I've been reading a number of posts regarding lottery picking systems, and believe the fundamental assumptions are missing a component. Statistical random number theory would indicate you need two variables to increase the probability of picking the highest probability numbers. Most of the posts I see deal with one of the two variables and ignores the other. Either the systems look at frequency (how often a given number is picked), or they deal with proximity (when was the last time a number was picked, the "hot" and "cold" number theory). I'm developing some software which factors both of these variables in, and am open to suggestions for improvement. The basic concept is: if a number is picked 1:10 times, and it's been 9 times, then the number is "due". If it's been 11 times, it's "overdue". So, by weighting proximity and frequency, I believe I can significantly improve the accuracy of my picks, and derive a list of the highest probability numbers. I'm using 11 years of PowerBall data to validate the hypothesis, and so far the results seem pretty impressive.
Any suggestions? Without "gimick" systems, any ideas of things I'm missing?
Thanks,
Jim