Using filters a different way

[Nick Koutras]

to Dennis...

Hello,

Quote:
How do you applied the resampling on your population ?...There are many ways of doing this resampling....What criterias did you use ?
Your program is very nice! I had a look at it this morning!

--------------------------

The population as it stands in this little program is random.

No criteria used.

I hope you can find it useful.
Thanks for the good words.


Nick

 

[Dennis Bassboss]

Thank you Nick! It is going to be usefull!

 

[Nick Koutras]

to Goswinus

Thanks for the offer.

Below is a LD(13,6,4,6,L=11)=65
with 3 missing

1 2 3 4 5 6
1 2 3 7 8 9
1 2 3 4 7 10
1 2 3 7 11 12
1 2 3 8 10 11
1 2 3 4 10 13
1 2 3 5 6 8
1 2 4 8 10 12
1 2 4 5 6 7
1 2 4 8 11 12
1 2 5 10 11 13
1 2 5 7 9 13
1 2 6 9 10 11
1 3 4 8 12 13
1 3 4 7 8 11
1 3 5 7 11 12
1 3 5 8 10 12
1 3 6 9 12 13
1 3 6 7 10 11
1 3 8 9 10 13
1 4 5 7 9 11
1 4 5 6 7 9
1 4 6 9 12 13
1 4 6 8 11 12
1 4 7 10 12 13
1 5 6 9 10 13
1 5 7 9 10 12
1 6 7 8 9 11
1 6 7 9 10 12
1 8 9 11 12 13
2 3 4 5 9 12
2 3 5 6 9 10
2 3 5 7 9 11
2 3 5 6 11 12
2 3 6 8 9 13
2 3 6 7 8 13
2 3 7 8 11 13
2 3 7 9 12 13
2 4 5 7 8 11
2 4 5 7 10 13
2 4 9 11 12 13
2 4 9 10 11 13
2 5 6 8 10 13
2 5 6 7 8 12
2 5 9 10 11 12
2 6 7 11 12 13
2 6 8 9 12 13
3 4 5 6 10 11
3 4 5 6 11 13
3 4 6 9 11 12
3 4 7 8 10 12
3 4 8 9 10 12
3 4 8 9 10 11
3 5 6 7 12 13
3 5 10 11 12 13
4 5 6 8 12 13
4 5 8 11 12 13
4 5 8 9 11 13
4 6 7 8 10 13
4 6 7 9 10 12
4 6 7 10 11 13
5 6 8 10 11 12
5 7 8 9 10 13
5 7 8 9 10 12
6 8 9 10 11 13

If you can improve it that will be fine!

Nick

 

[thornc]

This is starting to resemble genetic algorithms!

A Portuguese guy has a paper about using GA in a lotto program that it's very interesting... to bad the paper is written in our language! He has a different approach, but your's is getting close
to a GA as it comes!

Start with iniatial Population,
Evaluate Population
Crossover, Mutate
New population
New evaluation...and so on!

 

[dwoods99]

This is a very interesting thread... I sure hope LT has been doing his backups lately !

 

[gsobier]

To answer you and Beaker... ...I wrote my own software to do everything including the selection of sets via filtering. My feeling is, there was nothing I've seen on the internet which I would trust at that time months ago. Recently, I've found this forum to see what others are doing and they have been interesting.

My 30 number selection is TOTALLY based on the history of the previous draws. I rank each number and also wrote a routine to tie-break the 6 which did get drawn and the frequency of the number being drawn (a hot number) makes it into my 30 eventually. The thing I've realized the most since being involved in this forum is the value and more significant importance of repeating numbers which I will be introducing in to my process more effectively soon. I'm the author and user of my software I designed and wrote all of it on my own. There is nothing like it anywhere else. I don't use spreadsheets loaded with macros, everything is done with the programs I wrote.

Beaker is fearing I'm picking sets by random... ...no, no, no... ...by PATTERN in the 30 numbers drives my selection process which filters absurd combinations.

What you are asking me is the document ny work and me being a typical programmer, I'd rather write code instead of writing documentation.

Here is a sample of the pattern I'm refering to and the numbers belonging to them...



