General solution for special case

General solution for special case

Postby Mirel » Thu Aug 05, 2010 7:59 pm

I noticed that lotto expert team is open to changes.From the first day I saw great strength of the statistical filter.
But this strength program uses only 20%.I have the solution:consists of only a few extra instructions; Is the job for several days.
It is something that you structure the program ,and not the new filters.Exactly what is now follows.
I can use the program at 100% power using tricks. Maybe this I will be disappointed whether the solution is not adopted .
With what my solution would Padawan in "Needle in a hay stack" would be very simple .Is required a new way of working .
I come back with clearer specifications .
Mirel
 
Posts: 40
Joined: Sun Jan 25, 2009 5:58 am

Re: General solution for special case

Postby Mirel » Sun Apr 24, 2011 4:38 pm

What he wants Padawan in "Needle in a hay stack " : viewtopic.php?f=3&t=1427 ( and now again Jerzy in " Synchronicity of settings filter request " )
would have to work very easily : a complex filter,a group of five statistical filters,
each studied each position to a last digit .Why not work ? Because the group does not work synchronous ,each work separately on drawings from winning numbers !
When studying a single draw ,all five result is OK .
It is necessary to introduce a new way of working for statistics filter :step by step , back , draw to draw ,for latest K draws .
So , besides : all,latest,all except,oldest,range be and ONE BY ONE .
Now , if built complex filter which has more components which study 5 position for last digit ,each filter of a position
investigates past draws separately and leave or remove tickets of package according to the result.
In case when number examined of the latest draws is high and setting Ocurrence =1 or Ocurrence >0 ,one statistical filter(by setting remove), part of a complex filter
remove all combinations of package , because in those latest draws (many) statistical filter finds all possible situations .
For example Complex filter (remove ,minimum count of filters to pass 2) formed :
Statistical filter accept First digit -Position 1 Occurence =1 5 latest draws
Statistical filter accept Odd/Eve -Position 1 Occurence >0 5 latest draws
This way of area working ,and it is useful for ocurrence > 1;2;3;...but more statistical filters should focus on each of latest draws .
What he wants Padawan in "Needle in a hay stack " ,remove combination that have the same last digit in position with all past drawings .
Change is easy : Statistical filter set latest/ way ONE BY ONE ,all filters investigates last draw ,leave or remove combinations
which correspond ,and now to give instructions moving to the penultimate draws etc.
To me , this method works perfectly !
I)First I do a complex filter consists of group of six statistical filters (for 6/49):
a)Last digit position 1 Ocurrence =1 latest 1 draws
b)Last digit position 2 Ocurrence =1 latest 1 draws
c)...
d)...
e)
f)Last digit position 6 Ocurrence =1 latest 1 draws
(Each statistical filter Set for 1 draw , last)
Run , remove or leave tickets of package .

Then delete last draws from winning numbers and the first cycle is complete !
And the cycle repeats ,as needed.with other programs everything is automatic and easy to get draws 1,000 back.
Is exactly What he wants Padawan in "Needle in a hay stack " !!

Introduction method ONE BY ONE for Statistical filter (to work synchronously ) has great advantages :

1)Can be introduced quickly ,with little instruction
2)can be used effectively statistical power filter
3)not be built further complicated filters to study the properties ,but will be groups of filters (which is much easier)
Mirel
 
Posts: 40
Joined: Sun Jan 25, 2009 5:58 am

Re: General solution for special case

Postby Mirel » Sun Apr 24, 2011 4:41 pm

A good way for remove package tickets is looking for combinations which have the same properties ,properties that are rarely repeated .
In game 6 / 49 all six last digits ,in the same position rarely repeated ,so can be eliminated all combinations that have the same six last digits
in the same position with all past drawings .to 6 / 49 ro last two drawings are :
3 10 28 31 38 47
1 14 16 24 27 31
want to eliminate in whole package tickets that have the first two positions same last digit.
Make a complex filter with two statistical filters :

