
10 12 14 16 18 30 32 34 36 38 50 52 54 56 58 =ip
01 03 05 07 09 21 23 25 27 29 41 43 45 47 49 61 63 65 67 69=pi

draws=08 16 26 30 58

pp=08,26

ii=zero

ip=16,30,58

pi= zero
dr san

bgonÃ§alves
Brasil
Member #92,558
June 9, 2010
4,155 Posts
Offline

Nov 25, 2025, 4:44 am

Hello, so we have a pattern, okay, one of the 4 groups will have one or zero numbers in 100%, and another group will have 2 or 3 numbers, meaning one group is "cold" and the other is "hot" with one, two, or three numbers. How do we create this formula?
dr san

bgonÃ§alves
Brasil
Member #92,558
June 9, 2010
4,155 Posts
Offline

Nov 25, 2025, 5:02 am

hello=To create a formula that represents the condition where, when dividing the Powerball lottery into 4 groups according to the parity of the digits (PP, II, IP, PI), one of the groups has zero or one number and another group has 2 to 4 numbers, we can think of something like this:
1. Definitions of the groups:
• PP: first digit even, second digit even
• II: first digit odd, second digit odd
• IP: first digit odd, second digit even
• PI: first digit even, second digit odd
2. Let xPP,xII,xIP,xPI be the number of numbers drawn in each group.

3. The desired condition is:
• One group xgx_gxg (where g∈{PP,II,IP,PI}g \in \{PP, II, IP, PI\}g∈{PP,II,IP,PI}) has 000 or 111 numbers,
• Another group has between 222 and 444 numbers,
• The other two carry the remaining numbers.

4. To express this, we formulate:
∃g1,g2∈{PP,II,IP,PI},g1≠g2:(xg1=0 or 1)∧(2≤xg2≤4)\exists g_1, g_2 \in \{PP, II, IP, PI\}, g_1 \neq g_2 : (x_{g_1} = 0 \text{ or } 1) \land (2 \leq x_{g_2} \leq 4)∃g1,g2∈{PP,II,IP,PI},g1=g2:(xg1=0 or 1)∧(2≤xg2≤4)
Furthermore, the total sum must be equal to the total number of numbers drawn, which in Powerball is 555 numbers (not counting the ball). (Special):
xPP+xII+xIP+xPI=5x_{PP} + x_{II} + x_{IP} + x_{PI} = 5xPP+xII+xIP+xPI=5
5. Since one group "carries more" and another "carries less," the formula emphasizes that the numbers are unevenly distributed, but the total quantity is fixed.

6. If you want the formula to return true only under these conditions, you can structure the formula in code or mathematical logic checking the values of xgx_gxg with these restrictions.

In summary, the model is:
• For the counters xPP, xII, xIP, xPI, the number of numbers in each group,
• Valid if there is a group with 0 or 1 number,
• And there is a group with between 2 and 4 numbers,
• Totaling 5 numbers added in the four groups.

If needed, I can also help with the implementation of this in Excel, VBA, or another language.

If desired, I can detail the logic to filter or create functions that evaluate this distribution of numbers in the Powerball according to these criteria.

1. https://pypi.org/project/powerball/
2. https://lotterycodex.com/lottery-formula/
3. https://floridalottery.com/content/dam/floridalottery-web/files/game-reports/powerball/pwrball-odds.pdf
4. https://www.powerball.com/powerball-prize-chart
5. https://www.durangobill.com/PowerballOdds.html
6. https://lotterycodex.com/how-to-win-powerball/
7. https://www.palottery.pa.gov/Draw-Games/Powerball.aspx
8. http://www.senecanewsdaily.com/images/Powerball-Jackpot-Distribution.pdf
9. https://www.reddit.com/r/theydidthemath/comments/1n7wlit/what_am_i_missing_about_powerball_oddsrequest/
10. https://www.powerball.net/statistics
dr san

bgonÃ§alves
Brasil
Member #92,558
June 9, 2010
4,155 Posts
Offline

Nov 25, 2025, 5:06 am

Okay, it's proven that in 100% of any Powerball draw, one of the four groups will have a number or a zero. You can check it all you want; you can even rig any draw, and one of the four groups will have a one or a zero 100%. So where are those who defend randomness?
hypersoniq

Pennsylvania
United States
Member #1,340
April 6, 2003
2,940 Posts
Offline

Nov 25, 2025, 7:34 am

That actually was understandable because there was an example included!

