Your point is correct, but I'm not surprised by the difference. I think "legibility" is the term of art here. Writing a paper like this makes it almost[1] certain to the institution he is applying to that he is capable of writing a paper of this quality, while all the other metrics (GPA, GRE, etc) are much more probabilistic. Since someone incapable of writing such a paper is probably unsuited for a PhD, it seems entirely appropriate to choose applicants who have demonstrated ability to clear this bar over those that have not.
[1] "Almost" to account for the slight chance that he didn't actually author the paper but somehow managed to get his name put on it anyway.
[1] "Almost" to account for the slight chance that he didn't actually author the paper but somehow managed to get his name put on it anyway.