Open Discussion: Define/Refine Prediction

hypersoniq's avatar - binary

There is a never-ending supply of info on how to calculate odds, probabilities, wheels, filters etc.. but is there a specific, well-defined process or algorithm for prediction itself? If there is, I can't find one... perhaps as a group we can figure this out...

What are the tools? What's the math? where to begin?

I think that the best way to start is to have a clearly defined understanding of the pick 3 "problem"...

the way I see it...

1. overall odds are 1000:1 but each ball drawn is an isolated dvent, so in the big picture it's a matter of having 3 successful sequential predictions of a 1 in 10 game... (will be easy to apply anything learned to the 4 digit games this way)... all we need to do is come up with some way to predict one ball out of 10 (3 times... in this light, box hits don't count)

2. Nobody will ever really have the opportunity to study the apparatus that the states use,so we're more interested in modelling the RESULTS rather than the process...

3. the goal is to remain managaeable and affordable, with the ultimate goal being as few picks as possible (otherwise it's not really prediction, right?)

4. It's an open and free discussion, never resulting in a "new system" for sale... Of the posters, by the posters and for the posters...

5. it's more about the process than the actual picks, this way everyone can benefit.

Unlike a big jacpot game, we can all benefit from an open discussion of ideas...

anyone interested?

blue's avatar - bigears

it sounds ok so are you going to start this in motion i am sure the feed back will come in time. blue

Prometheus1's avatar - trace9


                              Would you like to nice game of Chess ?

             Start off learning the different pieces and how they move.

              When they wanted to make a computer that played chess, what did they do first ? Most folks don't even know the names of the pieces yet. 

                      We need to learn the game before we break the code.

              It all starts with math and ends with math and probability.       


 Sounds good to me but every state probably has a different formula.

 What us to work for me in my state does not work any more

Rick G's avatar - avatar 1766.jpg

Here is a mistake I think we are all making.  We approach the game in a logical fashion.  It is the antithesis of a logical game and should be played illogically.  Take the, hot digits, trends, due digits, etc. and turn all these 'logical' theories around.  Try doing the exact opposite of what you're doing and see if it works better.

hypersoniq's avatar - binary

Here's the basic concept...

take a string of numbers (representing a history of the FIRST digit drawn)

2, 4, 6, 8, 0, 2, 4, ???        That one's easy... 6 (+2) but how, mathematically, do we know that?


1,3,5,7,9,2,1,3,5,7,9,4,1,3,5,7,9,6,1,3,5,7,9, ????  step 1

1,3,5,7,9,2,1,3,5,7,9,4,1,3,5,7,9,6,1,3,5,7,9, ????  step 2, see the dmbedded pattern? gotta be an 8... but what formula could we plug numbers into to determine that? I don't know...


1,3,9,9,5,8,7,1,9,9 ????  what's next? and how do we know? what process could be used? (find out what's next at 7pm tonight in the PA evening 3 digit draw)

WIN  D's avatar - q05Q0

hypersonig, Sorry this may be off the subject... but if I were in Pa. I would play an even digit..24680 at the end of all those examples you gave. That digit 6 could come back tonight with a 2 . 

  Since you haven't had a 2 in the first position in 26 days......I'd favor him. Futher more your last 8 doubles were all Odd doubles.....Geez .. this goes back only 30 days .......I would be looking at Even doubles real hard ....your very P.G. in P.A. for all even everything ...LOL

  Sorry ...just couldn't help it   

hypersoniq's avatar - binary
Quote:Originally posted by WIN D on November 18, 2003

your very P.G. in P.A. for all even everything

6-6-0 good call!
hypersoniq's avatar - binary

I'm thinking a possible refinement of announcer numbers might help...

in the chain (being a sequential list of a particular position only, e.g. first digit, second digit etc...) ...

