History Differences filter

[Jerzy]

Stan,

I believe that the History Differences filter (In Filters --> Winning Numbers History) can perform a useful function but I do not know anything about it. I might be wrong if I try to find the answer by guessing when I do not have enough information. So could you explain it to me?

What is the Number count? What are the History differences? What are the Numbers in a single column? And how can I find the most appropriate or the most useful settings for this filter?

I would like to see this filter working.

I am also interested in opinions about this filter and practical experience with it that other EL users may have.

Thank you.

Jerzy

[Joe]

Hi Jerzy,

History difference filter is for Segment columns, This deals with how many Winning Numbers occurred in Rows from Column 0 to -10.

You can know about this filter if you see my Excel sheet for this with my results that I have posted.

Take the First Row in each column (column 0 to -10) and count the Winning Numbers.

History differences is nothing but the Rows.

Number count is nothing but the winning Numbers occurred in selected History differences i.e in selected Rows.

Single Column is nothing but the Winning Numbers occurred in Single column.

Hope this makes little clear, if not I am glad to explain it,

Thanks,

[Joe]

Jerzy,

You can see the Excel sheet here

http://www.expertlotto.com/en/forum/for ... age=4#5581

thanks,

[PadawanLotto]

[quote=Joe]
Hi Jerzy,

History difference filter is for Segment columns, This deals with how many Winning Numbers occurred in Rows from Column 0 to -10.
[/quote]

Joe, I'm sorry but I have to disagree with you on this one as the Winning Number History Differences have nothing to do with Segments. Winning Number History differences are the column value of the last drawn and the next drawn numbers based on where the numbers in the next draw came from in the past drawings. History differences have to do with where the numbers in the past drawings come from in the next drawing.

Segment values are determined by your default settings for segment 1 and segment 2 and segment 3 (whats left over). You can have the history differences set tight as -10/+10 and still cover all segment posibilities 5-0-0/0-0-5.

[Jerzy]

Hi Joe,

Thank you for your information. The Excel sheet shows your idea very well.

But this History Differences filter removes the jackpot ticket from the package, when I set Number count very strictly. So I am not sure whether I am using the filter incorrectly or whether there is something wrong with the filter.

"History differences" idea is used in many contexts which makes it sometimes difficult to follow and may cause confusion.

We have to remember what is stated in the Help file for History Differences data of the WN History section:

"Each table row represents one possible difference value".

And:

"Row headers show the difference value and the count of numbers in the row (in parenthesis), row cells contain all drawn numbers which have the same difference in the given column in the Number History table".

The rows which represent the history differences are grouped together according to the length of the history to form segments.

The question is how many rows we should use when working with this filter. One row only, or several rows? Which ones? All rows of a segment? Which segment? Maybe all rows of 3 segments?

I experimented with 6 rows of the first segment.

I counted the occurence of winning numbers in the first segment in all columns using the data from the previous draw.

The count of numbers in 11 columns (0 to -10) was 40 and, consequently, I set the filter's Number count as 40 to 40.

I set History difference as 1 to 6 to include 6 rows of the 1st segment. (Maybe 6 to 6 would be appropriate?).

I set Numbers in a single column as 0 to 6. In my lottery from 0 to 6 winning numbers can be drawn in the 1st segment of any column.

7,294,492 tickets were filtered out of the package (8,145,060 tickets in the package for 6/45 lotto). Unfortunately, the jackpot ticket was also filtered out. However, a lot of 2nd, 3rd, 4th, and 5th prize tickets still remained in the package.

Using Number count 39 to 41 also filtered the jackpot ticket out.

How to use this filter effectively and preserve the jackpot ticket?

Regards

Jerzy

[Joe]

Jerzy,

Its working fine for me.

Let me check this again and come back to you.

Are you sure about skipping the latest one from the WN Dabase ?.

[Joe]

[quote=PadawanLotto:1181512470]
[quote=Joe]
Hi Jerzy,

History difference filter is for Segment columns, This deals with how many Winning Numbers occurred in Rows from Column 0 to -10.
[/quote]

Joe, I'm sorry but I have to disagree with you on this one as the Winning Number History Differences have nothing to do with Segments. Winning Number History differences are the column value of the last drawn and the next drawn numbers based on where the numbers in the next draw came from in the past drawings. History differences have to do with where the numbers in the past drawings come from in the next drawing.

