White Ball 1::: low=1 high=32 ::: eliminate 33-49
White Ball 2::: low=2 high=45 ::: eliminate 46-50
White Ball 3::: low=6 high=50 ::: eliminate 3-5 and 51
White Ball 4::: low=8 high=52 ::: eliminate 4-7
White Ball 5::: low=10 high=53 ::: eliminate 5-9
If you see that the game has too many balls and so therefore is very hard to make good reductions, or that with your present reduction methods you can't reduce your picks enough and if you also know that no matter what you do, that you are not going to get the top jackpot prize for that particular game, then I think that it might be best to play a jackpot game that has fewer balls and better chances (chances are WIND) like the cash 5 (pick 5) game.
I might be wrong, but I think that "on the average" even on that particular jackpot game, that you can or should be able to do much better than on the example above.
I have already done the above for the Texas cash 5 game 5/37, but I will not put it here or tell people about it, It does reduce a lot, but of course by itself is not good enough, several reduction methods must be used for meaningfull reduction.
Filter(s) reduction(s) and statistics are tied together, you can't just make any reductions anyway that you want, the reduction methods must be statisticaly sound, the averages, medians and bound limits must be found and used and "filters" must be found and or developed (the main thing).
But filters for any lottery game are best used just for reducing the combos produced by some prediction method(s) and not as a means of getting the winning number just by themselves (with only filters).
Anyhow as always this is just my opinion.
Good Luck.