4,5,7,2,6,7,9,2,9,0,7,1,2,1,6,5,7,2,1,4,4,5,7... ???

looking at the 7 as announcer... has been followed by a 1, 2 and a 9, but taking it to the next level of refinement... what numbers announced the 7? 0, 5 and 6... but Twice, the 5 announced the 7 which then announced the 2... which would be my pick in this sequence (if it were real)...

Rick G's avatar - avatar 1766.jpg

Just a note here...

Since we're talking about PA,  WIN D was right on target with the all-even pair.  For the last couple weeks PA and IL numbers have been very similar.  One could have looked at IL and seen the 066 came in straight in IL Monday midday.

Based on WIN D's comments, I would think another double even pair is likely to show up again soon.  In IL Monday evening the number drawn was 088.  This combo might be worth watching for a few days.

Hypersoniq...your quest is a difficult one and I don't think there is an algorithm that could determine patterns such as the numbers drawn for PA that you posted.

The 'optical-to-mind computer' would probably be better at spotting the type of trends you describe, but you would have fewer opportunities and would it be more successful?

But as WIN D just demonstrated, the past is a good key to the future (i.e., prediction).


I don't know anything about math or programing and so therefore I probably should not say anything at all, but i think that there are an infinify number of possible prediction lines and that sequential number prediction doesn't have to be exactly

sequential you can jump and skip numbers, go on zig zag, criss cross, it can be vertical, horizontal, and horizontal + vertical.

By all this I mean that there are many avenues or roads or ways of linear number prediction, for example:

You have a history numbers with 3 digits:








You can take the 3 digits or numbers as being either just 1 big       

number or 3 smaller separate and independent numbers.

If taken as one big number it will be 785,345,805, and then you will try to predict the next number which will be 123 using the numbers that came before 123.

If taken as 3 separate independent numbers you can try predicting by columns such as:

You have 7,3,8,1,9,6 and you will try to predict the next number which will be 1 this is for the left column or the left position, for the middle would be: 8,4,0,2,8,7 and you would try to predict the next number which is another 7, and the same for the right position or column.

But there are many other ways of doing this,for example:

Let say that you take the oldest draw: 785 and go from left to right, 7,8,5 and then the next one 3,4,5 and also go from left to right, and then the next one, 8,0,5 and keep on going in the same way until you get to the last draw 17( ) and you try to predict the last number (4)

It would be like this: 7,8,5,3,4,5,8,0,5,1,2,3,9,8,4,6,7,4,1,7 and the number that you would try to predict is the last number "4"

For all this linear number predictions you need what I would call "A Linear Prediction Engine" or whatever you call it.

And you can also try predicting for some numbers diagonaly such as in a diagonaly tic tac toe fashion or criss cross or in any other way that you can think of including jumping like a horse and skipping some numbers in between,like skipping by ones all numbers or even by twos and in any given pattern or way {road} that you want or think of.

But they would all be linear number predictions regardless.

Do you understand what I mean? I know that it sounds crazy.

hypersoniq's avatar - binary

I'm going with the column approach, because they are independent numbers (3 machines, each containing the balls 0-9). If you look at all 3, then I guess the 3 digit would be considered a "replacement" drawing, because doubles and triples can and do happen.

Not that the interaction should be totally discounted, perhaps the same conditions that evoke a 5 from one machine may also indicate a 3 from the next, etc... but I'm hoping to get better results in the columns.

My son was doing a first grade homework paper last night, on it were a circle, square, triangle, circle square triangle and 3 blank lines with the instructions to continue the pattern... now later on, this gets harder to do when real numbers are involved, but from what I can recall, there was no "formula" for doing this... or is there?

hypersoniq's avatar - binary
Quote:Originally posted by Prometheus1 on November 17, 2003


                              It all starts with math and ends with math and probability.       

I see that term alot, but let's see a practical application. without some form of "weighting" based on previous history, each ball has a 1 in 10 chance of coming up. How does it get narrowed down?

