Double spend has a well defined meaning, so BitMEXResearch saying there was a double spend is wrong too. If double spend means the second one above then every RBF is a double spend and their never was a "double spend" problem to solve.
in this case there was a block reorganization, so a confirmed transaction (1 confirmation) was dropped from the blockchain and a conflicting transaction was mined instead. that doesn't apply to every RBF transaction.
...
Agreed of course, but every reorg drops every transaction from the losing block which then may be, or may not be, included in the winning block or subsequent blocks. Likewise every RBF transaction that works replaces another transaction, just ones that could have been in the mempool or in a block like this one. The confluence of both here then being called "a successful double spend" is the issue when by its very nature the block chain prevents double spending.
Anyway, I agree with what you are saying.