Not only is the test limited, it is entirely irrelevant. A major advantage of (La)TeX is that the journal has a good LaTeX template in its Instructions for Auth

Not only is the test limited, it is entirely irrelevant. A major advantage of (La)TeX is that the journal has a good LaTeX template in its Instructions for Authors, and you almost don't have to think about formatting. Certainly not about formatting references. These guys come across as high school students criticising Apple for having a flawed product design approach. The fact that they've never heard of version control systems says it all. I would like to propose a similar test, much more relevant to what scientists do in their jobs, where LaTeX would win hands down: Your paper was rejected from Journal A. You now want to submit it to Journal B. Your task is to convert the manuscript ao that it fits the Instructions for Authors for Journal B.