Google is pretty bad at handling specific technical queries, and I'm pretty sure it's because their internal metrics don't account for the possibility that a qu

Google is pretty bad at handling specific technical queries, and I'm pretty sure it's because their internal metrics don't account for the possibility that a query might have no useful results, or have only one result which requires some iteration on the search terms to find. What happens is, if you search for something that's specific enough that there are few results or no results, it will either ignore keywords, assume that you meant to put a space inside a multi-word identifier, or spellcheck-correct something that wasn't actually misspelled. This produces convincing-looking decoy results, and you have to look closely (or click through) to find out that it's wasting your time, them rerun the same query in verbatim mode or with more quotes. So then you've forced it to verbatim mode, and reached a query which, let's say, has one StackOverflow thread which you've already read and which failed to answer the question. Then your search results will be a couple pages of StackOverflow scraper sites. I never want to visit a StackOverflow scraper site. They should be easy to detect. Why aren't those domains being blocked properly?