.O..........O....,,,,,,,,E,,Z; ** (4) .......,,,,,,E,E;; Apr 19, 2003
45 49* 3 17 19 25 27 37 15 5 29 7 43* 1 33 21 23 32 16 44 46 36 38 8 6 12* 22 28 40* 20 | 35 31 41 9 39 11 13 47 18 2 34 42 24 26 48* 14 4* 30 10

........O....,,,,,,,,,E,,,,;;; - (2) ........OO..,,E,E,; Apr 23, 2003
35 3 17 43 25 37 11 5 29* 27 1 15 23 26 32 16 2 24 46 36 48 8 6* 12 22 38 28 40 10 30 | 49 31 7 41 13 19 39 21 45* 9* 33 47 18 44 34* 42 14* 4 20

..OO..O........,E,,,,,,,,,,;;Z *** (5) ..........,,,,,,,,Z Apr 26, 2003
45 25 3* 17* 49 21 35* 5 29 43 11 19 1 15 23 28 32* 22 26 48 36 8 2 46 12 38 4 40 10 30* | 47 31 41 7 13 27 39 9 33 37 24 44 34 42 6 16 14 18 20*

.O....O....O...,,,,,,,,,,,E;;; ** (4) ..O...O...,,,,,,,,; Apr 30, 2003
45 29* 25 3 17 35 19* 15 37 1 5 43* 11 31 27 4 26 48 36 46 8 32 22 12 38 2 6* 40 20 30 | 23 41 7* 49 39 21 47* 13 33 9 18 28 44 34 42 16 14 24 10

.............,,,,,,,,,,,,,,;;; (?) ............,,,,,,; May 3, 2003
29 11 3 49 33 27 1 5 35 19 15 21 31 24 4 14 48 26 22 36 46 12 44 6 8 38 28 40 30 20 | 37 23 9 7 25 39 43 47 13 17 45 41 18 32 34 42 16 2 10

Regards,
George

This is a very interesting thread... I sure hope LT has been doing his backups lately !

 

[peter]

So now George I would have to ask, how do you determine the ranking, and what is the pattern?

 

[Dennis Bassboss]

So now George I would have to ask, how do you determine the ranking, and what is the pattern?

My question would be why 30 numbers instead of 31 or 32 or 29 or 40???

 

[Beaker]

My question is why not start a new thread

Thanks for that George - if you go through the posts at this site you're going to find alot of gems that will help you.



30 numbers is alot Fortunately, I'm a set 'n wheel person rather that set-filter 'n play like yourself.

Good luck with that

George, you're old enough to code in Fortran - what language do you speak?

 

[gsobier]

Beaker:

I've mostly written code with FORTRAN, COBOL, PL/I, ASSEMBLER, (370 and 390), and Rexx (these are all for IBM mainframes).

Yes, you are right, 30 numbers is too much and scares the heck out of me. My number selection on the other hand is very intimite with the patterns of my reports so I'm wondering if that is getting me results. I honestly have no idea how valid my process is. I'm using logic and common sense in my approach. All the things I write, for this or for business are reliable and robust routines which strive to be as generic as possible.

Regards,
George

My question is why not start a new thread

Thanks for that George - if you go through the posts at this site you're going to find alot of gems that will help you.



30 numbers is alot Fortunately, I'm a set 'n wheel person rather that set-filter 'n play like yourself.

Good luck with that

George, you're old enough to code in Fortran - what language do you speak?

 

[Beaker]

The reason I ask is I'm not as 'seasoned' as you , but I started as a FORTRAN/ PL/1 coder way way back, SPF and CLISTS also with many JCL Procs for the 390. I never did learn ASSEMBLER - my coding career hit a brick wall - too much for me.

What do you code on the PC? I have an old FORTRAN compiler here and all my Q&D's for the lotto are done with that

 

[gsobier]

Dennis:

I picked 30 because the frequency of 6 numbers all being included was fairly reasonable. The 5 and 4 number population per draw was pretty good too. There were 3 occurances of 0 which means all 6 numbers were in the 19 numbers I tossed out. Playing those 19 numbers could pay off. Look at what Ben pulled off a few weeks ago.

If I could get this down to 24 or less than 20 then we really got something. I know the 30 I'm using now is on the verge of rediculous and I'm all ears for methods/techniques to improve. I'm constantly looking for improvement.

Regards,
George

My question would be why 30 numbers instead of 31 or 32 or 29 or 40???

 

[gsobier]

Beaker:

You just "dated" yourself my friend (no offense), SPF has been ISPF for more than a decade. Noone in thier right mind would code CLIST anymore (just to maintain crummy old code), its mostly in Rexx now. There are also PIPELINEs which blows away other processing at times, I use it constantly.

FYI, IBM now calls MVS z/OS and VM is z/VM. There still is CICS, that has not changed. Then there is dB2 for SQL and HSM, etc.

I've written C++ (I call it C--) and Pascal on PCs.

