Combination Groups Linked to WNH Differences - Statistics an

[Jerzy]

Stan,

Would you mind building a statistics with filtering possibility for groups of number combinations in the package linked to, or associated with, specific Winning Number History Differences.

The statistics would identify all the WNH Differences, based on the current data and the content of the package, and it would count the combinations in the package which are linked to each of the WNH Difference. The filter button would allow the removal of unwanted combination groups.

The statistics and filter would operate from the Statistics section of EL, and would work for each column separately (a spinbutton for the column selection would be better than checkboxes). This statistics would work only for the combinations in the package (not needed for Winning Numbers History).

The statistics would contain the list of WNH Differences present (for the column) with information on the number of combinations in the package which are linked to each of the Differences.

Selecting a line with a given Difference (for example the difference associated with the largest group of combinations in the package) and clicking on Filter, would initiate the filtering action.

Why I'am asking for this statistics and filter?

My experiments in this area show me that the WNH Differences which are linked to the greatest number of tickets in the package are not necessarily the differences for the next draw and can be removed.

For 6/45 lotto which has 8,145,060 combinations, removing just one WNH Difference in each of the columns would result in the removal of approximately 1,900,000 tickets. For example, in my 6/45 game the difference of 9 in the column 0, could be linked to 228,822 tickets.

And such statistics and filter would allow further research of the WNH Sums/Differences.

Best Regards

Jerzy

[stan]

there's 'winsum' plugin that shows stats for a group of wnh differences for winning numbers tickets which is similar to what you're asking for, i think.
would it help if i made it a regular stats table that would show values for pakcage tickets and enable filtering on it?

[CARBOB]

[quote=stan]
there's 'winsum' plugin that shows stats for a group of wnh differences for winning numbers tickets which is similar to what you're asking for, i think.
would it help if i made it a regular stats table that would show values for pakcage tickets and enable filtering on it?
[/quote]

Stan, I can't find Winsum plug-in, where is it located?

[stan]

[quote=CARBOB]
[quote=stan]
there's 'winsum' plugin that shows stats for a group of wnh differences for winning numbers tickets which is similar to what you're asking for, i think.
would it help if i made it a regular stats table that would show values for pakcage tickets and enable filtering on it?
[/quote]

Stan, I can't find Winsum plug-in, where is it located?
[/quote]

http://www.expertlotto.com/en/download/beta.htm

(scroll down that page)

[Jerzy]

Stan,

I have installed WinSum long ago (I followed manual installation procedure which worked well for other plugins).

Unfortunately, Winsum does not show at any page of EL, even though it is present in the Plugins folder and listed in Lottery --> Preferences --> Plugins.

So I do not know what it is.

Jerzy

[Maryland]

Same here Jersy, it shows in plugins under preferences but that's about it.

[Da80th]

hi Jerzy and Maryland,

The stats are locate in WH History. See if you have:

Neg/Pos Differences
Hit & Skip Differences


Da80th

[CARBOB]

[quote=Da80th]
hi Jerzy and Maryland,

The stats are locate in WH History. See if you have:

Neg/Pos Differences
Hit & Skip Differences


Da80th
[/quote]


Thank you for pointing that out.

[Jerzy]

Da80th and Stan,

I know Neg/Pos Differences and Hit & Skip Differences, but they do not seem to be WinSum. There is nothing about sums in them.

What I am asking Stan would look approximately like this (example only):

Col. -1

WNH Difference Number of tickets in the package linked to this difference
-72 1
-71 8
-70 15
[...]
-1 157,342
0 180,234
1 201,011
2 211,576
3 222,891
4 234,090
5 220,191
6 180,473
5 179,200
[...]
38 27
39 3

The range of WNH Differences is from -72 to 39. The Difference of 4 has the greatest representation in the package (234,090 tickets). This difference might be the next draw difference, but chances for that are rather low. There are many other differences present which have eaqual chance of being the next draw difference.

If we remove the difference of 4 we will achieve the greater package reduction compared with removing other single differences.

Selecting the line with the difference of 4 and clicking on the Filter button would start filtration.

Best Regards

Jerzy

[Da80th]

hi Jerzy,

OK, I understand now. I found a long way to do this and you need Excel to do the calculations, some day Stan can create this plugin and give you the summary automatically.

