But you are assuming we know
all about nature and natural laws, the rest of us are assuming that we don't.
If everyone looked at what they know right now and say "yep, that's enough for me, that's all there is" then society in general and science in particular would go nowhere.
I have heard it said that the phrase a scientist most wants to hear when conducting an experiment/research is not necessarily "hooray! I've proved it" but rather "hmmm, that's strange" as that is what leads to further investigation and even more discoveries.
But that is fine, there are many people who don't like to challenge/question what they know so you are certainly not alone.
The rest of us can enjoy terrifying others around the bonfire and have some fun.
I don't believe in something until there is evidence for it. You are mixing up your examples here.
A scientist doesn't get excited hearing about things with a very long history of having no evidence, or having evidence to the contrary.
Even to form a hypothesis there has to be some evidential basis. Scientific journal articles tend to begin with a literature review.
It would be very bad science to, for example, believe something without evidence and that goes against all established evidence for the laws of nature and then proceed in that belief without testing your beliefs and seeing whether or not there is evidence for that belief.
A scientist has an open mind based on logical hypothesis, logical observation, and previous established evidence. And then they will seek to make their hypothesis falsifiable so it can be proven incorrect.
If possible they will operationalise the variables, the independent variables that you will manipulate and the dependant variables that you will measure. They will then seek to find a correlation between those variables and eliminate extraneous variables in order to ascertain a theory around causation.
Extraneous variables are just "mess" variables. Stuff that affects the dependant variable (what you are measuring) but isn't what you want to test and manipulate.
For example in the ghost example, seeing things other people aren't seeing. To establish the relationship between seeing such a thing (the dependant variable) and the correlating independant variable i.e the presence of a ghost, you would have to rule out all extraneous variables. You do this by measuring them and seeing if they correlate with the dependant variable.
In this case you would need to get a mental health assessment. Measure for carbon monoxide leaks. Test for evidence of drug use. Test for evidence of issues with diet. And measure all sorts of other things such as stress levels, lack of sleep etc... etc... As all these extraneous variables can impact on the dependant variable.
Then a scientist, once all these "known" extraneous variables are accounted for, will proceed with a theory. If the method is good enough, it will get published in a scientific journal. Other scientists will then seek to replicate the previous findings and perhaps measure other things also in order to find out whether or not there are any other extraneous variables.
A scientist will have to have a large amount of evidence for ghosts before suspending all the decades of scientific evidence against ghosts.
Edit: Sorry that's all pretty long winded. To put it very simply, what you are describing i.e a scenario where scientists should get excited about the idea of ghosts, is similar in it's issues as an idea that scientists should get excited about the idea of rocks curing cancer.
(1) It goes against established laws of nature
(2) There is no scientific evidence for it
(3) Even if there were a case where a cancer patient recovered after being given a rock, it's a safe assumption that there is most likely another explanation for the recovery.
However a scientist would start to get excited if despite all of this, a certain type of rock consistently led to patients recovering from cancer and all other explanations were systematically ruled out.