So, here's a good example of a decision, just out today, that is an example of the types of things a UDRP panel might consider, when dealing with marks that were not registered at the time the domain name was acquired:
https://www.wipo.int/amc/en/domains/decisions/pdf/2023/d2023-2594.pdf
The difficulty confronting the Complainant is that, generally speaking, a finding that a domain name has been registered and is being used in bad faith requires an inference to be drawn that the respondent in question has registered and is using the disputed domain name to take advantage of its significance as a trademark owned by (usually) the complainant.
However, the Respondent acquired the disputed domain name a month or so before the Complainant was incorporated on March 30, 202,2 and even before the Complainant’s associated company applied to register the trademarks on March 28, 2022.
Perhaps recognising this difficulty, the Complainant has pointed out that Messrs Ovalle, Ibanez, and Delgado started a WhatsApp chat group under the name “Stable” on January 20, 2022. As Annex R to the Complaint (which is the proof the Complainant advances to corroborate the date the WhatsApp chat group was started) states on its face, however, WhatsApp “[m]essages and calls are end-toend encrypted. No one outside of this chat, not even WhatsApp, can read or listen to them. ….”
The Panel considers this feature of WhatsApp is well-known.
The Complainant has not made any attempt to suggest that the Respondent was, or became, a member of the WhatsApp chat group before the Respondent paid over USD one million to acquire the disputed domain name.
There does not appear from the record in this case to have been any other public circumstance from which the Respondent could have become aware of the plans for the Complainant’s project. Nor has there been any suggestion that the Respondent, or someone associated with it, was privy to information about the plans to use “Stable” being developed by the Complainant’s backers.
---
So, a UDRP panel will look at "how well known or widely used was the mark at the time" and "what is the likelihood the Respondent would have known of that use".
This was also a case where the domain name is a single short dictionary word - stable.com - so the strength of the evidence showing a bad faith intent by the Respondent would also have to be proportionately higher than if it was an obscure or made-up term.
Notice how that key part of the decision makes no reference to anything having to do with trademark registration.
I hope that someday, domainers and the "someone told me" consultancy gives up their insane and usually irrelevant obsession with trademark registration databases.