How does one, in the course of documenting what is known on a topic, "accidentally find out something that wasn't known before"?
IIRC something like that happened with [[en:color blindness]]. But there are several other possibilities I guess, especially when Wikipedia articles will delve deeper and deeper into subjects.
If one does indeed have brand new ideas about a subject, these ideas warrant the testing and maturation that comes with peer review and publication before they warrant being repeated in an encyclopedia.
Yes, that's why I think Wikiresearch is a good idea. It combines peer review and publication.