Sometimes one post (P1) is commented "possible duplicate of P2". P2 itself is commented "possible duplicate of P3".
Example:
https://askubuntu.com/questions/246610/how-to-increase-battery-duration-hybrid-graphics-notebook
is pointed to
Battery drain on ubuntu is fast,
which is a possible duplicate of
Tips to extend battery life for laptops and notebooks.
What can/should we do to improve the situation?
Follow the rabbit hole, find the canonical question, comment on each question that is improperly commented, flag each wrong comment for removal, vote to close if you can.
If the questions are already closed, so it's necessary to go through multiple levels of duplicates:
A [duplicate] -> B [duplicate] -> C [duplicate] -> D [duplicate] -> E (master)
Then:
The most serious problem is prevented by the system--the system will not allow a question to be closed as a duplicate under circumstances where it would cause the graph representing the "duplicate of" relation to be cyclic. In other words, questions are never closed as duplicates in a cycle that loops back on itself. There is always some master question. (The master question might itself be closed...but not as a duplicate.)
This is not always bad. Referring to the representation above, sometimes D might make it clear why C is really asking something answered by E.
This is often bad but hardly ever serious, because the problem is easily ameliorated by commenting on the duplicates further away from the master to save people the trouble of navigating a long chain.
For example, if A should just have been closed as a duplicate of E, there's no need to reopen and reclose A as a duplicate of E. Instead, just post a short comment on A pointing people to E.
