Has anybody tried this approach for Pick5 type games?
How many times did each number occur in..
I chose the 15 game spread above arbitrarily. So the first thing I'll try to do is to find the optimum spread where a consistency can be found. Then use the info for the last equivalent game spread to determine what is due.
It's been a while, but what I've done in the past is analyze blocks of 5 and 10 draws prior the the current draw, to look for patterns.
Most often in the blocks of 10 previous draws, between 5 and 10 of the 39 balls were not drawn. Typically only 2 or 3 of the winning numbers in the current draw came from the past 10 draws. The was an occaisional 4 of 5 and only one or two 5 of 5 (going back about 300 draws).
You might be interested in the following article: Frequency Distribution.
I post a similar article earlier. I assigned letters to the earliest previous drawing that each number in the winning combinations has appeared. For example A=previous drawing, B=2 drawings back, J=10 drawings back, a=27 drawings back and etc. I then look for the most common words (A-S) to pick combinations to play. I used the words because they do not focus on numbers and the numbers for my combinations fit similar words but are totally random. With MagaMillions, I haven't had much luck with an affordable amount of lines, but for Ohio RC5 a 5/39 game, I've had some decent returns with only 25 lines. I haven't won the big one yet, but my cost of playing has gone down. Right now the RC5 jackpot is $476K and the odd of winning it is 1:576K so now is a good time to test the theory to its maximum. The following is data for the MegaMillions last 2 months drawings.
Good luck on your efforts.
For the 6/49 I go back 4 or 5 draws and count how many times each number has hit; & play the lowest hitters. Often get 4 numbers that way.
With a 6/49 games, using only the previous 4 or 5 drawings means you're only including a maximum 24-30 of the 49 numbers in your number pool. If the pool have any repeats, it's even smaller so you should only have a few lines to play.
Thanks everybody for your input.
For 6/47 here is one of my summaries over 201 games. The top horizontal row identifies lottery ball (1-5). The vertical column identifies how many times each occurrence occurred. So, Paint1, your assumption appears to be basically correct. My chart below would say that the best bet over a 5 game span would be 0-times, 0-times, 0-times, 1-times, and 1 to 2 times for the last ocurrence value. This proves to be a valuable filter, however it also leaves you with a very large amount of combinations to choose from. I typically use 3 different game ranges using the analysis as below as a filter. But, it isn't enough, because I have been unable to find a method to make the correct numbers somehow spike so that I can identify them.
By keeping track of the last time each number occurred from actual picks for 5/39, I have the follow chart. The top horizontal row is each lottery ball (1-5). The vertical row is how many games back +1, each number hit. As you can see 23.5% of the actual numbers picked occurred in the last two games. Even though I track this way, I have yet to find anything really useful to do with it. There is so much varience in the data, that it is hard to lock anything down reliably.
I like your alpha substitution method as a way to identify patterns.
I guess I didn't communicate very well. Give me some time for programming and anlalysis and we'll see what comes out of my original comments.
I would have you consider this:
What is the best game range to use or sets of game ranges to use?
How exact or generalized should a filter/reduction method be?
Perhaps all the tools to pick the right lotto numbers are already here at the LP, but the right way to summarize the data isn't. There are so many methods, but none of them are making the correct numbers somehow standout from the rest. So you are left with guessing or quick picking from a large amount of filters/reduced combinations.