For a spreadsheet you could use the following...

=IF(ISEVEN(QUOTIENT(A1,10)),"E","O") & IF(ISEVEN(MOD(A1,10)),"E","O")

or to stick with the Spanish par/impar...

=IF(ISEVEN(QUOTIENT(A1,10)),"P","I") & IF(ISEVEN(MOD(A1,10)),"P","I")

This should assume a single digit has a leading zero. It also assumes the first number of the draw history is in A1, my sheets start in B because the date is in A.

You are essentially classifying the numbers by membership in one of four groups.

I don't think anyone is trying to "defend" randomness, just admitting it is there and trying to work with it and understand it.

Have an EXCELlent day!
dr san

bgonÃ§alves
Brasil
Member #92,558
June 9, 2010
4,155 Posts
Offline

Nov 25, 2025, 2:12 pm

Hello, well, if at least one of the four groups has at least two numbers, then we have 571 pairs at 100%, or a pair, while closing 01 to 69 = we have 2,346 at 100% pairs in the Powerball, a reduction of 1,900 pairs with the same 100%, because it takes advantage of the pattern of one of the four groups having at least two numbers. This is the type of pattern we are looking for, of course, then there are the basic and dynamic filters, etc.
Ney1

Houston
United States
Member #255,587
September 22, 2025
74 Posts
Offline

Nov 25, 2025, 5:52 pm

What's the track record on your system/filter, Dr. San? You have a lot to say on the knowledge of the game but can you show us winning tickets? It doesn't have to be a jackpot hit....just show where 2 to 5 numbers have hit on a regular basis. It doesn't have to be consistently, either. Just show us some tickets where you got 2 to 5 numbers right. Can you produce that, Dr. San? 😏

I Don't Guess Numbers. I Create Smart Filters To Draft My Numbers.
dr san

bgonÃ§alves
Brasil
Member #92,558
June 9, 2010
4,155 Posts
Offline

Nov 26, 2025, 5:01 am

I'm still testing, I'm not betting yet.
Player649

Canada
Member #218,115
October 17, 2021
171 Posts
Offline

Nov 26, 2025, 2:46 pm

Quote: Originally posted by dr san on Nov 25, 2025
Okay, it's proven that in 100% of any Powerball draw, one of the four groups will have a number or a zero. You can check it all you want; you can even rig any draw, and one of the four groups will have a one or a zero 100%. So where are those who defend randomness?

I think dr san is coming to the same conclusion as I did, based on statistical evidence: that importance of randomness in lotteries is grossly overstates. There are many statistical means to overcome it. Let's start from the beginning.

Randomness in lotteries is based on either PRNGs or TRNGs. Great majority of them have one dominant characteristic: they produce patterns. Gazillions of them, all the time. If in doubt, ask GPT: can random number generators create patterns? I did, for curiosity rather than need, and the answer was very clear.

Thus patterns are always present in random number generation process. Moreover, they are detectable and measurable by even simple statistics.

Patterns exist on 3 different levels:
1) Single numbers (RNGs are designed to control this very tightly)
2) Group of numbers (RNG control can be much weaker, depending on group size)
3) Distribution = wheels (if you use all numbers the randomness has no effect)

Many think that you have to predict patterns for the future in order to be successful. Mission impossible to me. There is much better way.

As stated earlier patterns are detectable and measurable. Many of them last for a while. So the effort should go not into predicting the future (usually in vain) but into detecting patterns that are already on. Due to their abundance it's actually easy to do. Such pattern catchers can be applied to any level. You may need many of them, in fact as many as you can invent, there is no limit. Using only a few will have little effect if any. Automation, obviously, will be required. But that's why they invented Python.

