If 6+ seasons of data collection isn't enough, then why bother at all? You've naged and moaned about Roger Bernadina since before the All-Star break and he didn't have any more PA's. If 300+ PA aren't enough to spot a trend, then maybe the data isn't being collected or analyzed properly. But to say 6 years isn't enough time is ridiculous.
I'm not a huge fan of 338 PA sample sizes and making conclusions off of them, but 338 PA in one season is infinitely better than 338 PA over 9 seasons which is what Crawford has played. That's an average of 38 Random PA/year over a 9 year span. You can see where the problems come. I naged about Bernadina because he hasn't been good whether it is in the big leagues or his indistinguishable minor league career. He hasn't shown anything. If someone shows that they are a good player over a 5383 PA sample size, I'm not taking a random 338 PA from that sample (6.28% of sample) against pretty good pitchers to make conclusions.