Complex filter All must pass accept

- statistical filter reject Last digit position 1 Ocurrence = 1 ( 2 latest draws)
- statistical filter reject Last digit position 2 Ocurrence = 1 ( 2 latest draws)
are properly remove combinations :
3 10 ......
3 20 ......
3 30 ....
3 40 .....
13 10....
....
33 40 .. corresponding to the first draw ,


1 4 ....
1 14 ...
1 24 ...
1 34 ...
1 44 ...
11 14 ...
...
41 44 ... corresponding second draw

but combinations that had not eliminated :
1 10 ...
3 4 ....
1 20 ...
3 14 ...
....
31 40 ..
43 44 ...
why ?
for the two statistical filter does not work synchronously (simultaneously) ,first works the first filter which analyzes last digit of the first position
and only after it ends second start.
which is the solution ? : The two filters should work synchronously the same draw .
where should start ? : as many statistical filters would complex filter if are all set on 1 latest draws ,they will work synchronously !!
example :
Complex filter All must pass accept
- statistical filter reject Last digit position 1 Ocurrence = 1 ( 1 latest draws)
- statistical filter reject Last digit position 2 Ocurrence = 1 ( 1 latest draws)
are properly remove combinations :
1 4 ....
1 14 ...
1 24 ...
1 34 ...
1 44 ...
11 14 ...
...
41 44
Not removed anything from the program ,have added a new way of working .

To :All , Latest ,latest range ,all except ,oldest,date range be added ONE BY ONE Latest draws .

In this way all statistical filters work first the last draw (same way by setting :1 latest draws, so synchronous) ,then remove or leave tickets in the package ,
then an instruction is given move to the penultimate draw (same way by setting :1 latest draws, so synchronous) ,or hide the last draw ,then remove or leave tickets in the package etc.
many cycles each show ONE BY ONE Latest draws .

In this way can be complex filter with as many statistical filters like :
Low/High + Sum
Odd/even + Sum
First digit + Low/High
Number distance + Sum + Last digit etc and All will work well together !(synchronously,simultaneously)

This way of working is possible because I made myself using a external program which controlling EL 5.

all statistical filters are set on 1 latest draws after run-leave or remove tichets -to package-external program remove last draw in winning numbers ,

were here first cycle is complete .then start the second cycle because penultimate draw is now last !etc.


I tested One by one for lastest 15,20,30 draws to 6/49 ro with complex filter: Sum + odd/even AND Sum + Low/High ,results are very good : Jackpot remains and
many millions of combinations eliminated , results were to make .


My opinion is the spot method One By One is more effective than current area method ,which is also useful.
Its efficiency will increase when there will be new statistical position filters and especially if accompanied by Back Test .
Mirel
 
Posts: 40
Joined: Sun Jan 25, 2009 5:58 am

Re: General solution for special case

Postby stan » Tue Apr 26, 2011 12:07 pm

sorry, i'm afraid i don't fully understand what you're asking for.

but there is a new 'first/last digits patterns' filter in the development builds that maybe works the way you request
Expert Lotto Team
User avatar
stan
Site Admin
 
Posts: 6338
Joined: Thu Sep 23, 2004 1:01 pm

Re: General solution for special case

Postby Mirel » Thu Apr 28, 2011 11:28 am

Ok , more specifically when , software may be included in next cycle :
in the complex filter / Statistical Filter / Select draws
cycle 1 now : Select draws / Latest 1 Draws / run /Winning Numbers /Edit /Hidden last draws (one draw ,last , virtual course )
cycle 2 now : Select draws / Latest 1 Draws / run /Winning Numbers /Edit /Hidden last draws (one draw ,last , virtual course )
etc . cycles are identical
User sets how many filters like , accept/reject and number of cycles , must appear in design /winning numbers /select/ nr cycles (So , besides : all ,latest , all except , oldest , range be and nr cycles)
Number of cycles is each extraction after method is applied .
This method is called skipping . I left the observation that multiple filters of Complex filter / Statistical Filter
work synchronously when set on a single draw ( Latest 1 Draws )
This method is simpler implement in software and allows work synchronously no matter how many filters !
Method of achieving filters synchronization proposed by Jerzy , by statistical tables ,is more complete ,
made skipping and something more , but it is more difficult integrated into software by programmers of Expert Lotto !
Mirel
 
