The journal already had documented evidence that the legitimate paper was submitted to the journal before the plagerism was submitted to ArXiv. I don't know why they would trust an ArXiv timestamp more than their own.