I completely agree with the OP. Questions often get marked as duplicates just because they look kind of the same to a transient visitor.
In many cases the "duplicate" post asks a different question about the same topic, but it gets shut down so the question never gets answered.
For example, recently I asked why Debian was backpatching openssl and how I could tell if the patch fixed a certain problem; the question was marked as a duplicate of another which asked how to upgrade openssl, which I already knew and was completely tangential to my question. Luckily, someone had already posted an answer for me before the duplicate wrecking crew arrived, but many other posters are not so lucky.
Many people essentially lose their question and never get answered just because their questions bears some resemblance to some other question, even if the answers to that question in no way answer the "duplicate" question. This is a significant source of frustration for a lot of posters.