From the scheme proposed by dr san you can create 1000s of group pattern catchers.
Apply 3 time frames to each catcher. Set up criteria a catcher has to meet for each time frame:
Long (40 draws) - the pattern is long lasting
Medium (30 draws) - the pattern in consistent, without major fluctuations (sudden dips)
Short (20 draws) - the pattern is still on

I opted for distribution level and currently use 1K of pattern catchers. Fully automated, obviously. I use the above time frames as well.

So much for "unbeatable" randomness.
Stat$talker

700 light yrs West of Milky Way Galaxy's Center
United States
Member #200,634
September 1, 2019
5,508 Posts
Offline

Nov 29, 2025, 10:32 am

Quote: Originally posted by Player649 on Nov 26, 2025
I think dr san is coming to the same conclusion as I did, based on statistical evidence: that importance of randomness in lotteries is grossly overstates. There are many statistical means to overcome it. Let's start from the beginning.

Randomness in lotteries is based on either PRNGs or TRNGs. Great majority of them have one dominant characteristic: they produce patterns. Gazillions of them, all the time. If in doubt, ask GPT: can random number generators create patterns? I did, for curiosity rather than need, and the answer was very clear.

Thus patterns are always present in random number generation process. Moreover, they are detectable and measurable by even simple statistics.

Patterns exist on 3 different levels:
1) Single numbers (RNGs are designed to control this very tightly)
2) Group of numbers (RNG control can be much weaker, depending on group size)
3) Distribution = wheels (if you use all numbers the randomness has no effect)

Many think that you have to predict patterns for the future in order to be successful. Mission impossible to me. There is much better way.

As stated earlier patterns are detectable and measurable. Many of them last for a while. So the effort should go not into predicting the future (usually in vain) but into detecting patterns that are already on. Due to their abundance it's actually easy to do. Such pattern catchers can be applied to any level. You may need many of them, in fact as many as you can invent, there is no limit. Using only a few will have little effect if any. Automation, obviously, will be required. But that's why they invented Python.

From the scheme proposed by dr san you can create 1000s of group pattern catchers.
Apply 3 time frames to each catcher. Set up criteria a catcher has to meet for each time frame:
Long (40 draws) - the pattern is long lasting
Medium (30 draws) - the pattern in consistent, without major fluctuations (sudden dips)
Short (20 draws) - the pattern is still on

I opted for distribution level and currently use 1K of pattern catchers. Fully automated, obviously. I use the above time frames as well.

So much for "unbeatable" randomness.

"So the effort should go not into predicting the future (usually in vain) but into detecting patterns that are already on."

A lottery theory, wrapped inside an enigma..

Playa649 haz a goose that drank wine,

Hiz Uncle chews tobacco from a Streetcar line,

The line broke, hiz Uncle got choked,

Insurance money lotto

Bought him a leaky row boat ....
Player649

Canada
Member #218,115
October 17, 2021
171 Posts
Offline

Nov 30, 2025, 12:20 am

Quote: Originally posted by Player649 on Nov 26, 2025
I think dr san is coming to the same conclusion as I did, based on statistical evidence: that importance of randomness in lotteries is grossly overstates. There are many statistical means to overcome it. Let's start from the beginning.

Randomness in lotteries is based on either PRNGs or TRNGs. Great majority of them have one dominant characteristic: they produce patterns. Gazillions of them, all the time. If in doubt, ask GPT: can random number generators create patterns? I did, for curiosity rather than need, and the answer was very clear.

Thus patterns are always present in random number generation process. Moreover, they are detectable and measurable by even simple statistics.

Patterns exist on 3 different levels:
1) Single numbers (RNGs are designed to control this very tightly)
2) Group of numbers (RNG control can be much weaker, depending on group size)
3) Distribution = wheels (if you use all numbers the randomness has no effect)

Many think that you have to predict patterns for the future in order to be successful. Mission impossible to me. There is much better way.

