I'm currently working on reporting methods as my emphasis.
The current method I'm working on is what are the most occuring individual numbers occuring in the remaining combinations. And do those most occuring numbers correspond to what's being picked. Well, some of the results are interesting. If I use a game history range of 2 along with another extended history parameter, I don't get a very good correlation. Of the highest values, I'm only getting one number. But, when I get to a 7 game history, 10 game history, and extended history combined I'm getting two to four of the most occurring numbers showing up in the next draw. So far this is the best. It does make a big difference which game history you use...5 games back, 10 games back...etc.
The reason I'm looking at the most occuring numbers after filtering should be somewhat obvious. Upon, using any filter or reduction method you've arbitrarily said some number are going to come up more than others. It will show in the remaing combinations. And even if you use a random pick technique of the remaning numbers, the more occuring ones will show up more in your answers. Therefore, why not test a small combination of parameters for their accuracy. This methods hopefully will also prove useful in deciding which numbers to wheel in the future.
The work goes on. Trying one combination after another.