Let try something small so your pc won't freeze :-p

Create a 20 numbers (1 to 20) full wheel and go to Package Statistics, then click the "Click Here For Options", uncheck everything except the Diff 0, Diff -1,...Diff -10. Export the file and save it. When you open with Excel you can sort the Col 0 (one at a time) and do the quick count. Write down the total for each Differences and it will give you the summary of what you looking for.

If you want to do a full package counts then you will need to create a second full wheel of 19 numbers (21 to 39), that will give you 39 numbers for 5/39 lotto game. I think you get the idea.

Best bet is to have Stan to create the plugin to do this calculations for us.


Da80th

[Jerzy]

Da80th,

Thank you for thinking about it and presenting this idea. It renewed my interest in the Package Statistics.

But a wheel based on only 20 numbers, which contains only 38,760 combinations, would be unsatisfactory representation of the full package, which has all 45 numbers, each number appearing equal number of times, and which contains 8,145,060 combinations. I'm afraid that the results obtained this way would be incorrect.

When probing the groups of numbers in the package which are linked to distinct WNH Differences, I use a file of 100,000 randomly generated combinations (the file was created according to the instructions for Strategy 1 & 2). In these 100,000 combinations there are all 45 numbers and the occurrences of numbers are approximately equal).

Working with the package of 100,000 combinations, using the WNH Sums/Differences filter is sufficiently quick, and the results are accurate or almost accurate. To make sure that the result is correct, it needs to be confirmed by filtering the full package for 1 -2 or 3 differences in question.

The biggest groups of tickets are usually those that are linked to positive differences. In certain columns there could be only one biggest group of tickets, while in other columns we can find several large groups of similar size.

I'm asking for a tool capable of calculating the statistics for all the WNH Differences in one go. It would still take some time, but much less time than doing calculations separately. And the filter button present on the statistic page would make it even more functional.

Best Regards

Jerzy

[Da80th]

hi Jerzy,

I used 20 numbers full wheel as an example, Excel cannot handle any such large file like the 45 number full wheel.

One idea is to wheel all 45 numbers in a 5if5 or 5if6 wheel, do some backtesting and see if any trend in the WN Differences that we can use for future prediction.

You can find the 5if5 wheel for pick 6 lotto here: http://www.weefs-lottosysteme.de/systeme,6,5506,en.htm

The 45 numbers wheel with 5if5 condition = 226,184 tickets

Good lucks, and please let us know how you do.


Da80th

[stan]

[quote=Jerzy:1199508090]
I know Neg/Pos Differences and Hit & Skip Differences, but they do not seem to be WinSum. There is nothing about sums in them.
[/quote]
it's called 'winsum' because the original author of the plugin chose that name. i just re-implemented his idea in java

What I am asking Stan would look approximately like this (example only):

Col. -1

WNH Difference Number of tickets in the package linked to this difference
-72 1
-71 8
-70 15
[...]
-1 157,342
0 180,234
1 201,011
2 211,576
3 222,891
4 234,090
5 220,191
6 180,473
5 179,200
[...]
38 27
39 3

that's what hit & skip differences table shows (i think). it just doesn't include data for package tickets

[Jerzy]

Stan,

Hit & skip differences shows the occurrences of differences in the game history.

Sometimes I use this filter to remove the differences, which despite long, (more than 20 years) game history have the occurrence of 0. There are many such differences, but they are located on the periphery of the differences spectrum and are linked to a small number of tickets. The number of the tickets removed this way is relatively small.

I am looking for a possibility of finding:

1. the spectrum of differences from the most positive to the most negative in each of the columns
2. how many tickets in the package are are linked to each of the differences
3. which difference(s) is or are linked to the largest group(s) of tickets in the package

Having this information I would like to remove the differences which are linked to the largest groups of tickets (and consequently, to remove the largest possible number of tickets).

I would like to be able to do it in one go and within approximately 1- 2 hours time.

It can be done using the existing functions of the current EL build but it is very tedious and takes approximaly 10 - 12 hours of continuous work.

Jerzy

[stan]

i think we're on the same wavelength:)
it should be fairly easy for me to provide such stats, so stay tuned pls:)