Gatsby wrote:Hi Stan and Jerzy.
for sometime i´ve been advocating the use of circularity, or if you want, circular patterns. This is the case for a permutations builder, but it applies also in this filter. If you look at a pattern like +++-- and aplply it to the last draw you allways will have combinations to remove, but it must permite circularity and rearranjement of those combinations:
Example:
Last draw is: 1, 13, 22, 24, 45
applying the movement (+++--) will give:
2, 14, 23, 25, 46
3, 15, 23, 26, 47
4, 16, 24, 27, 48,
5, 17, 25, 28, 49
6, 18, 26, 29, 50
but also:
7, 19, 27, 30, 1 or what is the same: 1, 7, 19, 27, 30 *
8, 20, 28, 31, 2 .......................: 2, 8, 20, 28, 31 *
These are combinations that are worth to remove (that is 5, 4 or 3 numbers from each of them), as the rule icontinues the same. If you try this transformations and applied to your last draw and look at the next, you will see that normally only 1 or 2 numbers (exceptionally 3) appear in that next draw. Mind that this also is observable for the combinations identified by * ... as if they were part of the same family of transformations, so i think they should be considered.
Sorry if english isn´t very good or to my difficulty to express an idea in a more clear way. I think, though, this is something to explore. A filter that looks at the movement shouldn´t consider this? Isn´t this logical?
Regards,
F.
If after your example (1, 13, 22, 24, 45) is to propose a new draw (3, 10, 23, 29, 41) then we have:
47, 09, 18, 20, 41 => 1 number
48, 10, 19, 21, 42 => 1 number
49, 11, 20, 22, 43 => 0
50, 12, 21, 23, 44 => 1 number
01, 13, 22, 24, 45
02, 14, 23, 25, 46 => 1 number
03, 15, 24, 26, 47 => 1 number
04, 16, 25, 27, 48 => 0
05, 17, 26, 28, 49 => 0
06, 18, 27, 29, 50 => 1 number
where we have tickets to 0, 1 numbers of proposed draw, while filters Adjacent Numbers and Knight's Move will choose even itself draw 3, 10, 23, 29, 41. In conclusion, your proposal is not more efficient than existing current filters in EL 5.