Two days ago I submitted my first, recently accepted, paper to the widely used preprint server arXiv/astro-ph. People have studied the dependence of long-term citation rates of astrophysical papers put on astro-ph as a function of position in the daily listing. Even astro-ph seem to have realized that and offer a page where part of the relevant information to be first in the listing is published. They do not tell you, however, at which time one has to submit to be first and how resubmissions are treated. So I had to find out myself.

I waited until the evening (22:00 CEST = 16:00 EDT) with uploading my paper to the server and finally submitted it at 22:02:00 CEST. However, due to differences in the TeX system on astro-ph and on my computer, the paper didn’t appear as it should have and I had to edit a number of things. I resubmitted the final, corrected version on the next day, still before the deadline of course, at 16:29:51 CEST.

The paper appeared on position 21 (out of 25) in the Cosmology and Extragalactic Astrophysics section. This is what Konrad and I conclude:

• Papers are sorted in descending order, the on submitted last appears on top of the page.
• Numbers are given in ascending order as soon as the paper is the first time accepted by the system (later submission means higher number) and the date and time when you change your paper after the first acceptance by the arXiv system does not have any influence on the position of the paper

The best time to have a paper accepted by astro-ph is therefore 21:59:59 CEST.

Update: As Konrad found out, the inverse is true for the position of the paper in the daily mailing. So if you want to be first in the mailing, submit at 22:00:00 CEST. If you want to be first on the web page, submit at 21:59:59 CEST.