I thought about that more than a decade ago, but:
Georgia (GA) Cash 3 Midday
Tue, Feb 09, 2016 1-1-9
Mon, Feb 08, 2016 0-2-5
Sun, Feb 07, 2016 4-5-0
Sat, Feb 06, 2016 6-2-6
Fri, Feb 05, 2016 4-7-8
Thu, Feb 04, 2016 9-3-5
Wed, Feb 03, 2016 8-0-6
Tue, Feb 02, 2016 1-1-8
Mon, Feb 01, 2016 0-1-6
Sun, Jan 31, 2016 7-9-2
Sat, Jan 30, 2016 4-8-2
Fri, Jan 29, 2016 5-4-0
Thu, Jan 28, 2016 1-4-4
Wed, Jan 27, 2016 0-6-4
Tue, Jan 26, 2016 7-2-7
Mon, Jan 25, 2016 9-1-4
Sun, Jan 24, 2016 4-0-5
Sat, Jan 23, 2016 9-7-5
Fri, Jan 22, 2016 2-4-9
Thu, Jan 21, 2016 8-3-1
Wed, Jan 20, 2016 1-2-9
Tue, Jan 19, 2016 0-1-7
Mon, Jan 18, 2016 6-9-6
Sun, Jan 17, 2016 4-5-2
Sat, Jan 16, 2016 7-3-1
Fri, Jan 15, 2016 0-5-6
Thu, Jan 14, 2016 2-1-9
Wed, Jan 13, 2016 9-4-3
Tue, Jan 12, 2016 9-0-6
Mon, Jan 11, 2016 4-9-3
Sun, Jan 10, 2016 2-3-0
Sat, Jan 09, 2016 5-1-2
Fri, Jan 08, 2016 2-6-4
Thu, Jan 07, 2016 2-0-1
Wed, Jan 06, 2016 9-2-4
Tue, Jan 05, 2016 2-9-3
Mon, Jan 04, 2016 4-7-8
Sun, Jan 03, 2016 0-6-4
Sat, Jan 02, 2016 0-3-4
While it wasn't too bad there by itself, when added to other filters failures, well, failures of course accumulate, after all, we might use several filters and each of them will sometimes have failures.
Since we were using just singles then about 4 failures on 38 past draws, but each time that a Triple and a Double comes out that is also a failure as we would not win, so there were no fewer than 5 Doubles there so 4 + 5 = 9 Failures on about 38 past draws, while that is not too bad, it really is, cause the advantage that you might get by filtering out about 200 straight numbers you might lose when you don't win and there would be many more failures as more filters-patterns are used, still a very good try, sometimes might work better than at some other times as seen on these past draws, when I looked at it on the Texas past draws very many years ago it didn't look very good, it was a lot worse than what we see here.
-----------