> Good journals, in theory, are more selective about what they publish

It should be noted that this selectivity is not just on quality, but also e.g. on how "groundbreaking" the results are. This results in papers in top journals actually being retracted more often than in others, since they're more likely to be false.