James, I think you have hit upon my biggest issue in genealogy. When do you decide something is a chimera and end your research?Unfortunately, it is not uncommon for online family trees to included unsubstantiated ancestral connections. First of all, it is important to look at the dates involved in this particular issue. No matter what country this particular question refers to, research back into the 1600s is going to start encountering a significant lack of source material, particularly if the researcher is searching for evidence of a specific event, such as a baptismal record. Normally, we would search for alternate records to support a calculated date, but in this time period the options may be extremely limited.
I have a family patriarch who was baptized in 1671, according to my family's records. I'm going back and verifying all of the links, and I'm really stuck here, as I can't find the baptism in question, which calls into question the rest of the research the family has done, linking back to the ascending three generations.
Do I just end the line and say "This is what the rest of the family believes, however, data does not exist to prove its veracity" or do I keep going trying to find that piece of data to make the system work?
In this case, it appears that someone has entered a date without a source. In my opinion, any unsourced information in an online family tree is pure speculation. The burden of supporting the date in question is not on the commentator, but on the person who entered the date in the first place. Unless there is substantiating evidence in sources that support this particular connection, the line is unproved and should be discounted. The researcher here can "keep trying to find that piece of data" but I would suggest that it might be more productive to step back a generation or more and see if the rest of the line is well documented. If it happens to be documented, then the focus should be on finding more supporting data for the target individual's descendants and family rather than focusing on the narrow issue of one event.
I hope this helps clarify this type of question.