I'm considering the announcer theory... look at the previous draw, for example 123... then look at what came after each number in that position... for example, 1 announces a 9 more than any other number... BUT that still doesn't help when you have several thousand results... perhaps looking at the number drawn before the 123 (say, 567) may help... then the question would be.. what number most follows the 1 in that position WHEN THE 1 was preceded by a 5...

perhaps that could help add some weight and focus to what will most likely be the next number (or at least narrow down the choices).... ]

any comments/corrections/additions to this possible theory? as I have well over 9,000 results to sift thru, this will take awhile before I can take a crack at some guesses for PA... but if I can figure out a way to automate the process in excel, I'll post the macro....

JAP69's avatar - scene sunovermountains.jpg

Create or use a software that tracks roots and sums.
Treat each chamber as a separate game and enter the drawn digit in the order of being drawn. When you search the root or sum which will be the same anyhow you will find a hit skip pattern .  Mine has a percentile drawn at different skip levels which will show a higher or lower percentile from which I decide to use that digit for that chamber.
I create straight wheels for my plays after choosing my digits.

hypersoniq's avatar - binary

by root, do you mean the last digit(LD) of the sum of the 3 drawn numbers?

Last20Show page 18Page 19 of 24
emilyg's avatar - cat anm.gif

that's what i'm trying on pick 4 - going to keep a list for a month to see which workout hits best. love your idea. thanks.

hypersoniq's avatar - binary

testing and tweaking and observation will take a long looooong time, even with excel...

as soon as I can figure anything out (especially if it dventually has merit) I'll gladly share the info... as for an algorithm...

LastNightsDraw + RundownNumber = 1stRundownResult

1stRundownResult + RundownNumber = 2ndRundownResult


NthRundownResult + RundownNumber = (Nth+1)RundownResult



the RundownResult answers are the numbers that you play for the next draw.

RundownNumber is the 111 (or whatever number you want to use)

ADDING is the system shown, you could easily use Division, Multiplication, Subtraction...

THat's all I know for now, I'm still figuring it out...

hypersoniq's avatar - binary
Quote:Originally posted by emilyg on December 27, 2003

that's what i'm trying on pick 4 - going to keep a list for a month to see which workout hits best. love your idea. thanks.

good luck, hope it brings you some $$$



hypersonig =Too many math computations for me lol......I will stick to Alpha Patterns, Sum Totals with "Elimination for Determination" and let my software do the hard work for me of coming  up with combinations since it has a "controlled number generator" and various filters.

What I've seen in years of tracking is that "most" of the digits that play will replay before they get 5 or more games away from the last time they play.  If I can determine that all digits will come from within the last 4 games I can eliminate all digits 5 games or more out.  If there is a 1/2, 2/1 or 0/3 split, then my software will give me combinations based upon that scenario which will cut down the possible combinations considerably.

I have over 40 charts of record keeping the software even updates for me after 1 entry into the Database of what plays.


 Food for thought:  Most of you might know that there are 120 possible 3 digit combinations that can be written 6 different ways each (6x120=720 exacts).  There are 90 doubles combinations that can be written 3 different ways each....3x90=270.    With 10 Triples+270 Doubles+720 3-digits you get the 1000 EXACTS.

Since there are 3 times more 3-digit combinations then Pures, Doubles & Triples combined most of the time you can eliminate them from playing.  When you do that you will find of the possible 100 combinations left that 68 of the combinations have a Sum Total from 11-20.  There are 26 3-digit combinations that Sum Total 0-10 and 6 3-digit combinations that Sum Total 21+.