As stated earlier patterns are detectable and measurable. Many of them last for a while. So the effort should go not into predicting the future (usually in vain) but into detecting patterns that are already on. Due to their abundance it's actually easy to do. Such pattern catchers can be applied to any level. You may need many of them, in fact as many as you can invent, there is no limit. Using only a few will have little effect if any. Automation, obviously, will be required. But that's why they invented Python.

From the scheme proposed by dr san you can create 1000s of group pattern catchers.
Apply 3 time frames to each catcher. Set up criteria a catcher has to meet for each time frame:
Long (40 draws) - the pattern is long lasting
Medium (30 draws) - the pattern in consistent, without major fluctuations (sudden dips)
Short (20 draws) - the pattern is still on

I opted for distribution level and currently use 1K of pattern catchers. Fully automated, obviously. I use the above time frames as well.

So much for "unbeatable" randomness.

Here is a statistical example of wheels used as pattern catchers.

Based on latest data ending Nov 26/25.
Percentages represent ROIs.
All wheels have all 49 numbers in 10 tickets.
All processing is fully automated.
Results are computed in less than 10 seconds.

Loading all draws...
3531 draws loaded
Loading wheels for playing...
1000 wheels loaded

Backtracking 40 draws (for long term patterns)

Wheel-49N-10T(15) - 63%
Wheel-49N-10T(53) - 77%
Wheel-49N-10T(55) - 62%
Wheel-49N-10T(103) - 73%
Wheel-49N-10T(175) - 61%
Wheel-49N-10T(208) - 83%
Wheel-49N-10T(217) - 67%
Wheel-49N-10T(223) - 65%
Wheel-49N-10T(224) - 60%
Wheel-49N-10T(287) - 60%
Wheel-49N-10T(363) - 66%
Wheel-49N-10T(376) - 61%
Wheel-49N-10T(393) - 62%
Wheel-49N-10T(395) - 90%
Wheel-49N-10T(403) - 60%
Wheel-49N-10T(412) - 63%
Wheel-49N-10T(454) - 65%
Wheel-49N-10T(470) - 68%
Wheel-49N-10T(471) - 61%
Wheel-49N-10T(544) - 67%
Wheel-49N-10T(572) - 67%
Wheel-49N-10T(586) - 63%
Wheel-49N-10T(612) - 60%
Wheel-49N-10T(630) - 72%
Wheel-49N-10T(637) - 63%
Wheel-49N-10T(670) - 62%
Wheel-49N-10T(681) - 63%
Wheel-49N-10T(687) - 62%
Wheel-49N-10T(719) - 61%
Wheel-49N-10T(752) - 68%
Wheel-49N-10T(857) - 78%
Wheel-49N-10T(906) - 63%
Wheel-49N-10T(908) - 88%

Were there no detectable patterns all these percentages would be very similar with little variation.
Had I only few wheels most likely I would have caught nothing and easily come to conclusion patterns are undetectable.
Some of these returns are simply amazing, hard to believe they can happen but they are there.

Clear evidence that randomness is not as smooth and unpredictable as some want to see it.
Stat$talker

700 light yrs West of Milky Way Galaxy's Center
United States
Member #200,634
September 1, 2019
5,508 Posts
Offline

Nov 30, 2025, 12:31 am

Quote: Originally posted by Player649 on Nov 30, 2025
Here is a statistical example of wheels used as pattern catchers.

Based on latest data ending Nov 26/25.
Percentages represent ROIs.
All wheels have all 49 numbers in 10 tickets.
All processing is fully automated.
Results are computed in less than 10 seconds.