Posts: 40
Joined: Sun Jan 25, 2009 5:58 am

Re: General solution for special case

Postby stan » Sat Apr 30, 2011 10:51 pm

ok, i think i know what you mean now. but unfortunately, it isn't easy to implement this type of filtering (even though it might seem simple:). i would have to rewrite a big part of filtering infrastructure in expert lotto 5.
but as i posted before - there are several new 'pattern' filters in the development builds that provide similar results.
Expert Lotto Team
User avatar
stan
Site Admin
 
Posts: 6338
Joined: Thu Sep 23, 2004 1:01 pm

Re: General solution for special case

Postby Mirel » Sun May 01, 2011 2:22 pm

Seem simple to me , a non programmer which give me opinion :) .
I discovered the new filters of development version , number of filtering possibilities , more increased .
An important filter which currently lacks is ticket index digits positions ( all to statistical filter :D ).
In the future when there will be time , can be made and this filter .
I noticed that the EL works in Windows 7 on 4 cores .I checked .Would full work on 6 , 8 or 12 cores ?!!?
Mirel
 
Posts: 40
Joined: Sun Jan 25, 2009 5:58 am

Re: General solution for special case

Postby stan » Sun May 01, 2011 2:41 pm

Mirel wrote:I noticed that the EL works in Windows 7 on 4 cores .I checked .Would full work on 6 , 8 or 12 cores ?!!?

yes
Expert Lotto Team
User avatar
stan
Site Admin
 
Posts: 6338
Joined: Thu Sep 23, 2004 1:01 pm

Re: General solution for special case

Postby Mirel » Sun May 08, 2011 4:51 pm

Good job ! Thanks Stan . All problems solved elegantly and shortly , by statistical tables .
He left a little problem : when using combined statistics , sometimes appears memory error .
System has 8 gb memory / Windows 7 64 bit , intel quad core .
Strangle is from java ? I installed java 64 , same error . How to get the use of larger amounts of memory ?
Number of rows in the table is too large ? passes 1 million lines ? everything is limited to 2 GB ? (especially if you combine the two sums to the analysis )
There are settings in Java / EL , to use more memory ?
Thanks in advance .
Mirel
 
Posts: 40
Joined: Sun Jan 25, 2009 5:58 am

Re: General solution for special case

Postby stan » Sun May 08, 2011 8:31 pm

Mirel wrote:Good job ! Thanks Stan . All problems solved elegantly and shortly , by statistical tables .
He left a little problem : when using combined statistics , sometimes appears memory error .
System has 8 gb memory / Windows 7 64 bit , intel quad core .
Strangle is from java ? I installed java 64 , same error . How to get the use of larger amounts of memory ?
Number of rows in the table is too large ? passes 1 million lines ? everything is limited to 2 GB ? (especially if you combine the two sums to the analysis )
There are settings in Java / EL , to use more memory ?
Thanks in advance .

you need to open c:/program files/expert lotto 5/etc/expertlotto.conf file in notepad and add "-J-Xmx=2048m" to the line starting with 'default_options'. this way expert lotto will use up to 2gb of ram. you can naturally specify even more available memory.
Expert Lotto Team
User avatar
stan
Site Admin
 
Posts: 6338
Joined: Thu Sep 23, 2004 1:01 pm


Return to Comments, suggestions, feature requests

Who is online

Users browsing this forum: No registered users and 5 guests

cron