NMD
It seems that most of my free time goes to analyzing data looking for stuff that can be used
day to day. Filters can be just about anything but like you say, there are filters and then there
are filters. Some change wildly day to day and are hard to pin down. The consistent filters are
easier to hit but don't have quite the filtering effect. I guess it breaks down to what the filtering
needs are for the sets being filtered. Back several years ago I had the idea to auto-filter the
lexie based on groupings. I still think it's possible but coding it makes me want to pull my hair
out. It's a sorting tool that then focuses on one group out-putted by the sorting algorithm.
The optimizer or auto filter included in many of my software's is just the first stage of the new
filter but is geared towards a one off hit. Another tool that I don't have time to work on.
Maybe someday I will finish it but it's way down on the list. IMHO the ultimate software won't
need filtering. I don't like the word prediction when coupled with lottery but I do think there is
a smarter method lurking around somewhere. AI's, NN's and a host of other like minded tools
have their place in prediction but won't predict random. In the math forum I see stoopendaal's
topic is still going where people are setting/adjusting around 10 values trying to improve the lines
generated. I could build a dozen programs where 10 user inputs would reduce the matrix to a
single line winner. Generate all the lines for your 5-36 and then pass them through the stack
filter setting the first column to single values. Or better yet set the first bank of filters in DMP all
to single values and see what happens. Here are the results of a 10 filter 5-39 run. All but 3
of the lines generated would have won a prize. Is it easier to set stoopendaal's AI programs
variables? Even is set correctly will it produce the winning set in 26 lines. Anyway, running on
here so I will close.
RL