Loading all draws...
3531 draws loaded
Loading wheels for playing...
1000 wheels loaded

Backtracking 40 draws (for long term patterns)

Wheel-49N-10T(15) - 63%
Wheel-49N-10T(53) - 77%
Wheel-49N-10T(55) - 62%
Wheel-49N-10T(103) - 73%
Wheel-49N-10T(175) - 61%
Wheel-49N-10T(208) - 83%
Wheel-49N-10T(217) - 67%
Wheel-49N-10T(223) - 65%
Wheel-49N-10T(224) - 60%
Wheel-49N-10T(287) - 60%
Wheel-49N-10T(363) - 66%
Wheel-49N-10T(376) - 61%
Wheel-49N-10T(393) - 62%
Wheel-49N-10T(395) - 90%
Wheel-49N-10T(403) - 60%
Wheel-49N-10T(412) - 63%
Wheel-49N-10T(454) - 65%
Wheel-49N-10T(470) - 68%
Wheel-49N-10T(471) - 61%
Wheel-49N-10T(544) - 67%
Wheel-49N-10T(572) - 67%
Wheel-49N-10T(586) - 63%
Wheel-49N-10T(612) - 60%
Wheel-49N-10T(630) - 72%
Wheel-49N-10T(637) - 63%
Wheel-49N-10T(670) - 62%
Wheel-49N-10T(681) - 63%
Wheel-49N-10T(687) - 62%
Wheel-49N-10T(719) - 61%
Wheel-49N-10T(752) - 68%
Wheel-49N-10T(857) - 78%
Wheel-49N-10T(906) - 63%
Wheel-49N-10T(908) - 88%

Were there no detectable patterns all these percentages would be very similar with little variation.
Had I only few wheels most likely I would have caught nothing and easily come to conclusion patterns are undetectable.
Some of these returns are simply amazing, hard to believe they can happen but they are there.

Clear evidence that randomness is not as smooth and unpredictable as some want to see it.

"Here is a statistical example of wheels used"

BLACKBOARD BS..!!

Wheaa are YO posted lottery ticketz PROOVIN that dat crap werkz Playa649?.It'z uzzzeless if ya cain't pick CORRECT Numbaz.. which obviously YOU cain't do..!

It'z hard to sell a "Fly- Free Flyswatter
Cmoore50

United States
Member #194,171
December 3, 2018
621 Posts
Offline

Nov 30, 2025, 9:16 am

Quote: Originally posted by Player649 on Nov 30, 2025
Here is a statistical example of wheels used as pattern catchers.

Based on latest data ending Nov 26/25.
Percentages represent ROIs.
All wheels have all 49 numbers in 10 tickets.
All processing is fully automated.
Results are computed in less than 10 seconds.

Loading all draws...
3531 draws loaded
Loading wheels for playing...
1000 wheels loaded

Backtracking 40 draws (for long term patterns)

Wheel-49N-10T(15) - 63%
Wheel-49N-10T(53) - 77%
Wheel-49N-10T(55) - 62%
Wheel-49N-10T(103) - 73%
Wheel-49N-10T(175) - 61%
Wheel-49N-10T(208) - 83%
Wheel-49N-10T(217) - 67%
Wheel-49N-10T(223) - 65%
Wheel-49N-10T(224) - 60%
Wheel-49N-10T(287) - 60%
Wheel-49N-10T(363) - 66%
Wheel-49N-10T(376) - 61%
Wheel-49N-10T(393) - 62%
Wheel-49N-10T(395) - 90%
Wheel-49N-10T(403) - 60%
Wheel-49N-10T(412) - 63%
Wheel-49N-10T(454) - 65%
Wheel-49N-10T(470) - 68%
Wheel-49N-10T(471) - 61%
Wheel-49N-10T(544) - 67%
Wheel-49N-10T(572) - 67%
Wheel-49N-10T(586) - 63%
Wheel-49N-10T(612) - 60%
Wheel-49N-10T(630) - 72%
Wheel-49N-10T(637) - 63%
Wheel-49N-10T(670) - 62%
Wheel-49N-10T(681) - 63%
Wheel-49N-10T(687) - 62%
Wheel-49N-10T(719) - 61%
Wheel-49N-10T(752) - 68%
Wheel-49N-10T(857) - 78%
Wheel-49N-10T(906) - 63%
Wheel-49N-10T(908) - 88%

Were there no detectable patterns all these percentages would be very similar with little variation.
Had I only few wheels most likely I would have caught nothing and easily come to conclusion patterns are undetectable.
Some of these returns are simply amazing, hard to believe they can happen but they are there.

