What we do know is that the reviewers disagreed with the conclusions of the paper. That might have been a red flag to a careful editor.
Not all editors are careful. Some have an agenda; some are just careless.
A paper that I reviewed some time ago was still published even though I had recommended against it; there were incorrect mathematical statements; some made it into the published version. The authors tried to create a biased stream of bits by combining a random string (probability .5 for 0 and 1) with blocks of 1-bits. They claimed this would allow them to get any probability (false, exercise for the reader), and that the stream would look like random stream with non equal probabilities (false, exercise for the advanced reader.) This had little to do with the main point of the paper (parallization of some Monte Carlo code), but one never knows what a reader will try to use from an article.