Good question! I've been asking myself the same thing, but haven't been able to come up with an answer that I like!!
Since October of 2015, #18 is ranked second with 29 hits, but when you factor in 13 of those hits came in 2020, then you can see that it hasn't been all that strong since 10/7/2015 when the matrix last changed.
The top ranked PowerBall since October of 2015 is #24 with 30 hits. I like #24 because in 2016, and 2017, it was the third ranked PowerBall, in 2018 it was the top hitter with 8 hits. In 2019 it cooled off, but came back strong in 2020 with 8 hits. (#24 is ranked second in 2020. I'd also point out usually 8 hits in any year is good enough for a number to be the top hitter)
I also look at a moving or "rolling 104 drawings". For instance the first drawing in July until the last drawing in June of the following year. Typically the top ranked hitter in 104 drawings will have 7 or 8 hits.
It's fair to say that a sampling of 104 drawings is arbitrary and isn't considered by probability experts to be a large enough sample. I chose it simply because it was a convenient way for me to compare calendar year over calendar year differences. The roots of that come from my career when I was always asked by exec's to provide them with charts and graphs comparing the current years results to the previous year's. That's OK to do in business, but it may not work well for to statistics and probability analysis.
Fair enough, but who cares??? I obviously don't. And I sure as Hell will care even less if I win a lot of $$$ by doing what I do. G5