The hitting analogy can be expanded further. I would argue a .295 hitter is the same as a .300 hitter with the caveat that there are numerous other stats to help separate them. I think you agree a .295/.380/.550 hitter is better than a .300/.330/.415 hitter.
As with CERA, there is a danger with small sample sizes regarding opponents. A pitcher and catcher who have 100PAs together against the three worst offensive teams can be very misleading. Or in 3 pitcher-friendly parks. Like all stats, we hope the larger sample sizes even it out. But like with a lot of other stats, it doesn’t always.
And more important, fearing the unknown and accepting the status quo simply because it worked once is dangerous. Cherington found this out in 2014, and it lead to bad decisions that plague the Sox to this day...