When reading reviews I often go straight for the one-stars.

There’s often two clusters of them: high signal or ~worthless.

I discount the latter but update heavily on the former. I’ve found this to be useful for areas with high expertise caps.