by **Gatsby** » Wed Jan 04, 2017 10:16 pm

Hi Stan,

Thanks for your ready reply.

I Wish i could better express what i´m trying to expose.

"Why 50 numbers? Why not just those 5 latest numbers?"

Well, this is the exact point, because you just look at the last five numbers from the last draw, but those numbers can be any numbers from 1 to 50 (wich are the main numbers from Euromillions). So, for these set of 5 numbers we need theirs transformations into other numbers from 1 to 50. So, each transformation "rule" has allways 50 transformations.

Example:

Transformation rule number 1 (the simplest case: each number converts to itself

1-1

2-2

...

50-50.

In these case whatever numbers are drawn in the last draw, they are kept the same, and we filter all the combinations in the package with that numbers acocordingly. If the set with transformations would be composed of only this rule, then matching 5 would remove all the combinations matching 5 numbers from the last draw.

Then a second rule:

1-2

2-3

3-4

50-1

In this case all numbers drawn in the last draw will be converted from odds to evens and from evens to odds. If in the last draw the numbers were 12,15,27,35,40 then the transformed draw will be 13,16,28,35,41. If the matching five numbers match remove option would be maintained, then all the combinations with 13,16,28,35,41 would be removed from package.

Then a third rule;

1-3

2-2

3-5

4-4

5-7

...

49-1

50-50

In this case, all even numbers would be maitained and all the odds would change between themselves. Again, for the same last draw as in the previous exemple 12,15,27,35,40 the transformed draw would be 12,17,29,37,40.

What i discovered is that normally from this transformations and other i created is that normally no more than 3 numbers are drawn from each of these

transformated set (each one transformation corresponding to a line in a transformated set). This can help me to populate the package with more probable numbers to outcome)

As for the "Do you want to have separate 'Match x' settings for each transformation rule or should those match settings apply to all transformed tickets?", it should apply to all transformed "tickets" or what is the same all the derived combinations from the last draw.

And for "It would be probably faster to have a filter that applies the transformation rules on the fly instead of generating transformed tickets and then compare those with package tickets", i agree tottaly.

I hope i could make it clear what is that i´m looking for, because i need a way to introduce the transformations and save them. It just takes, for each rule to be created, a board with 1 to 50 numbers with a empty square in front of each number to convert into another number. Then save the rule under some name, then aks "Do you want to save more transformation rules?" if yes, then another identical window/board with 1 to 50 numbers with correponding spaces ahead to introduce the transformed number, save this as a 2nd rule, etc, etc. The filter part would read the last draw main numbers, apply each of the transformation rules, and use the result to filter. Maybe this imply the creation of a file ( where each line would correspond to the application of a rule of transformation) to filter from package, or maybe just a temporary file just to as you say "applies the transformation rules on the fly" and compare them to with the package tickets. Either way it is necessary a interface to create and save the transformation rules.

I let to your best judgement the better way to achieve this. Maybe i didn´t understand what you meant by the use of a file text. Are you suggesting a file with some indications as i refered up is enough, like

# Tranformation rule 1

1-1

2-2

...

50-50

#Transformation 2

1-2

2-3

...

49-50

50-1

If this is the case i much prefer an interface like the one we use with "Numbers Group" , inside EL, which allow us to create and manage numbers groups. The difference in here is that instead of "Number Groups" we would have "Tranformation Rules" (or something else that you might prefer to cal them).

But i must stress that if you manage to indicate a way to achieve this with a text file with the appropriate syntax, if it is also simple and ready to use in a filter, if that is easier, you could provided it and we see how it works from there.

Many thanks,

Regards