Clear evidence that randomness is not as smooth and unpredictable as some want to see it.

So you're getting back around 65% to 70% of every dollar spent in this example? If that's the case. That's darn good wheeling all numbers. How many numbers do you need to eliminate to get this to even?

GET IT DONE
Player649

Canada
Member #218,115
October 17, 2021
171 Posts
Offline

Dec 1, 2025, 1:01 am

Quote: Originally posted by Cmoore50 on Nov 30, 2025
So you're getting back around 65% to 70% of every dollar spent in this example? If that's the case. That's darn good wheeling all numbers. How many numbers do you need to eliminate to get this to even?

You are correct - percentages represent cash returns on a dollar e.g. 70% means 70 cent return. You play for only 30 cents.

I always play all numbers in all wheels. 40 draws backtracks generally do not reach above 80% due to fluctuations. To break even you need shorter term backtracks. 10 draws backtracks show many break evens.

Backtracking 20 draws (90% threshold)
Wheel-49N-10T(15) - 97%
Wheel-49N-10T(208) - 92%
Wheel-49N-10T(209) - 95%
Wheel-49N-10T(300) - 95%
Wheel-49N-10T(507) - 95%
Wheel-49N-10T(533) - 90%
Wheel-49N-10T(544) - 127%
Wheel-49N-10T(586) - 97%
Wheel-49N-10T(630) - 97%
Wheel-49N-10T(670) - 95%
Wheel-49N-10T(698) - 90%
Wheel-49N-10T(753) - 90%
Wheel-49N-10T(857) - 102%
Wheel-49N-10T(908) - 107%

Backtracking 10 draws (100% threshold)
Wheel-49N-10T(15) - 180%
Wheel-49N-10T(56) - 114%
Wheel-49N-10T(95) - 105%
Wheel-49N-10T(98) - 100%
Wheel-49N-10T(138) - 114%
Wheel-49N-10T(164) - 120%
Wheel-49N-10T(177) - 100%
Wheel-49N-10T(187) - 100%
Wheel-49N-10T(194) - 100%
Wheel-49N-10T(197) - 100%
Wheel-49N-10T(207) - 110%
Wheel-49N-10T(208) - 185%
Wheel-49N-10T(209) - 110%
Wheel-49N-10T(232) - 105%
Wheel-49N-10T(254) - 114%
Wheel-49N-10T(270) - 114%
Wheel-49N-10T(286) - 105%
Wheel-49N-10T(300) - 100%
Wheel-49N-10T(314) - 100%
Wheel-49N-10T(395) - 155%
Wheel-49N-10T(411) - 125%
Wheel-49N-10T(454) - 120%
Wheel-49N-10T(476) - 100%
Wheel-49N-10T(505) - 110%
Wheel-49N-10T(528) - 100%
Wheel-49N-10T(544) - 170%
Wheel-49N-10T(592) - 105%
Wheel-49N-10T(612) - 110%
Wheel-49N-10T(637) - 135%
Wheel-49N-10T(649) - 110%
Wheel-49N-10T(651) - 110%
Wheel-49N-10T(682) - 105%
Wheel-49N-10T(687) - 110%
Wheel-49N-10T(706) - 110%
Wheel-49N-10T(722) - 100%
Wheel-49N-10T(724) - 105%
Wheel-49N-10T(761) - 110%
Wheel-49N-10T(832) - 120%
Wheel-49N-10T(857) - 105%
Wheel-49N-10T(882) - 110%
Wheel-49N-10T(908) - 100%
Wheel-49N-10T(929) - 100%
Wheel-49N-10T(973) - 100%
Wheel-49N-10T(997) - 100%

New Topic New Poll

Subscribe to this topic

sums de patternsPrev TopicNext Topic

sums de patternsPrev Topic Next Topic