Segment values are determined by your default settings for segment 1 and segment 2 and segment 3 (whats left over). You can have the history differences set tight as -10/+10 and still cover all segment posibilities 5-0-0/0-0-5.
[/quote]

Padawan,

I am sure about my explanation. because stan has done this as I explained my idea on Hd filter. I am sorry if I mis-understood your views on this filter.

[Jerzy]

Joe,

Today I repeated my experiment with the same settings as yesterday and the History Differences filter filtered out a different number of tickets, leaving the jackpot ticket in the package this time.

Yesterday I did my experiment with the same setting several times and each time the result were incorrect even though the Number count that I used in the filter was correct. I do not know what went wrong.

Yesterday and today I skipped the last drawing in Edit Winning Numbers and counted the last draw winning numbers in the first segments of all columns in the History Differences table. Also I used the data from the Segments table which gave me the same number count.

Today I repeated my experiment with two different draws and in both cases the result was OK.

The filter probably can remove a large number of tickets while leaving the jackpot ticket in the package if you use the most common range of number count. I need to see how it works with other filters, particularly with segment filters. Sometimes the different filters overlap with their action and together they remove a smaller than expected number of tickets.

I am sorry for reporting a problem with this filter which seems to work correctly.

And thank you Joe for your help.

Regards

Jerzy

[CARBOB]

This filter is confusing for me. Would someone please post an example, thanks.

[Jerzy]

Carbob,

Using History Differences filter we are trying to find the winning tickets for the next draw using the data from the previous draws. This is initial approach which is not different from working with other filters. But unlike with other filters, currently there is no statistical information in EL to support this filter (no number count statistics). I hope this will be provided in the future. Meanwile we can create some simple statistics for ourselves.

As I understand it, the right answer for the next draw could be located at the cross section of rows and columns, and the History Differences filter probably makes it easier to find it.

The filter uses 3 parameters.

History differences parameter should be considered first, not second, and its name should be changed to Range of rows to decrease the confusion it may cause.
When we are using this filter our first consideration is which range or rows we want to use. We can use a single row or groups of rows, for example the rows of the specific segment. I do not know yet which approach would bring about the best outcome.

The number count parameter- means the count of winning numbers in all 11 columns within the selected row or range of rows. Numbers are counted separately for each columns and than added to obtain the count for the filter. This way the same numbers are counted several times, as they appear in columns, but that is how it should be.

No statistics for number count are available in EL yet, but you can prepare it for yourself.

If you want to work with the same range of rows as those which are by default set in segments, preparation of statistics could be easy.

In the Summary table → Segments you can find the characteristics of all default segments. And you can translate segments statistics into number count statistics for a given segment. After doing it for a number of drawings you would find the most commonly occuring range of number count for a given segment. In my lotto (a 6/45 number game) the most commonly occuring number count range seems to be from 30 an something to 40 and something (based on several examples only, more research is needed).

Example:

If there are the following segments:

2-2-2, 5-0-1, 4-1-1, 3-2-1, 3-2-1, 2-3-1, 6-0-0, 5-1-0, 4-0-2, 5-1-0, 2-3-1
the number count for the first segment is 2 + 5 + 4 + 3 +3 +2 + 6 +5 + 4 + 5 + 2 = 41

The third parameter is Numbers in a single column.

Normally all winning numbers occur in a column, but using this filter we are working with parts of columns, maybe segments. No winning numbers, or some of those numbers or maybe all winnig number can occur in a given section of any column, and the filter allows for a wide or narrow setting.

Regards and good luck.

Jerzy

[CARBOB]

Jerzy,

Thank you for a very good explanation on how this filter works.

[Joe]

Excellent Jerzy,

please note, we have not yet finished the HD filter effectively, there are some features have to be implemented.
Still I am waiting for stan to implement, you can feel worth of this filter once we get all the features in this.

thanx,