Most of the time I use Winning Numbers History to find what works for some game, find my method, then I  consider it to be a sample,  and based on it I can play real game in hopes trend works. However, sometimes this trend works only that one time, which was a sample. From here I concluded that it is OK sometimes to come up with method and instead of creating sample,  simply play it. Say I play it for real, I spend 60 or 120 Straight tickets and it did not work. I can play second time and suddenly it works great I get $500. After the win it is wise to stop given trend, because most of the time they are short- lasting. 

Say, I had 

490

203

My method I use without preliminary successful calculations : 490–203=297  I calculate all possible  arrangement of those numbers: 29a, 27a, 97a, then I flip them 92a, 72a, 79a; a29, a27, a97 and flip a92, a72, a79; 2a9, 2a7, 9a7, then flip 9a2, 7a2, 7a9. That is 180 tickets. Too much. I most likely will skip and not play.

Next game ends up being 733, so I would loose if I played.

I then calculate for the next game by the same method: 

733

490

733–490=363   Possible arrangements I prefer:  36a, 33a, flip 63a, 33a; a36, a33, a63, a33; 3a6, 3a3;  6a3, 3a3. 120 Straight tickets.  Next game combination is 367.  my formula 36a worked. $500.

After the win I'd stop and move on to another invented trend.

What if I took the risk and played both games. I spent $180 and $120, or $300, I still won $500 or got $200 pure profit.