My gut says that you are much better at math than me so in that you have an advantage. I've just learned what math I needed to as I went along and have spent countless hours brute forcing my way through an equation that could have taken someone else minutes to figure out. I like to say that when I first started I was basically doing the computer equivalent of counting on my fingers. In some cases I may very well still be.
My motivations for posting were very similar to yours as normally I am a fairly private person. And like you I don't like the idea of just picking a number because it happened to appear in the last X drawings.
Ironically enough we are also in the same age range although I'll be 50 this year.
Where we are different is that I have banged my head on this long enough to realize I'm not going to discover some magical cipher that will allow me to predict what numbers are going to appear in the next drawing. I've moved onto trying a method of picking that works for me that offers a realistic chance of picking a winning number in the next 28 drawings as 20 dollars a day.
To me that means making as few controlled choices as possible. As I mentioned before I was a big fan of the by bit method and had worked it out to making a 1 in 6 choice 10 times. The problem with that method is that every choice had to be exactly right and that even getting 9 out of 10 right could potentially yield no money. The other downside of that method is that by pick 5 there was usually only one possible winning combination.
The way that I got it down to 1 in 6 per position was to pick an ideal set of bits form the previous drawing and look for combinations where two bits changed.
Here is a screen shot of how that worked.
The yellow represents the bits that were changed.
Right now I'm bouncing back and forth between a method where I divide the 80 numbers into 10 sections of 8 and another one where I look at common drawings.
The problem with the segment method is that winning numbers only occur once every 8 drawings or so. I overcome this by re-sequencing the numbers 20 times which produces winning numbers 99.99% of the time with about 1,100 potential winning combinations per drawing. It sounds simple but there are over 1 billion potential combinations. Divide that by 1000 and it gets knocked down to 1 in a million.
On the flip side if I just played 20 bucks my odds would be 1 in 450K. But so far with some tweaking I've been able to get the odds down to about 1 in 25K before hitting a road block.
I've yet to build a pick tool for this method because like the bit method by about pick 6 or 7 I am locked into trying to hit a very specific pattern although unlike the bit method the is usually more than one path. The problem is that too many of the segment only have 1 possible digit that can be a winner while another section might have 3 or 4.
The common drawing method has potential but right now it involves making too many picks per number. While most fo the time 1 is the correct pick there are about 8-9 instances where it is not and the value can go as high as 15. It is very similar in performance to the pick by split method that I referred you to.
In short what it does is pick numbers by what drawings they share in common. These patterns are fairly repetitive and most numbers can be isolated by identifying 4-5 common drawings. The problem is that there are occasionally numbers that require picking 6 or 9 drawings in order to isolate them.
What I like about that method is that there is no single path to picking a winner. Each pick has multiple potential winner paths. It is just too many picks to have to make per digit to make it viable. I did go as far as to build a pick screen for this method because I wanted to see what the flow was like. Here is a sample of that screen.