At the moment I'm working on the numbers at position filter.

I use the previous total historical appearances of each number.

I then sort the numbers according to how many times they have appeared.

Now as an example.

1st position is for the ten numbers which are at sorted positions 11 to 20

2nd position is for the ten numbers which are at sorted positions 31 to 40

3rd position is for the ten numbers which are at sorted positions 1 to 10

4th position is for the ten numbers which are at sorted positions 31 to 40

5th position is for the ten numbers which are at sorted positions 11 to 20

6st position is for the ten numbers which are at sorted positions 21 to 30

There are 4096 possible ways for this happen. 4x4x4x4x4x4

The example I tested wih gave me around 1000 combinations.

Unfortunately there is only one winning way with one set of lines.

So I have added one to every number in the history to give another set of numbers.

For 1st draw 3-14-27-29-32-40 the 1st draw in the second set would be 4-15-28-30-33-1

and the 1st draw of the 3rd set would be 5-16-29-31-34-2.

I do this 40 times so I have 40 unique sets of history lines which means in a perfect world it

would take 102.4 draws for all ofthe 4096 ways to occur.

I expect that I will be left with around 40,000 combinations or probably less due to repeat combinations occuring.

If I used 80 random sets of lines that would of course mean 80,000 combinations and the expectation to hit within 50 draws.

I guess random sets would be just as useful for what I'm doing as the 40 sets like I have created above.

I would probably tweak that 80,000 with elimination filters to remove lines I don't expect to be drawn and then use the Package Reduce Filter

to give me a systematic selection of lines - the number of lines I want to play.

Well, that's my fun for the moment.

Probably time wasting by me but I'm determined to find something that might win from any one of multiple ways.