In this scenario you can see there are over twice as many 3-digit combinations that total 11-20 then there are in the combined number of Singles & Twenties at 32.   It's on that basis that I say the below group of digit combinations will play most of the time......68 of them.  These are what I concentrate on unless my charts indicate to me that either a Pures, Singles or Twenties is about to play....or whether there will be Doubles instead.  Triples are so remote to me that I rarely even give them consideration although I know they do play.  In one year in Texas they played less then .5% though <G>

 11    12  13   14  15  16  17  18    19  20
056 156 067 167 186 187 386 567 856 857
038 318 085 185 087 367 098 387 487 956
416 057 418 356 096 385 467 918 496 938
407 093 436 095 456 916 485 936 298 497
218 417 409 437 438 097 287 495       
236 435 256 491 267 457 296 297       
209 237 238 257 285 493               
245 291 247 293 249 295 

hypersoniq's avatar - binary

Still trying to figure out how to count skips in excel... Had been reading about using a COUNTIF formula with the INDIRECT formula, but no luck so far in actually implementing it (they didn't provide an example). Anyone have any other ideas?


Barely now how to use Excel lol  Something I've never learned but would like to learn.  I have a program in Excel for Lottery and dont even know how to use the thing :( 

Another reason for developing my own software lol

emilyg's avatar - cat anm.gif

clipper - same here - have it and don't know how to use it. one of my resolutions to learn.


Well....I took a one night Community Education Course on Beginner Excel so I know some basics.  Just cant fully maneuver in the software.  Am thinking about checking the "Professor" CD's <G>

knot.ben.lucky's avatar - eye storm.jpg

......  just a post to try  to keep this thread alive and back on the first pages .... lots of good tips here..... hoping to see more

hypersoniq's avatar - binary

let's make '04 the year...

for all of you wanting to get down and dirty with Microsoft Excel...

the above is a site jam packed with free (FREE) tips and tutorials on using MS Excel...

'04 is the year I combine Excel with MS Access database power... Anyone wishing to follow along...

the above link is free tips and tutorials for using the MS Access Database...

If I come up with anything for excel I'll pst it here as a simple cut and paste formula with instructions, same with access...

let's do this thing already! Time to go back over 6 pages of threads and recap what we have so far...

hypersoniq's avatar - binary

Everyone throw your hats into the ring... It doesn't necessarily have to be something mathematical...

Are you using astrology? number "vibratility"? Dream numbers? Remote Viewing? Travelling Numbers? Just share how and why you choose certain numbrs over others... dventually we should be able to come up with something...

Remember, we're not trapping here, that can be the next phase, right now we're concerned with prediction... Trapping with a wheel will be better if we have a better way to narrow down the field... right now we're narrowing down the field...


Thanks for the links!  I went and bought one of those "Professor" Videos for Excel.  Knew there was online Tutorial stuff....just didnt go there and as much trouble as I've had this past week with online connection I needed something I could work with offline <G>

JAP69's avatar - scene sunovermountains.jpg

I posted back in the thread on the use of box order entry for selecting numbers in that manner. This is some data I put together to help in looking over the columns to select the numbers from the columns.
The first digit in a column will always stay the same. Work the last two digits in each  column  to find the repetiveness of a digit in postion 2 or 3. Look for the digit in position 2 and 3 that is due come back in that position.
Below is the south carolina data count of digits drawn by position.Check your state to see the data count for your state. The data count is the one in blue for the position count.

  south carolina box order dec
089 129 236 334 468 569 679 799 889     
034 159 259 389 457 558               899     
025 124 267 347 469 567               
077 124 245 356 469 578                 
048 125 235 356 459                       
014 118 225 334 446                       
017 179 248                             
018 135                           
047 128                         
035 189                               
013 156                               
057 145                           
                Jan 04
   total draws in systen 824
__drawn by position___box order_
_#_pos 1___pos 2____pos 3______

hypersoniq's avatar - binary
Quote:Originally posted by Clipper on January 02, 2004

Thanks for the links!  I went and bought one of those "Professor" Videos for Excel.  Knew there was online Tutorial stuff....just didnt go there and as much trouble as I've had this past week with online connection I needed something I could work with offline <G>

The nice thing about excel is that you have the ability to list results and backtest various systems. Things can be color coded for ease of readability and with autofill, wntire formulas can be changed and updated instantly.


Welcome Guest

Your last visit: Sun, May 29, 2022, 12:20 am

Log In

Log InCancel

Forgot your username?

Forgot your password?