Why will you look back when the keyword is Prediction(looking into the future).Have it occurred to you that initiators of most system threads hardly predict the next picks. All am saying is, if your think add 999, -1,+1, mirror, etc works, then do some predictions with few picks. Validating a system after the fact is delusional, bias. The ideal about date, month etc does not hold, why? Read the rules of any lottery game !
The draws of yesterday has no impact on today's draw!, this simple statement is place to start analysing data.
Let's take P3 draws for example, do you understand the rules of this game?
- How much data are you analysing, 1000,500,200 etc ?
- Why do you think 1000,500 etc data is relevant to the above statement?
- What is your 'PRIOR' Data, when every draw start with 1000 combos
I do understand people trying to defend their ideal, but analysing a 'control system' against a ' surname' of data is delusional ! .
Remedy:
- Flip your own coin
- Create your own data
- Create your own patterns
These three points are used by many business, Banks use randomized trials to woo customers for credit cards, Amazon used to guess your purchase pattern' People who bought this also like...', you get the drift.
When you flip a coin several times, the distribution overtime will even out, in randomize trial, you are not just flipping for flipping sake, you asking why H or T is behaving certain way, so you come with different assumptions for H and T, you create patterns for H or T irrespective of current flip trend. The same thing for any lottery . I create my own patterns irrespective of draw trend(Rules of the game), then I focus on triggers (it could be parameters like sums, digits, etc).