[quote=Jerzy]
Lexus,
You give us examples to show how it could work. If we use 8 variants in 4 columns, there will be 4096 combinations (8*8*8*8 = 4096). Probably too many combinations. As a way around the problem you suggest adding the variants 8+8+8+8 = 32
I do not know what you mean by uniting the results. What's that? Why do you do it? How do you do it? Did you try to use the Joe's filter for testing your ideas or a different piece of software?
[/quote]
In previous Post, there is an answer to your remark
[quote=Jerzy]
Lexus,
Prefiltering the package with different filters to get approximately 1,000,000 tickets before using the new filter, as you propose, would decrease the time of filtering.
[/quote]
Certainly.
And as for me having improved Match Number filtre it is possible to reduce the size of a package lost-free. It is possible to read under the reference in more details:
http://www.expertlotto.com/en/forum/forum.php? req=thread&id=1059&page=9&nocount=1
To reflexion.
All filters which are in EL can be used in two ways which are reduced to: Remove or Leave.
Starting with statistics to guess on 100 % a probable outcome of event it is impossible! Then it is better to exclude the least probable event.
As the result a package decreases on 1000-10000 etc. combinations.
As in EL it is a little filters if to compare to our requirements and desires enough big package of combinations on an exit is received. Further we already risk to be mistaken in our choice, what further to filter.
Reading a forum, I see that many use Excell for conducting statistics and many have those or other ideas. But not all of us programmers and create the filter we CAN not!
But easily we embody the ideas in a reality by means of formulas Excell!
Here what for batch operation in Match Number filtre is necessary
[quote=Jerzy]
Lexus,
In point 3 of your remarks you suggest that it is possible to remove the majority of tickets from the package by using 650 - 700 number order variants. But it is important to know how many columns can be used or should be used with so many number order variants to achieve acceptable results. Probably we do not know that yet.
You mentioned quite rightly that there should be no correct number order variant at least in one of the columns to keep the jackpot ticket in the package. But you did not mentioned clearly that the filter should be set for Remove option.
[/quote]
In this case too it is necessary to use option Leave because we filter at first on a column 0, we unite, then on a column-1, we unite, etc.
If in a column-5 we use a set in which there is no correct value on an exit in an incorporated file, WIN combinations any more will not be. Other columns (-6,-7...) values will already not have
[quote=Jerzy]
Lexus,
The same result can be achieved if we use 20 - 70 number order variants, with the correct variant present, for the same columns, and with the filter setting Leave. Probably the processing time would be the same as for Remove option, unless different algorithms are used, but finding 20 - 70 number order patterns to leave seems to be easier than finding 650 - 700 to remove.
[/quote]
Certainly it is possible and so. Only we use option Remove
In the attached file I have drawn the scheme (saw already such at a forum) - a principle of filtering which supplements the document with my theoretical calculations
If to think, the scheme also shows a principle of work Winning History Number Filtre, WN History Differences Segments and to it similar
You do not have the required permissions to view the files attached to this post.