Regards,
George

The reason I ask is I'm not as 'seasoned' as you , but I started as a FORTRAN/ PL/1 coder way way back, SPF and CLISTS also with many JCL Procs for the 390. I never did learn ASSEMBLER - my coding career hit a brick wall - too much for me.

What do you code on the PC? I have an old FORTRAN compiler here and all my Q&D's for the lotto are done with that

 

[Beaker]

Beaker:

You just "dated" yourself my friend (no offense), SPF has been ISPF for more than a decade. Noone in thier right mind would code CLIST anymore (just to maintain crummy old code), its mostly in Rexx now. There are also PIPELINEs which blows away other processing at times, I use it constantly.

FYI, IBM now calls MVS z/OS and VM is z/VM. There still is CICS, that has not changed. Then there is dB2 for SQL and HSM, etc.

I've written C++ (I call it C--) and Pascal on PCs.

Regards,
George

I'm not that old George been out of that business for a long time as you can tell. Actually, what I was doing I never really was in the business - it was a very old shop.

I'm walking upright now

 

[peter]

All I've got to say is with all this heady expertise here, we should all be cashing in very soon.
All kidding aside George, if you looked at some of the things I just posted in the discussions forum, ie: LD's, consecutives, odd/evens, etc, whittled the selections down, wrote some outstanding code, let those of us who wheel , continue to wheel, and you could line filter, I see $$$$$$$$$$$$$$$$$$$$$$$$$

 

[gsobier]

Peter:

There is lots to read and think about here lately... ...I hope we make some progress.

Regards,
George

All I've got to say is with all this heady expertise here, we should all be cashing in very soon.
All kidding aside George, if you looked at some of the things I just posted in the discussions forum, ie: LD's, consecutives, odd/evens, etc, whittled the selections down, wrote some outstanding code, let those of us who wheel , continue to wheel, and you could line filter, I see $$$$$$$$$$$$$$$$$$$$$$$$$

 

[winhunter]

Thread died?

Where did everyone go? I was reading this thread and it seems to have died.


George-

Im not sure if your aware, but WINHunter does number elimination, not line elimination. With some very good optimizations and a good Filter Stack design, you can eliminate down to your 30 numbers and still maintain the jackpot a good percentage of the time.

But ill go one step further. WINHunter can analyze the success rates of its stacks, and show you how well or not so well it performed. All guesswork aside, you get to see how it did in the past based on your configuration parameters.


Andrew

 

[gsobier]

Re: Thread died?

Andrew:

I'm still here... ...I have never seen WINHunter run anywhere. I've got Windows XP and nothing else like VisualBasic, C++, etc. right now.

With WINHuner, do you need to decide which numbers to use? In other words, would I need to toss out lets say, 29 numbers?

Regards,
George

Where did everyone go? I was reading this thread and it seems to have died.


George-

Im not sure if your aware, but WINHunter does number elimination, not line elimination. With some very good optimizations and a good Filter Stack design, you can eliminate down to your 30 numbers and still maintain the jackpot a good percentage of the time.

But ill go one step further. WINHunter can analyze the success rates of its stacks, and show you how well or not so well it performed. All guesswork aside, you get to see how it did in the past based on your configuration parameters.


Andrew

 

[winhunter]

Decides for you

WINHunter allows you to stack/group (for series/parallel functioning) filters as you see fit. These Groups of Filters have outputs that are passed into a selection method that makes the final decision of what numbers to pass into the next group of filters for processing. WINHunter so far has had mixed successes in actual predicting. But, WINHunter gives the user the ability to test ANY design against the entire past history and display the results on screen. You instantly (maybe with some delays on older PC's, or more delays with a huge stack design) can see the amount of balls predicted and how well those predictions performed against the drawing history.

WINHunter comes loaded with user configurable options, and the greatest thing of all is that it is Opensource Freeware. I have made every attempt to make the sourcecode as simple as possible to allow for other users to jump into the project, but so far that has yielded 0 contributors, other than a few users who submitted processor ideas.

WINHunter is a number picker, not a line picker. The best performance I have seen (granted, the prediction was after the fact) was a prediction of 11 numbers which yielded a 6 0f 6 match out of 53 balls.

I wont go into to much more detail about WINHunter, because there is a whole forum set aside here just for it. Perhaps you should download it and poke around a bit.

On a more personal note, I have come to understand the power of wheeling, but wheels are useless unless you pick the RIGHT set of numbers. Thus, WINHunter is born.


Oh, one more thing. A recent month long test of WINHunter revealed that so far it performs at least as good as random expectation. Thanks to Brad, Beaker and Combo who submitted daily and weekly predictions and prediction results from their own stack designs. You really should go check out the WINHunter section....


Andrew