I AM BRAZIL, HELP,HELP , HELP

[lista0800]

[pre]I need a lot of help from you, to create a program in visual foxpro, I am from the time of the clipper and do not program for a long time so I come here humbly to ask for help.

I need a visual foxpro program, which takes 60 tens and combines lines in 50 tens, will give 75,394,027,566 records, so I need a program that runs as fast as possible because I know it will take time ... preferably be in dbf, but as it has the size limitation, can be txt ... I am willing to pay for this service, I need it to fix a bug in an old program of the company I work ... here in Brazil, we do not have a lot of visual foxpro reference[/b][/pre]

 

[lista0800]

this program i am running is an old program of mega sena (the brazilian lottery), which is made up of 60 dozens ..
combining the 60 tens into 6 tens lines, we have just over 50 million lines ...
but I want to combine the 60 tens into 50 tens lines, which will generate over 75 billion 50 tens lines
What is the purpose: (this information can be useful to make the program faster)
I generated about 1666 lines from 24 dozen, covering all of the more than 50 million mega sena combinations and noticed that of these 1666 lines, 6 of these lines never scored the 6 dozen together, so I used these 6 lines as a filter ....
How does this filter work? my program that I made to generate lottery predictions, generates a lot of combinations of 6 tens (about 2 million), but I can find the 6 balls drawn in about 2 million combinations (much better than playing with 1 change in 50 million), By using these 6 lines of 24 dozen as a filter, I was able to reduce about 10% of these combinations.
How do I use the filter? of the about 2 million lines that I generate, I check if the dozens are in these 6 lines of 24, if I delete.If I delete this line, since in 20 years there is the lottery in Brazil, these 6 lines never scored the six dozen, their chances of scoring is very weak ...
Now I want to do the same with lines of 50 dozen, an idea for the program to get faster, would be, before writing the information, take the database of past games already drawn, and check if the line of 50 generated, contains the 6 dozen, from any of the 2191 games already drawn so far ... if you have eliminates this line before recording (it can consist of 1 to 5 dozen but never 6 dozen).

arquivos dos resultados da loteria do brazil

https://files.engineering.com/getfi...66e-ac4b-0674ffcac70f&file=megasenabrazil.DBF

 

[tomk3]

Hello,

1) If everything is fair each number has exactly the same chance to appear.
So what should the program be good for ?

2) VFP has a limit of 2GB for a file.
Not sure on how you want to store how many data, but you should check recordsize * number of records before.

For fast search you may need to index each column to speed up, see help for rushmore.

And/or add a table "invalid numbers" (just one columns or maybe an addition info why) and search this when creating.
Then when creating new numbers check each one directly after creating, if in invalid table, take a new one and do not forget to check for doubles.
Ehen you have 50 or 60. stop

Best regards
tom

 

[Koen Piller]

Hi,
Your assumption on their chances of scoring is very weak ... because they have never been scored is to my humble opinion not correct. If you follow 'the law of the big numbers' it means that each and every number has the same chance to be scored again. Meaning that after a while you will notice that most numbers have equally scored, if now by any change a number is significant lower in score the change that it will score on the next ballot is higher, the numbers with the lowest score have the biggest change. But remember change, it is not 100% they will score there is always the factor of randomness.
Your dbf with 2193 records, are these the numbers which have scored over the last 2193 ballots?
Interesting.
If that is so, I have a routine which will tell you how many times each number was drawn, i.o.w. what is more interesting, what are the numbers with the least drawns consequently which the highest change to be drawn on next ballot.
Regards,
Koen

 

[lista0800]

Actually this idea that numbers have the same chance ... or statistics that more numbers come out, columns etc .. I do not need ... I apologize ... Study the Brazilian lottery to 14 years ... and already tested and used several techniques, I have a good idea of ​​what works and what does not work ... I was able to reduce from 50 million to 2 million be a great victory ... more unfortunately in Brazil the awards start at $ 750,000 which compensate people to join play (we are united the initial games start with 10,000,000 dollars, yes it would compensate) ... if alche alci can help me, I just need a routine that takes 60 tens and generates combinations of 50 tens ...