Hi Kimberlee,
at this moment there are not enough rules "known".
One thing that the algorithm
btw, i put them on the board very late, for some states i was too late.
The whole thing was "under construction" and i had to in a hurry make it running again to even be able to post on the board.
And, it kinda failed on the other states.
But i still have to go over them one by one to see why it brought, in most cases, the pairs correctly, but not the third digit, and vice versa.
This is still a job that needs to be done with pencil and paper, so it'll take quite some time.
One condition however is very strong (i have written it before, but it is very important, or at least up to here it is utmost important): Repeating digits
Assume, for example, the serie is 123
The last draw was 275 and the draw before was 829.
Now, we are dealing with 2 strings: the serie and the "not serie", meaning:
- 123
- 0456789
One of the series will bring a repeater to the draw. That is almost a certainty!
In the example it might happen that the next draw is : 035.
It appears that there has been 1 repeating digit. In fact there have been 2.
But, the serie should be concidered as 1 entity, created of 3 different digits.
So, we can see then that there in fact has been 2 repeaters, namely digit 5 and the serie itself --> digit 2 turned into digit 3.
This means that we know, sort of, (unless the serie ends ofcourse) what will come to the draw:
- digit 2 came two draws in a row. The rule is that it will not come back the next draw (rules get violated lol)
- digit 1 and/or 3 will come to the draw.
- A repeating digit from the previous draw (a digit that does not belong to the serie)
This meant the following in this example:
Draw 1: 829
Draw 2: 275
Draw 3: ???
Not a 2, so a 1 or 3
Thus far, relying on repeating digits: 5 or 7
Created primary pairs: 1 & 5, 1 & 7, 3 & 5, 3 & 7 = 15, 17, 35, 37
Now, the algorithm, as it is now, will go through history and collect all hits and skips of the remaining digits: 0, 3, 6, 8 and 9
So far it pretty much picks out the "longest out" digits to create the numbers.
This is only 1 condition ofcourse.
Here's another one: doubles
Has a running serie created doubles, full numbers and not full numbers?
Full number would be: 112 (all digits from the serie)
Not full number would be: 118 (only digit 1 is from the serie)
I noticed that if 2 of the 3 digits have been the matching pairs in Full number doubles, the third digit from the serie is desperate to create a double too where it will be the digit to create the matching pair. But there's a trap, sometimes the serie ends for 1 or 2 draws and then that digit shows up again to make the Full number double. (i have no idea, yet, how this can/could/would no, should be anticipated on)
Now, to make things more complicated, the algorithm must deal with several running series at the same time.
This can be a downside, but it can also be action that will make sure the correct number is created.
It is complicated and it is simple at the same time: the algorithm must look for the serie that meets the condition, and then it must decide what serie to ride with...
To check what data the algorithm is producing and looping through is... well, it is pretty much a "no done job", i have set counters to it and you'd be amazed how many numbers it creates and filters out again, because the condition of another serie tells it to.
I think this midday was luck, i hope to get the code back together by this evening so i can post new numbers, but i wouldn't expect too much of it.
There's tons of research left... but if it only comes up with 1 digit per draw, and it is the correct digit, than that would be more than enough to go with. (the critic inside me screams out that that is impossible lol)
cheers
Ricky